Upload folder using huggingface_hub
Browse files- README.md +60 -0
- all_results.json +8 -0
- config.json +30 -0
- generation_config.json +6 -0
- merges.txt +0 -0
- model.safetensors +3 -0
- special_tokens_map.json +43 -0
- tokenizer.json +0 -0
- tokenizer_config.json +172 -0
- train_results.json +8 -0
- trainer_log.jsonl +0 -0
- trainer_state.json +4151 -0
- training_args.bin +3 -0
- vocab.json +0 -0
README.md
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: transformers
|
| 3 |
+
license: other
|
| 4 |
+
base_model: HuggingFaceTB/SmolLM2-1.7B
|
| 5 |
+
tags:
|
| 6 |
+
- llama-factory
|
| 7 |
+
- full
|
| 8 |
+
- generated_from_trainer
|
| 9 |
+
model-index:
|
| 10 |
+
- name: smollm2-1.7B-oh-dcft
|
| 11 |
+
results: []
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 15 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 16 |
+
|
| 17 |
+
# smollm2-1.7B-oh-dcft
|
| 18 |
+
|
| 19 |
+
This model is a fine-tuned version of [HuggingFaceTB/SmolLM2-1.7B](https://huggingface.co/HuggingFaceTB/SmolLM2-1.7B) on the mlfoundations-dev/oh-dcft-v3.1-gpt-4o-mini dataset.
|
| 20 |
+
|
| 21 |
+
## Model description
|
| 22 |
+
|
| 23 |
+
More information needed
|
| 24 |
+
|
| 25 |
+
## Intended uses & limitations
|
| 26 |
+
|
| 27 |
+
More information needed
|
| 28 |
+
|
| 29 |
+
## Training and evaluation data
|
| 30 |
+
|
| 31 |
+
More information needed
|
| 32 |
+
|
| 33 |
+
## Training procedure
|
| 34 |
+
|
| 35 |
+
### Training hyperparameters
|
| 36 |
+
|
| 37 |
+
The following hyperparameters were used during training:
|
| 38 |
+
- learning_rate: 5e-06
|
| 39 |
+
- train_batch_size: 16
|
| 40 |
+
- eval_batch_size: 8
|
| 41 |
+
- seed: 42
|
| 42 |
+
- distributed_type: multi-GPU
|
| 43 |
+
- num_devices: 32
|
| 44 |
+
- total_train_batch_size: 512
|
| 45 |
+
- total_eval_batch_size: 256
|
| 46 |
+
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
|
| 47 |
+
- lr_scheduler_type: constant
|
| 48 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 49 |
+
- num_epochs: 3.0
|
| 50 |
+
|
| 51 |
+
### Training results
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
### Framework versions
|
| 56 |
+
|
| 57 |
+
- Transformers 4.48.0
|
| 58 |
+
- Pytorch 2.5.0a0+b465a5843b.nv24.09
|
| 59 |
+
- Datasets 3.0.2
|
| 60 |
+
- Tokenizers 0.21.1
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"total_flos": 2506240622592000.0,
|
| 4 |
+
"train_loss": 0.754048796695819,
|
| 5 |
+
"train_runtime": 27383.3094,
|
| 6 |
+
"train_samples_per_second": 109.726,
|
| 7 |
+
"train_steps_per_second": 0.214
|
| 8 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "/leonardo_work/EUHPC_E03_068/marianna/models/HuggingFaceTB/SmolLM2-1.7B",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"LlamaForCausalLM"
|
| 5 |
+
],
|
| 6 |
+
"attention_bias": false,
|
| 7 |
+
"attention_dropout": 0.0,
|
| 8 |
+
"bos_token_id": 0,
|
| 9 |
+
"eos_token_id": 0,
|
| 10 |
+
"head_dim": 64,
|
| 11 |
+
"hidden_act": "silu",
|
| 12 |
+
"hidden_size": 2048,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 8192,
|
| 15 |
+
"max_position_embeddings": 8192,
|
| 16 |
+
"mlp_bias": false,
|
| 17 |
+
"model_type": "llama",
|
| 18 |
+
"num_attention_heads": 32,
|
| 19 |
+
"num_hidden_layers": 24,
|
| 20 |
+
"num_key_value_heads": 32,
|
| 21 |
+
"pretraining_tp": 1,
|
| 22 |
+
"rms_norm_eps": 1e-05,
|
| 23 |
+
"rope_scaling": null,
|
| 24 |
+
"rope_theta": 130000,
|
| 25 |
+
"tie_word_embeddings": true,
|
| 26 |
+
"torch_dtype": "bfloat16",
|
| 27 |
+
"transformers_version": "4.48.0",
|
| 28 |
+
"use_cache": false,
|
| 29 |
+
"vocab_size": 49152
|
| 30 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 0,
|
| 4 |
+
"eos_token_id": 0,
|
| 5 |
+
"transformers_version": "4.48.0"
|
| 6 |
+
}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c46db117a58d26901f4467675ae53225c005f39ae00bc7e9127a0034e561d455
|
| 3 |
+
size 3422777952
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,43 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"additional_special_tokens": [
|
| 3 |
+
"<|endoftext|>",
|
| 4 |
+
"<|im_start|>",
|
| 5 |
+
"<|im_end|>",
|
| 6 |
+
"<repo_name>",
|
| 7 |
+
"<reponame>",
|
| 8 |
+
"<file_sep>",
|
| 9 |
+
"<filename>",
|
| 10 |
+
"<gh_stars>",
|
| 11 |
+
"<issue_start>",
|
| 12 |
+
"<issue_comment>",
|
| 13 |
+
"<issue_closed>",
|
| 14 |
+
"<jupyter_start>",
|
| 15 |
+
"<jupyter_text>",
|
| 16 |
+
"<jupyter_code>",
|
| 17 |
+
"<jupyter_output>",
|
| 18 |
+
"<jupyter_script>",
|
| 19 |
+
"<empty_output>"
|
| 20 |
+
],
|
| 21 |
+
"bos_token": {
|
| 22 |
+
"content": "<|endoftext|>",
|
| 23 |
+
"lstrip": false,
|
| 24 |
+
"normalized": false,
|
| 25 |
+
"rstrip": false,
|
| 26 |
+
"single_word": false
|
| 27 |
+
},
|
| 28 |
+
"eos_token": {
|
| 29 |
+
"content": "<|im_end|>",
|
| 30 |
+
"lstrip": false,
|
| 31 |
+
"normalized": false,
|
| 32 |
+
"rstrip": false,
|
| 33 |
+
"single_word": false
|
| 34 |
+
},
|
| 35 |
+
"pad_token": "<|im_end|>",
|
| 36 |
+
"unk_token": {
|
| 37 |
+
"content": "<|endoftext|>",
|
| 38 |
+
"lstrip": false,
|
| 39 |
+
"normalized": false,
|
| 40 |
+
"rstrip": false,
|
| 41 |
+
"single_word": false
|
| 42 |
+
}
|
| 43 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,172 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"added_tokens_decoder": {
|
| 4 |
+
"0": {
|
| 5 |
+
"content": "<|endoftext|>",
|
| 6 |
+
"lstrip": false,
|
| 7 |
+
"normalized": false,
|
| 8 |
+
"rstrip": false,
|
| 9 |
+
"single_word": false,
|
| 10 |
+
"special": true
|
| 11 |
+
},
|
| 12 |
+
"1": {
|
| 13 |
+
"content": "<|im_start|>",
|
| 14 |
+
"lstrip": false,
|
| 15 |
+
"normalized": false,
|
| 16 |
+
"rstrip": false,
|
| 17 |
+
"single_word": false,
|
| 18 |
+
"special": true
|
| 19 |
+
},
|
| 20 |
+
"2": {
|
| 21 |
+
"content": "<|im_end|>",
|
| 22 |
+
"lstrip": false,
|
| 23 |
+
"normalized": false,
|
| 24 |
+
"rstrip": false,
|
| 25 |
+
"single_word": false,
|
| 26 |
+
"special": true
|
| 27 |
+
},
|
| 28 |
+
"3": {
|
| 29 |
+
"content": "<repo_name>",
|
| 30 |
+
"lstrip": false,
|
| 31 |
+
"normalized": false,
|
| 32 |
+
"rstrip": false,
|
| 33 |
+
"single_word": false,
|
| 34 |
+
"special": true
|
| 35 |
+
},
|
| 36 |
+
"4": {
|
| 37 |
+
"content": "<reponame>",
|
| 38 |
+
"lstrip": false,
|
| 39 |
+
"normalized": false,
|
| 40 |
+
"rstrip": false,
|
| 41 |
+
"single_word": false,
|
| 42 |
+
"special": true
|
| 43 |
+
},
|
| 44 |
+
"5": {
|
| 45 |
+
"content": "<file_sep>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false,
|
| 50 |
+
"special": true
|
| 51 |
+
},
|
| 52 |
+
"6": {
|
| 53 |
+
"content": "<filename>",
|
| 54 |
+
"lstrip": false,
|
| 55 |
+
"normalized": false,
|
| 56 |
+
"rstrip": false,
|
| 57 |
+
"single_word": false,
|
| 58 |
+
"special": true
|
| 59 |
+
},
|
| 60 |
+
"7": {
|
| 61 |
+
"content": "<gh_stars>",
|
| 62 |
+
"lstrip": false,
|
| 63 |
+
"normalized": false,
|
| 64 |
+
"rstrip": false,
|
| 65 |
+
"single_word": false,
|
| 66 |
+
"special": true
|
| 67 |
+
},
|
| 68 |
+
"8": {
|
| 69 |
+
"content": "<issue_start>",
|
| 70 |
+
"lstrip": false,
|
| 71 |
+
"normalized": false,
|
| 72 |
+
"rstrip": false,
|
| 73 |
+
"single_word": false,
|
| 74 |
+
"special": true
|
| 75 |
+
},
|
| 76 |
+
"9": {
|
| 77 |
+
"content": "<issue_comment>",
|
| 78 |
+
"lstrip": false,
|
| 79 |
+
"normalized": false,
|
| 80 |
+
"rstrip": false,
|
| 81 |
+
"single_word": false,
|
| 82 |
+
"special": true
|
| 83 |
+
},
|
| 84 |
+
"10": {
|
| 85 |
+
"content": "<issue_closed>",
|
| 86 |
+
"lstrip": false,
|
| 87 |
+
"normalized": false,
|
| 88 |
+
"rstrip": false,
|
| 89 |
+
"single_word": false,
|
| 90 |
+
"special": true
|
| 91 |
+
},
|
| 92 |
+
"11": {
|
| 93 |
+
"content": "<jupyter_start>",
|
| 94 |
+
"lstrip": false,
|
| 95 |
+
"normalized": false,
|
| 96 |
+
"rstrip": false,
|
| 97 |
+
"single_word": false,
|
| 98 |
+
"special": true
|
| 99 |
+
},
|
| 100 |
+
"12": {
|
| 101 |
+
"content": "<jupyter_text>",
|
| 102 |
+
"lstrip": false,
|
| 103 |
+
"normalized": false,
|
| 104 |
+
"rstrip": false,
|
| 105 |
+
"single_word": false,
|
| 106 |
+
"special": true
|
| 107 |
+
},
|
| 108 |
+
"13": {
|
| 109 |
+
"content": "<jupyter_code>",
|
| 110 |
+
"lstrip": false,
|
| 111 |
+
"normalized": false,
|
| 112 |
+
"rstrip": false,
|
| 113 |
+
"single_word": false,
|
| 114 |
+
"special": true
|
| 115 |
+
},
|
| 116 |
+
"14": {
|
| 117 |
+
"content": "<jupyter_output>",
|
| 118 |
+
"lstrip": false,
|
| 119 |
+
"normalized": false,
|
| 120 |
+
"rstrip": false,
|
| 121 |
+
"single_word": false,
|
| 122 |
+
"special": true
|
| 123 |
+
},
|
| 124 |
+
"15": {
|
| 125 |
+
"content": "<jupyter_script>",
|
| 126 |
+
"lstrip": false,
|
| 127 |
+
"normalized": false,
|
| 128 |
+
"rstrip": false,
|
| 129 |
+
"single_word": false,
|
| 130 |
+
"special": true
|
| 131 |
+
},
|
| 132 |
+
"16": {
|
| 133 |
+
"content": "<empty_output>",
|
| 134 |
+
"lstrip": false,
|
| 135 |
+
"normalized": false,
|
| 136 |
+
"rstrip": false,
|
| 137 |
+
"single_word": false,
|
| 138 |
+
"special": true
|
| 139 |
+
}
|
| 140 |
+
},
|
| 141 |
+
"additional_special_tokens": [
|
| 142 |
+
"<|endoftext|>",
|
| 143 |
+
"<|im_start|>",
|
| 144 |
+
"<|im_end|>",
|
| 145 |
+
"<repo_name>",
|
| 146 |
+
"<reponame>",
|
| 147 |
+
"<file_sep>",
|
| 148 |
+
"<filename>",
|
| 149 |
+
"<gh_stars>",
|
| 150 |
+
"<issue_start>",
|
| 151 |
+
"<issue_comment>",
|
| 152 |
+
"<issue_closed>",
|
| 153 |
+
"<jupyter_start>",
|
| 154 |
+
"<jupyter_text>",
|
| 155 |
+
"<jupyter_code>",
|
| 156 |
+
"<jupyter_output>",
|
| 157 |
+
"<jupyter_script>",
|
| 158 |
+
"<empty_output>"
|
| 159 |
+
],
|
| 160 |
+
"bos_token": "<|endoftext|>",
|
| 161 |
+
"chat_template": "{% set system_message = 'You are a helpful assistant.' %}{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% else %}{% set loop_messages = messages %}{% endif %}{% if system_message is defined %}{{ '<|im_start|>system\n' + system_message + '<|im_end|>\n' }}{% endif %}{% for message in loop_messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ '<|im_start|>user\n' + content + '<|im_end|>\n<|im_start|>assistant\n' }}{% elif message['role'] == 'assistant' %}{{ content + '<|im_end|>' + '\n' }}{% endif %}{% endfor %}",
|
| 162 |
+
"clean_up_tokenization_spaces": false,
|
| 163 |
+
"eos_token": "<|im_end|>",
|
| 164 |
+
"extra_special_tokens": {},
|
| 165 |
+
"model_max_length": 8192,
|
| 166 |
+
"pad_token": "<|im_end|>",
|
| 167 |
+
"padding_side": "right",
|
| 168 |
+
"split_special_tokens": false,
|
| 169 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 170 |
+
"unk_token": "<|endoftext|>",
|
| 171 |
+
"vocab_size": 49152
|
| 172 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"total_flos": 2506240622592000.0,
|
| 4 |
+
"train_loss": 0.754048796695819,
|
| 5 |
+
"train_runtime": 27383.3094,
|
| 6 |
+
"train_samples_per_second": 109.726,
|
| 7 |
+
"train_steps_per_second": 0.214
|
| 8 |
+
}
|
trainer_log.jsonl
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
trainer_state.json
ADDED
|
@@ -0,0 +1,4151 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 5871,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.005109862033725089,
|
| 13 |
+
"grad_norm": 57.87140382113745,
|
| 14 |
+
"learning_rate": 5e-06,
|
| 15 |
+
"loss": 2.3817,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.010219724067450179,
|
| 20 |
+
"grad_norm": 22.753974610258698,
|
| 21 |
+
"learning_rate": 5e-06,
|
| 22 |
+
"loss": 1.7816,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.015329586101175269,
|
| 27 |
+
"grad_norm": 10.200028852824934,
|
| 28 |
+
"learning_rate": 5e-06,
|
| 29 |
+
"loss": 1.4328,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.020439448134900357,
|
| 34 |
+
"grad_norm": 10.263199576103078,
|
| 35 |
+
"learning_rate": 5e-06,
|
| 36 |
+
"loss": 1.2383,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.025549310168625446,
|
| 41 |
+
"grad_norm": 4.8615267890156915,
|
| 42 |
+
"learning_rate": 5e-06,
|
| 43 |
+
"loss": 1.131,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.030659172202350538,
|
| 48 |
+
"grad_norm": 4.118120498213385,
|
| 49 |
+
"learning_rate": 5e-06,
|
| 50 |
+
"loss": 1.0356,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.03576903423607562,
|
| 55 |
+
"grad_norm": 2.8275806453598338,
|
| 56 |
+
"learning_rate": 5e-06,
|
| 57 |
+
"loss": 1.0006,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.040878896269800714,
|
| 62 |
+
"grad_norm": 2.315792934153761,
|
| 63 |
+
"learning_rate": 5e-06,
|
| 64 |
+
"loss": 0.9655,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.045988758303525806,
|
| 69 |
+
"grad_norm": 2.165672978783245,
|
| 70 |
+
"learning_rate": 5e-06,
|
| 71 |
+
"loss": 0.9467,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.05109862033725089,
|
| 76 |
+
"grad_norm": 1.5488035180810096,
|
| 77 |
+
"learning_rate": 5e-06,
|
| 78 |
+
"loss": 0.9281,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.05620848237097598,
|
| 83 |
+
"grad_norm": 1.7243341087325224,
|
| 84 |
+
"learning_rate": 5e-06,
|
| 85 |
+
"loss": 0.9128,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.061318344404701075,
|
| 90 |
+
"grad_norm": 1.5359895854005425,
|
| 91 |
+
"learning_rate": 5e-06,
|
| 92 |
+
"loss": 0.9031,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.06642820643842616,
|
| 97 |
+
"grad_norm": 1.435850065371256,
|
| 98 |
+
"learning_rate": 5e-06,
|
| 99 |
+
"loss": 0.8922,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.07153806847215124,
|
| 104 |
+
"grad_norm": 1.2790521202391505,
|
| 105 |
+
"learning_rate": 5e-06,
|
| 106 |
+
"loss": 0.8765,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.07664793050587634,
|
| 111 |
+
"grad_norm": 1.5917071745489466,
|
| 112 |
+
"learning_rate": 5e-06,
|
| 113 |
+
"loss": 0.883,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.08175779253960143,
|
| 118 |
+
"grad_norm": 1.4686990635112733,
|
| 119 |
+
"learning_rate": 5e-06,
|
| 120 |
+
"loss": 0.8874,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.08686765457332651,
|
| 125 |
+
"grad_norm": 1.2956583323307207,
|
| 126 |
+
"learning_rate": 5e-06,
|
| 127 |
+
"loss": 0.8411,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.09197751660705161,
|
| 132 |
+
"grad_norm": 1.2357587318309058,
|
| 133 |
+
"learning_rate": 5e-06,
|
| 134 |
+
"loss": 0.8726,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.0970873786407767,
|
| 139 |
+
"grad_norm": 1.1075519305139339,
|
| 140 |
+
"learning_rate": 5e-06,
|
| 141 |
+
"loss": 0.8548,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.10219724067450178,
|
| 146 |
+
"grad_norm": 1.086727474202395,
|
| 147 |
+
"learning_rate": 5e-06,
|
| 148 |
+
"loss": 0.8493,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.10730710270822688,
|
| 153 |
+
"grad_norm": 1.0270014162441539,
|
| 154 |
+
"learning_rate": 5e-06,
|
| 155 |
+
"loss": 0.8545,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.11241696474195197,
|
| 160 |
+
"grad_norm": 0.8505362329385927,
|
| 161 |
+
"learning_rate": 5e-06,
|
| 162 |
+
"loss": 0.8388,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.11752682677567705,
|
| 167 |
+
"grad_norm": 1.109098072362057,
|
| 168 |
+
"learning_rate": 5e-06,
|
| 169 |
+
"loss": 0.8384,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.12263668880940215,
|
| 174 |
+
"grad_norm": 1.2809391093206264,
|
| 175 |
+
"learning_rate": 5e-06,
|
| 176 |
+
"loss": 0.8471,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.12774655084312725,
|
| 181 |
+
"grad_norm": 0.9245822751345097,
|
| 182 |
+
"learning_rate": 5e-06,
|
| 183 |
+
"loss": 0.836,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.13285641287685232,
|
| 188 |
+
"grad_norm": 0.8873874884959316,
|
| 189 |
+
"learning_rate": 5e-06,
|
| 190 |
+
"loss": 0.8283,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.13796627491057742,
|
| 195 |
+
"grad_norm": 0.6839627248028421,
|
| 196 |
+
"learning_rate": 5e-06,
|
| 197 |
+
"loss": 0.8191,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.1430761369443025,
|
| 202 |
+
"grad_norm": 0.8006134542844254,
|
| 203 |
+
"learning_rate": 5e-06,
|
| 204 |
+
"loss": 0.82,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.1481859989780276,
|
| 209 |
+
"grad_norm": 0.7249813175028869,
|
| 210 |
+
"learning_rate": 5e-06,
|
| 211 |
+
"loss": 0.8254,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.1532958610117527,
|
| 216 |
+
"grad_norm": 0.611471064933341,
|
| 217 |
+
"learning_rate": 5e-06,
|
| 218 |
+
"loss": 0.8148,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.15840572304547776,
|
| 223 |
+
"grad_norm": 0.8403753613185879,
|
| 224 |
+
"learning_rate": 5e-06,
|
| 225 |
+
"loss": 0.8091,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.16351558507920286,
|
| 230 |
+
"grad_norm": 0.5250234838001376,
|
| 231 |
+
"learning_rate": 5e-06,
|
| 232 |
+
"loss": 0.8328,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.16862544711292796,
|
| 237 |
+
"grad_norm": 0.5411140081911021,
|
| 238 |
+
"learning_rate": 5e-06,
|
| 239 |
+
"loss": 0.8171,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.17373530914665303,
|
| 244 |
+
"grad_norm": 0.7002721154510084,
|
| 245 |
+
"learning_rate": 5e-06,
|
| 246 |
+
"loss": 0.8073,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.17884517118037813,
|
| 251 |
+
"grad_norm": 0.7366019714677299,
|
| 252 |
+
"learning_rate": 5e-06,
|
| 253 |
+
"loss": 0.8139,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.18395503321410323,
|
| 258 |
+
"grad_norm": 0.4960107622038265,
|
| 259 |
+
"learning_rate": 5e-06,
|
| 260 |
+
"loss": 0.8195,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.1890648952478283,
|
| 265 |
+
"grad_norm": 0.590921742782821,
|
| 266 |
+
"learning_rate": 5e-06,
|
| 267 |
+
"loss": 0.8131,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.1941747572815534,
|
| 272 |
+
"grad_norm": 0.43046345355299165,
|
| 273 |
+
"learning_rate": 5e-06,
|
| 274 |
+
"loss": 0.8166,
|
| 275 |
+
"step": 380
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.1992846193152785,
|
| 279 |
+
"grad_norm": 0.37156340060616266,
|
| 280 |
+
"learning_rate": 5e-06,
|
| 281 |
+
"loss": 0.8082,
|
| 282 |
+
"step": 390
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.20439448134900357,
|
| 286 |
+
"grad_norm": 0.5029827349762587,
|
| 287 |
+
"learning_rate": 5e-06,
|
| 288 |
+
"loss": 0.8187,
|
| 289 |
+
"step": 400
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.20950434338272866,
|
| 293 |
+
"grad_norm": 0.39268177137347837,
|
| 294 |
+
"learning_rate": 5e-06,
|
| 295 |
+
"loss": 0.8132,
|
| 296 |
+
"step": 410
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.21461420541645376,
|
| 300 |
+
"grad_norm": 0.38787210509179787,
|
| 301 |
+
"learning_rate": 5e-06,
|
| 302 |
+
"loss": 0.8153,
|
| 303 |
+
"step": 420
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.21972406745017883,
|
| 307 |
+
"grad_norm": 0.5073537687615494,
|
| 308 |
+
"learning_rate": 5e-06,
|
| 309 |
+
"loss": 0.8032,
|
| 310 |
+
"step": 430
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.22483392948390393,
|
| 314 |
+
"grad_norm": 0.36232141966438186,
|
| 315 |
+
"learning_rate": 5e-06,
|
| 316 |
+
"loss": 0.786,
|
| 317 |
+
"step": 440
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.22994379151762903,
|
| 321 |
+
"grad_norm": 0.3362374048828632,
|
| 322 |
+
"learning_rate": 5e-06,
|
| 323 |
+
"loss": 0.8019,
|
| 324 |
+
"step": 450
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.2350536535513541,
|
| 328 |
+
"grad_norm": 0.3661481336901712,
|
| 329 |
+
"learning_rate": 5e-06,
|
| 330 |
+
"loss": 0.8012,
|
| 331 |
+
"step": 460
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.2401635155850792,
|
| 335 |
+
"grad_norm": 0.3186478374299407,
|
| 336 |
+
"learning_rate": 5e-06,
|
| 337 |
+
"loss": 0.8088,
|
| 338 |
+
"step": 470
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.2452733776188043,
|
| 342 |
+
"grad_norm": 0.3742412176358808,
|
| 343 |
+
"learning_rate": 5e-06,
|
| 344 |
+
"loss": 0.795,
|
| 345 |
+
"step": 480
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.2503832396525294,
|
| 349 |
+
"grad_norm": 0.3687134051282934,
|
| 350 |
+
"learning_rate": 5e-06,
|
| 351 |
+
"loss": 0.7899,
|
| 352 |
+
"step": 490
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.2554931016862545,
|
| 356 |
+
"grad_norm": 0.2844684124400157,
|
| 357 |
+
"learning_rate": 5e-06,
|
| 358 |
+
"loss": 0.7992,
|
| 359 |
+
"step": 500
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.26060296371997954,
|
| 363 |
+
"grad_norm": 0.3445705328013273,
|
| 364 |
+
"learning_rate": 5e-06,
|
| 365 |
+
"loss": 0.7908,
|
| 366 |
+
"step": 510
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.26571282575370464,
|
| 370 |
+
"grad_norm": 0.3100104564359116,
|
| 371 |
+
"learning_rate": 5e-06,
|
| 372 |
+
"loss": 0.7874,
|
| 373 |
+
"step": 520
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.27082268778742974,
|
| 377 |
+
"grad_norm": 0.32495110543783357,
|
| 378 |
+
"learning_rate": 5e-06,
|
| 379 |
+
"loss": 0.7982,
|
| 380 |
+
"step": 530
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.27593254982115484,
|
| 384 |
+
"grad_norm": 0.27997656185948167,
|
| 385 |
+
"learning_rate": 5e-06,
|
| 386 |
+
"loss": 0.7954,
|
| 387 |
+
"step": 540
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.28104241185487994,
|
| 391 |
+
"grad_norm": 0.4055280072220847,
|
| 392 |
+
"learning_rate": 5e-06,
|
| 393 |
+
"loss": 0.7895,
|
| 394 |
+
"step": 550
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.286152273888605,
|
| 398 |
+
"grad_norm": 0.2751499734062912,
|
| 399 |
+
"learning_rate": 5e-06,
|
| 400 |
+
"loss": 0.7931,
|
| 401 |
+
"step": 560
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.2912621359223301,
|
| 405 |
+
"grad_norm": 0.3232277546201772,
|
| 406 |
+
"learning_rate": 5e-06,
|
| 407 |
+
"loss": 0.7972,
|
| 408 |
+
"step": 570
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.2963719979560552,
|
| 412 |
+
"grad_norm": 0.2915040535303487,
|
| 413 |
+
"learning_rate": 5e-06,
|
| 414 |
+
"loss": 0.7912,
|
| 415 |
+
"step": 580
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.3014818599897803,
|
| 419 |
+
"grad_norm": 0.2637267252402153,
|
| 420 |
+
"learning_rate": 5e-06,
|
| 421 |
+
"loss": 0.7903,
|
| 422 |
+
"step": 590
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.3065917220235054,
|
| 426 |
+
"grad_norm": 0.329139944193737,
|
| 427 |
+
"learning_rate": 5e-06,
|
| 428 |
+
"loss": 0.7906,
|
| 429 |
+
"step": 600
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.3117015840572305,
|
| 433 |
+
"grad_norm": 0.24558162809212852,
|
| 434 |
+
"learning_rate": 5e-06,
|
| 435 |
+
"loss": 0.7947,
|
| 436 |
+
"step": 610
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.3168114460909555,
|
| 440 |
+
"grad_norm": 0.27224578177967845,
|
| 441 |
+
"learning_rate": 5e-06,
|
| 442 |
+
"loss": 0.7998,
|
| 443 |
+
"step": 620
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.3219213081246806,
|
| 447 |
+
"grad_norm": 0.2379116871443806,
|
| 448 |
+
"learning_rate": 5e-06,
|
| 449 |
+
"loss": 0.7702,
|
| 450 |
+
"step": 630
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.3270311701584057,
|
| 454 |
+
"grad_norm": 0.2609800304013704,
|
| 455 |
+
"learning_rate": 5e-06,
|
| 456 |
+
"loss": 0.7799,
|
| 457 |
+
"step": 640
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.3321410321921308,
|
| 461 |
+
"grad_norm": 0.23810454322283608,
|
| 462 |
+
"learning_rate": 5e-06,
|
| 463 |
+
"loss": 0.7897,
|
| 464 |
+
"step": 650
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.3372508942258559,
|
| 468 |
+
"grad_norm": 0.252487995051141,
|
| 469 |
+
"learning_rate": 5e-06,
|
| 470 |
+
"loss": 0.7726,
|
| 471 |
+
"step": 660
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.342360756259581,
|
| 475 |
+
"grad_norm": 0.24925719360242624,
|
| 476 |
+
"learning_rate": 5e-06,
|
| 477 |
+
"loss": 0.7978,
|
| 478 |
+
"step": 670
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.34747061829330605,
|
| 482 |
+
"grad_norm": 0.2656149332274455,
|
| 483 |
+
"learning_rate": 5e-06,
|
| 484 |
+
"loss": 0.7915,
|
| 485 |
+
"step": 680
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.35258048032703115,
|
| 489 |
+
"grad_norm": 0.2661464494899381,
|
| 490 |
+
"learning_rate": 5e-06,
|
| 491 |
+
"loss": 0.78,
|
| 492 |
+
"step": 690
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.35769034236075625,
|
| 496 |
+
"grad_norm": 0.20145837945729328,
|
| 497 |
+
"learning_rate": 5e-06,
|
| 498 |
+
"loss": 0.7734,
|
| 499 |
+
"step": 700
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.36280020439448135,
|
| 503 |
+
"grad_norm": 0.26148666494272177,
|
| 504 |
+
"learning_rate": 5e-06,
|
| 505 |
+
"loss": 0.7792,
|
| 506 |
+
"step": 710
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.36791006642820645,
|
| 510 |
+
"grad_norm": 0.24346683962909085,
|
| 511 |
+
"learning_rate": 5e-06,
|
| 512 |
+
"loss": 0.7816,
|
| 513 |
+
"step": 720
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.37301992846193155,
|
| 517 |
+
"grad_norm": 0.22278218589963927,
|
| 518 |
+
"learning_rate": 5e-06,
|
| 519 |
+
"loss": 0.7827,
|
| 520 |
+
"step": 730
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.3781297904956566,
|
| 524 |
+
"grad_norm": 0.2303051334399858,
|
| 525 |
+
"learning_rate": 5e-06,
|
| 526 |
+
"loss": 0.7707,
|
| 527 |
+
"step": 740
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.3832396525293817,
|
| 531 |
+
"grad_norm": 0.2543406962739334,
|
| 532 |
+
"learning_rate": 5e-06,
|
| 533 |
+
"loss": 0.7721,
|
| 534 |
+
"step": 750
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.3883495145631068,
|
| 538 |
+
"grad_norm": 0.22568609222745584,
|
| 539 |
+
"learning_rate": 5e-06,
|
| 540 |
+
"loss": 0.7836,
|
| 541 |
+
"step": 760
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.3934593765968319,
|
| 545 |
+
"grad_norm": 0.2294845661519681,
|
| 546 |
+
"learning_rate": 5e-06,
|
| 547 |
+
"loss": 0.7872,
|
| 548 |
+
"step": 770
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.398569238630557,
|
| 552 |
+
"grad_norm": 0.2197759747155408,
|
| 553 |
+
"learning_rate": 5e-06,
|
| 554 |
+
"loss": 0.7681,
|
| 555 |
+
"step": 780
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.4036791006642821,
|
| 559 |
+
"grad_norm": 0.22484907906751414,
|
| 560 |
+
"learning_rate": 5e-06,
|
| 561 |
+
"loss": 0.7797,
|
| 562 |
+
"step": 790
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.40878896269800713,
|
| 566 |
+
"grad_norm": 0.21381439198625019,
|
| 567 |
+
"learning_rate": 5e-06,
|
| 568 |
+
"loss": 0.7891,
|
| 569 |
+
"step": 800
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.41389882473173223,
|
| 573 |
+
"grad_norm": 0.21726381466920616,
|
| 574 |
+
"learning_rate": 5e-06,
|
| 575 |
+
"loss": 0.7779,
|
| 576 |
+
"step": 810
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.4190086867654573,
|
| 580 |
+
"grad_norm": 0.21218649602843365,
|
| 581 |
+
"learning_rate": 5e-06,
|
| 582 |
+
"loss": 0.7756,
|
| 583 |
+
"step": 820
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.4241185487991824,
|
| 587 |
+
"grad_norm": 0.2321703903701098,
|
| 588 |
+
"learning_rate": 5e-06,
|
| 589 |
+
"loss": 0.778,
|
| 590 |
+
"step": 830
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.4292284108329075,
|
| 594 |
+
"grad_norm": 0.21376686462742833,
|
| 595 |
+
"learning_rate": 5e-06,
|
| 596 |
+
"loss": 0.7754,
|
| 597 |
+
"step": 840
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.4343382728666326,
|
| 601 |
+
"grad_norm": 0.20773121342846562,
|
| 602 |
+
"learning_rate": 5e-06,
|
| 603 |
+
"loss": 0.7729,
|
| 604 |
+
"step": 850
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.43944813490035767,
|
| 608 |
+
"grad_norm": 0.23075209429146118,
|
| 609 |
+
"learning_rate": 5e-06,
|
| 610 |
+
"loss": 0.7743,
|
| 611 |
+
"step": 860
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.44455799693408277,
|
| 615 |
+
"grad_norm": 0.21889813942522324,
|
| 616 |
+
"learning_rate": 5e-06,
|
| 617 |
+
"loss": 0.7872,
|
| 618 |
+
"step": 870
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.44966785896780787,
|
| 622 |
+
"grad_norm": 0.23158111213452945,
|
| 623 |
+
"learning_rate": 5e-06,
|
| 624 |
+
"loss": 0.7729,
|
| 625 |
+
"step": 880
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.45477772100153296,
|
| 629 |
+
"grad_norm": 0.23759359366214816,
|
| 630 |
+
"learning_rate": 5e-06,
|
| 631 |
+
"loss": 0.7781,
|
| 632 |
+
"step": 890
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.45988758303525806,
|
| 636 |
+
"grad_norm": 0.20737512188117463,
|
| 637 |
+
"learning_rate": 5e-06,
|
| 638 |
+
"loss": 0.7733,
|
| 639 |
+
"step": 900
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.46499744506898316,
|
| 643 |
+
"grad_norm": 0.21978896076440746,
|
| 644 |
+
"learning_rate": 5e-06,
|
| 645 |
+
"loss": 0.7724,
|
| 646 |
+
"step": 910
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.4701073071027082,
|
| 650 |
+
"grad_norm": 0.21326048741734183,
|
| 651 |
+
"learning_rate": 5e-06,
|
| 652 |
+
"loss": 0.7779,
|
| 653 |
+
"step": 920
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.4752171691364333,
|
| 657 |
+
"grad_norm": 0.2127733995734116,
|
| 658 |
+
"learning_rate": 5e-06,
|
| 659 |
+
"loss": 0.7862,
|
| 660 |
+
"step": 930
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.4803270311701584,
|
| 664 |
+
"grad_norm": 0.20545254783011138,
|
| 665 |
+
"learning_rate": 5e-06,
|
| 666 |
+
"loss": 0.7741,
|
| 667 |
+
"step": 940
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.4854368932038835,
|
| 671 |
+
"grad_norm": 0.22401938020775877,
|
| 672 |
+
"learning_rate": 5e-06,
|
| 673 |
+
"loss": 0.7879,
|
| 674 |
+
"step": 950
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.4905467552376086,
|
| 678 |
+
"grad_norm": 0.20313108726519555,
|
| 679 |
+
"learning_rate": 5e-06,
|
| 680 |
+
"loss": 0.7795,
|
| 681 |
+
"step": 960
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.4956566172713337,
|
| 685 |
+
"grad_norm": 0.20676236489485517,
|
| 686 |
+
"learning_rate": 5e-06,
|
| 687 |
+
"loss": 0.7658,
|
| 688 |
+
"step": 970
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.5007664793050588,
|
| 692 |
+
"grad_norm": 0.1997349100496237,
|
| 693 |
+
"learning_rate": 5e-06,
|
| 694 |
+
"loss": 0.7665,
|
| 695 |
+
"step": 980
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.5058763413387839,
|
| 699 |
+
"grad_norm": 0.18582330695996696,
|
| 700 |
+
"learning_rate": 5e-06,
|
| 701 |
+
"loss": 0.7678,
|
| 702 |
+
"step": 990
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.510986203372509,
|
| 706 |
+
"grad_norm": 0.20480578385489007,
|
| 707 |
+
"learning_rate": 5e-06,
|
| 708 |
+
"loss": 0.7916,
|
| 709 |
+
"step": 1000
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.516096065406234,
|
| 713 |
+
"grad_norm": 0.2068595244192038,
|
| 714 |
+
"learning_rate": 5e-06,
|
| 715 |
+
"loss": 0.7728,
|
| 716 |
+
"step": 1010
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.5212059274399591,
|
| 720 |
+
"grad_norm": 0.19202510728167327,
|
| 721 |
+
"learning_rate": 5e-06,
|
| 722 |
+
"loss": 0.7875,
|
| 723 |
+
"step": 1020
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.5263157894736842,
|
| 727 |
+
"grad_norm": 0.20142193391262955,
|
| 728 |
+
"learning_rate": 5e-06,
|
| 729 |
+
"loss": 0.7691,
|
| 730 |
+
"step": 1030
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.5314256515074093,
|
| 734 |
+
"grad_norm": 0.19153357968487428,
|
| 735 |
+
"learning_rate": 5e-06,
|
| 736 |
+
"loss": 0.7717,
|
| 737 |
+
"step": 1040
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.5365355135411344,
|
| 741 |
+
"grad_norm": 0.21070221575444148,
|
| 742 |
+
"learning_rate": 5e-06,
|
| 743 |
+
"loss": 0.771,
|
| 744 |
+
"step": 1050
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.5416453755748595,
|
| 748 |
+
"grad_norm": 0.19388651883167224,
|
| 749 |
+
"learning_rate": 5e-06,
|
| 750 |
+
"loss": 0.7674,
|
| 751 |
+
"step": 1060
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.5467552376085846,
|
| 755 |
+
"grad_norm": 0.19235966239305358,
|
| 756 |
+
"learning_rate": 5e-06,
|
| 757 |
+
"loss": 0.7639,
|
| 758 |
+
"step": 1070
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.5518650996423097,
|
| 762 |
+
"grad_norm": 0.2020128053771939,
|
| 763 |
+
"learning_rate": 5e-06,
|
| 764 |
+
"loss": 0.7838,
|
| 765 |
+
"step": 1080
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.5569749616760348,
|
| 769 |
+
"grad_norm": 0.19898211373916952,
|
| 770 |
+
"learning_rate": 5e-06,
|
| 771 |
+
"loss": 0.7734,
|
| 772 |
+
"step": 1090
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.5620848237097599,
|
| 776 |
+
"grad_norm": 0.19556108367687597,
|
| 777 |
+
"learning_rate": 5e-06,
|
| 778 |
+
"loss": 0.7668,
|
| 779 |
+
"step": 1100
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.567194685743485,
|
| 783 |
+
"grad_norm": 0.19409586285487196,
|
| 784 |
+
"learning_rate": 5e-06,
|
| 785 |
+
"loss": 0.7778,
|
| 786 |
+
"step": 1110
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.57230454777721,
|
| 790 |
+
"grad_norm": 0.20681598650606026,
|
| 791 |
+
"learning_rate": 5e-06,
|
| 792 |
+
"loss": 0.779,
|
| 793 |
+
"step": 1120
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.5774144098109351,
|
| 797 |
+
"grad_norm": 0.2017546177375358,
|
| 798 |
+
"learning_rate": 5e-06,
|
| 799 |
+
"loss": 0.7573,
|
| 800 |
+
"step": 1130
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.5825242718446602,
|
| 804 |
+
"grad_norm": 0.1913755514728684,
|
| 805 |
+
"learning_rate": 5e-06,
|
| 806 |
+
"loss": 0.7625,
|
| 807 |
+
"step": 1140
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.5876341338783853,
|
| 811 |
+
"grad_norm": 0.18303523769215124,
|
| 812 |
+
"learning_rate": 5e-06,
|
| 813 |
+
"loss": 0.7752,
|
| 814 |
+
"step": 1150
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.5927439959121104,
|
| 818 |
+
"grad_norm": 0.19646699079126273,
|
| 819 |
+
"learning_rate": 5e-06,
|
| 820 |
+
"loss": 0.7734,
|
| 821 |
+
"step": 1160
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.5978538579458355,
|
| 825 |
+
"grad_norm": 0.18589450588787854,
|
| 826 |
+
"learning_rate": 5e-06,
|
| 827 |
+
"loss": 0.7631,
|
| 828 |
+
"step": 1170
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.6029637199795606,
|
| 832 |
+
"grad_norm": 0.19141416094028266,
|
| 833 |
+
"learning_rate": 5e-06,
|
| 834 |
+
"loss": 0.7721,
|
| 835 |
+
"step": 1180
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.6080735820132857,
|
| 839 |
+
"grad_norm": 0.2010700371236207,
|
| 840 |
+
"learning_rate": 5e-06,
|
| 841 |
+
"loss": 0.7723,
|
| 842 |
+
"step": 1190
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.6131834440470108,
|
| 846 |
+
"grad_norm": 0.209693144160535,
|
| 847 |
+
"learning_rate": 5e-06,
|
| 848 |
+
"loss": 0.7709,
|
| 849 |
+
"step": 1200
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.6182933060807358,
|
| 853 |
+
"grad_norm": 0.19006264249604685,
|
| 854 |
+
"learning_rate": 5e-06,
|
| 855 |
+
"loss": 0.7731,
|
| 856 |
+
"step": 1210
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.623403168114461,
|
| 860 |
+
"grad_norm": 0.19625031587081004,
|
| 861 |
+
"learning_rate": 5e-06,
|
| 862 |
+
"loss": 0.7683,
|
| 863 |
+
"step": 1220
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.628513030148186,
|
| 867 |
+
"grad_norm": 0.19078272960681916,
|
| 868 |
+
"learning_rate": 5e-06,
|
| 869 |
+
"loss": 0.773,
|
| 870 |
+
"step": 1230
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.633622892181911,
|
| 874 |
+
"grad_norm": 0.2007688157803181,
|
| 875 |
+
"learning_rate": 5e-06,
|
| 876 |
+
"loss": 0.7533,
|
| 877 |
+
"step": 1240
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.6387327542156361,
|
| 881 |
+
"grad_norm": 0.20004504142951482,
|
| 882 |
+
"learning_rate": 5e-06,
|
| 883 |
+
"loss": 0.7721,
|
| 884 |
+
"step": 1250
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.6438426162493612,
|
| 888 |
+
"grad_norm": 0.18940428161938747,
|
| 889 |
+
"learning_rate": 5e-06,
|
| 890 |
+
"loss": 0.7525,
|
| 891 |
+
"step": 1260
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.6489524782830863,
|
| 895 |
+
"grad_norm": 0.19192245014604287,
|
| 896 |
+
"learning_rate": 5e-06,
|
| 897 |
+
"loss": 0.7743,
|
| 898 |
+
"step": 1270
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.6540623403168114,
|
| 902 |
+
"grad_norm": 0.18803291404020508,
|
| 903 |
+
"learning_rate": 5e-06,
|
| 904 |
+
"loss": 0.7781,
|
| 905 |
+
"step": 1280
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.6591722023505365,
|
| 909 |
+
"grad_norm": 0.19034763073417169,
|
| 910 |
+
"learning_rate": 5e-06,
|
| 911 |
+
"loss": 0.7609,
|
| 912 |
+
"step": 1290
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.6642820643842616,
|
| 916 |
+
"grad_norm": 0.2000260745898994,
|
| 917 |
+
"learning_rate": 5e-06,
|
| 918 |
+
"loss": 0.7493,
|
| 919 |
+
"step": 1300
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.6693919264179867,
|
| 923 |
+
"grad_norm": 0.19721212735320226,
|
| 924 |
+
"learning_rate": 5e-06,
|
| 925 |
+
"loss": 0.759,
|
| 926 |
+
"step": 1310
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.6745017884517118,
|
| 930 |
+
"grad_norm": 0.1840594216152482,
|
| 931 |
+
"learning_rate": 5e-06,
|
| 932 |
+
"loss": 0.7554,
|
| 933 |
+
"step": 1320
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.6796116504854369,
|
| 937 |
+
"grad_norm": 0.19294533218500587,
|
| 938 |
+
"learning_rate": 5e-06,
|
| 939 |
+
"loss": 0.7601,
|
| 940 |
+
"step": 1330
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.684721512519162,
|
| 944 |
+
"grad_norm": 0.2003437533474394,
|
| 945 |
+
"learning_rate": 5e-06,
|
| 946 |
+
"loss": 0.7665,
|
| 947 |
+
"step": 1340
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.6898313745528871,
|
| 951 |
+
"grad_norm": 0.19522682692250634,
|
| 952 |
+
"learning_rate": 5e-06,
|
| 953 |
+
"loss": 0.7682,
|
| 954 |
+
"step": 1350
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.6949412365866121,
|
| 958 |
+
"grad_norm": 0.19114649796975278,
|
| 959 |
+
"learning_rate": 5e-06,
|
| 960 |
+
"loss": 0.7675,
|
| 961 |
+
"step": 1360
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.7000510986203372,
|
| 965 |
+
"grad_norm": 0.18795490185304542,
|
| 966 |
+
"learning_rate": 5e-06,
|
| 967 |
+
"loss": 0.7514,
|
| 968 |
+
"step": 1370
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.7051609606540623,
|
| 972 |
+
"grad_norm": 0.19982770311498202,
|
| 973 |
+
"learning_rate": 5e-06,
|
| 974 |
+
"loss": 0.7587,
|
| 975 |
+
"step": 1380
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.7102708226877874,
|
| 979 |
+
"grad_norm": 0.2094868748722276,
|
| 980 |
+
"learning_rate": 5e-06,
|
| 981 |
+
"loss": 0.7424,
|
| 982 |
+
"step": 1390
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.7153806847215125,
|
| 986 |
+
"grad_norm": 0.18141939898186682,
|
| 987 |
+
"learning_rate": 5e-06,
|
| 988 |
+
"loss": 0.7668,
|
| 989 |
+
"step": 1400
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.7204905467552376,
|
| 993 |
+
"grad_norm": 0.184317543793234,
|
| 994 |
+
"learning_rate": 5e-06,
|
| 995 |
+
"loss": 0.7659,
|
| 996 |
+
"step": 1410
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.7256004087889627,
|
| 1000 |
+
"grad_norm": 0.19482359157799217,
|
| 1001 |
+
"learning_rate": 5e-06,
|
| 1002 |
+
"loss": 0.766,
|
| 1003 |
+
"step": 1420
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.7307102708226878,
|
| 1007 |
+
"grad_norm": 0.20609416559465576,
|
| 1008 |
+
"learning_rate": 5e-06,
|
| 1009 |
+
"loss": 0.7591,
|
| 1010 |
+
"step": 1430
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.7358201328564129,
|
| 1014 |
+
"grad_norm": 0.1733817918744796,
|
| 1015 |
+
"learning_rate": 5e-06,
|
| 1016 |
+
"loss": 0.7657,
|
| 1017 |
+
"step": 1440
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.740929994890138,
|
| 1021 |
+
"grad_norm": 0.20231819059208814,
|
| 1022 |
+
"learning_rate": 5e-06,
|
| 1023 |
+
"loss": 0.7693,
|
| 1024 |
+
"step": 1450
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.7460398569238631,
|
| 1028 |
+
"grad_norm": 0.19384075901742115,
|
| 1029 |
+
"learning_rate": 5e-06,
|
| 1030 |
+
"loss": 0.7677,
|
| 1031 |
+
"step": 1460
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.7511497189575882,
|
| 1035 |
+
"grad_norm": 0.20242534213073207,
|
| 1036 |
+
"learning_rate": 5e-06,
|
| 1037 |
+
"loss": 0.7658,
|
| 1038 |
+
"step": 1470
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.7562595809913132,
|
| 1042 |
+
"grad_norm": 0.18992152096280124,
|
| 1043 |
+
"learning_rate": 5e-06,
|
| 1044 |
+
"loss": 0.7604,
|
| 1045 |
+
"step": 1480
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.7613694430250383,
|
| 1049 |
+
"grad_norm": 0.20300312286644698,
|
| 1050 |
+
"learning_rate": 5e-06,
|
| 1051 |
+
"loss": 0.7654,
|
| 1052 |
+
"step": 1490
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.7664793050587634,
|
| 1056 |
+
"grad_norm": 0.2110214114358105,
|
| 1057 |
+
"learning_rate": 5e-06,
|
| 1058 |
+
"loss": 0.7579,
|
| 1059 |
+
"step": 1500
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.7715891670924885,
|
| 1063 |
+
"grad_norm": 0.18507470375993035,
|
| 1064 |
+
"learning_rate": 5e-06,
|
| 1065 |
+
"loss": 0.7535,
|
| 1066 |
+
"step": 1510
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.7766990291262136,
|
| 1070 |
+
"grad_norm": 0.20773697020515,
|
| 1071 |
+
"learning_rate": 5e-06,
|
| 1072 |
+
"loss": 0.7681,
|
| 1073 |
+
"step": 1520
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.7818088911599387,
|
| 1077 |
+
"grad_norm": 0.1846942481775099,
|
| 1078 |
+
"learning_rate": 5e-06,
|
| 1079 |
+
"loss": 0.7589,
|
| 1080 |
+
"step": 1530
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.7869187531936638,
|
| 1084 |
+
"grad_norm": 0.19873266556747132,
|
| 1085 |
+
"learning_rate": 5e-06,
|
| 1086 |
+
"loss": 0.7701,
|
| 1087 |
+
"step": 1540
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.7920286152273889,
|
| 1091 |
+
"grad_norm": 0.1889722343751879,
|
| 1092 |
+
"learning_rate": 5e-06,
|
| 1093 |
+
"loss": 0.7637,
|
| 1094 |
+
"step": 1550
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.797138477261114,
|
| 1098 |
+
"grad_norm": 0.2160555515101488,
|
| 1099 |
+
"learning_rate": 5e-06,
|
| 1100 |
+
"loss": 0.7459,
|
| 1101 |
+
"step": 1560
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.8022483392948391,
|
| 1105 |
+
"grad_norm": 0.18958425843094595,
|
| 1106 |
+
"learning_rate": 5e-06,
|
| 1107 |
+
"loss": 0.7499,
|
| 1108 |
+
"step": 1570
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.8073582013285642,
|
| 1112 |
+
"grad_norm": 0.18917128647246217,
|
| 1113 |
+
"learning_rate": 5e-06,
|
| 1114 |
+
"loss": 0.7572,
|
| 1115 |
+
"step": 1580
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.8124680633622893,
|
| 1119 |
+
"grad_norm": 0.17914225162735836,
|
| 1120 |
+
"learning_rate": 5e-06,
|
| 1121 |
+
"loss": 0.7506,
|
| 1122 |
+
"step": 1590
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.8175779253960143,
|
| 1126 |
+
"grad_norm": 0.19189511614416635,
|
| 1127 |
+
"learning_rate": 5e-06,
|
| 1128 |
+
"loss": 0.7613,
|
| 1129 |
+
"step": 1600
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.8226877874297394,
|
| 1133 |
+
"grad_norm": 0.19059344363394998,
|
| 1134 |
+
"learning_rate": 5e-06,
|
| 1135 |
+
"loss": 0.7604,
|
| 1136 |
+
"step": 1610
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.8277976494634645,
|
| 1140 |
+
"grad_norm": 0.20852250489781288,
|
| 1141 |
+
"learning_rate": 5e-06,
|
| 1142 |
+
"loss": 0.7737,
|
| 1143 |
+
"step": 1620
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.8329075114971896,
|
| 1147 |
+
"grad_norm": 0.1877964215413997,
|
| 1148 |
+
"learning_rate": 5e-06,
|
| 1149 |
+
"loss": 0.7725,
|
| 1150 |
+
"step": 1630
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.8380173735309147,
|
| 1154 |
+
"grad_norm": 0.18419774730049385,
|
| 1155 |
+
"learning_rate": 5e-06,
|
| 1156 |
+
"loss": 0.757,
|
| 1157 |
+
"step": 1640
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.8431272355646398,
|
| 1161 |
+
"grad_norm": 0.1926784804428757,
|
| 1162 |
+
"learning_rate": 5e-06,
|
| 1163 |
+
"loss": 0.7515,
|
| 1164 |
+
"step": 1650
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.8482370975983649,
|
| 1168 |
+
"grad_norm": 0.19627214391258455,
|
| 1169 |
+
"learning_rate": 5e-06,
|
| 1170 |
+
"loss": 0.7718,
|
| 1171 |
+
"step": 1660
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.85334695963209,
|
| 1175 |
+
"grad_norm": 0.1944045450874202,
|
| 1176 |
+
"learning_rate": 5e-06,
|
| 1177 |
+
"loss": 0.7449,
|
| 1178 |
+
"step": 1670
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.858456821665815,
|
| 1182 |
+
"grad_norm": 0.18125259142482736,
|
| 1183 |
+
"learning_rate": 5e-06,
|
| 1184 |
+
"loss": 0.7604,
|
| 1185 |
+
"step": 1680
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.8635666836995401,
|
| 1189 |
+
"grad_norm": 0.17817637475202877,
|
| 1190 |
+
"learning_rate": 5e-06,
|
| 1191 |
+
"loss": 0.7641,
|
| 1192 |
+
"step": 1690
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.8686765457332652,
|
| 1196 |
+
"grad_norm": 0.20471658407314103,
|
| 1197 |
+
"learning_rate": 5e-06,
|
| 1198 |
+
"loss": 0.7734,
|
| 1199 |
+
"step": 1700
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.8737864077669902,
|
| 1203 |
+
"grad_norm": 0.1878161208418682,
|
| 1204 |
+
"learning_rate": 5e-06,
|
| 1205 |
+
"loss": 0.7696,
|
| 1206 |
+
"step": 1710
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.8788962698007153,
|
| 1210 |
+
"grad_norm": 0.18334210713527221,
|
| 1211 |
+
"learning_rate": 5e-06,
|
| 1212 |
+
"loss": 0.7536,
|
| 1213 |
+
"step": 1720
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.8840061318344404,
|
| 1217 |
+
"grad_norm": 0.18076699199021762,
|
| 1218 |
+
"learning_rate": 5e-06,
|
| 1219 |
+
"loss": 0.7522,
|
| 1220 |
+
"step": 1730
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.8891159938681655,
|
| 1224 |
+
"grad_norm": 0.19181861440876702,
|
| 1225 |
+
"learning_rate": 5e-06,
|
| 1226 |
+
"loss": 0.7625,
|
| 1227 |
+
"step": 1740
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.8942258559018906,
|
| 1231 |
+
"grad_norm": 0.17965104525055625,
|
| 1232 |
+
"learning_rate": 5e-06,
|
| 1233 |
+
"loss": 0.754,
|
| 1234 |
+
"step": 1750
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 0.8993357179356157,
|
| 1238 |
+
"grad_norm": 0.19942739585455946,
|
| 1239 |
+
"learning_rate": 5e-06,
|
| 1240 |
+
"loss": 0.7663,
|
| 1241 |
+
"step": 1760
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.9044455799693408,
|
| 1245 |
+
"grad_norm": 0.19441177912329213,
|
| 1246 |
+
"learning_rate": 5e-06,
|
| 1247 |
+
"loss": 0.7643,
|
| 1248 |
+
"step": 1770
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.9095554420030659,
|
| 1252 |
+
"grad_norm": 0.1989153914067934,
|
| 1253 |
+
"learning_rate": 5e-06,
|
| 1254 |
+
"loss": 0.7538,
|
| 1255 |
+
"step": 1780
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.914665304036791,
|
| 1259 |
+
"grad_norm": 0.18779250602110079,
|
| 1260 |
+
"learning_rate": 5e-06,
|
| 1261 |
+
"loss": 0.7582,
|
| 1262 |
+
"step": 1790
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.9197751660705161,
|
| 1266 |
+
"grad_norm": 0.18880847142963628,
|
| 1267 |
+
"learning_rate": 5e-06,
|
| 1268 |
+
"loss": 0.7834,
|
| 1269 |
+
"step": 1800
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.9248850281042412,
|
| 1273 |
+
"grad_norm": 0.19100671149679851,
|
| 1274 |
+
"learning_rate": 5e-06,
|
| 1275 |
+
"loss": 0.7633,
|
| 1276 |
+
"step": 1810
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.9299948901379663,
|
| 1280 |
+
"grad_norm": 0.19156019315031683,
|
| 1281 |
+
"learning_rate": 5e-06,
|
| 1282 |
+
"loss": 0.7644,
|
| 1283 |
+
"step": 1820
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.9351047521716913,
|
| 1287 |
+
"grad_norm": 0.18435442552610007,
|
| 1288 |
+
"learning_rate": 5e-06,
|
| 1289 |
+
"loss": 0.7597,
|
| 1290 |
+
"step": 1830
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 0.9402146142054164,
|
| 1294 |
+
"grad_norm": 0.2051724242213117,
|
| 1295 |
+
"learning_rate": 5e-06,
|
| 1296 |
+
"loss": 0.7504,
|
| 1297 |
+
"step": 1840
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.9453244762391415,
|
| 1301 |
+
"grad_norm": 0.18536587248191086,
|
| 1302 |
+
"learning_rate": 5e-06,
|
| 1303 |
+
"loss": 0.7396,
|
| 1304 |
+
"step": 1850
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.9504343382728666,
|
| 1308 |
+
"grad_norm": 0.17780024213447235,
|
| 1309 |
+
"learning_rate": 5e-06,
|
| 1310 |
+
"loss": 0.7506,
|
| 1311 |
+
"step": 1860
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.9555442003065917,
|
| 1315 |
+
"grad_norm": 0.18708967218779626,
|
| 1316 |
+
"learning_rate": 5e-06,
|
| 1317 |
+
"loss": 0.7455,
|
| 1318 |
+
"step": 1870
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 0.9606540623403168,
|
| 1322 |
+
"grad_norm": 0.17747103600840475,
|
| 1323 |
+
"learning_rate": 5e-06,
|
| 1324 |
+
"loss": 0.7444,
|
| 1325 |
+
"step": 1880
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.9657639243740419,
|
| 1329 |
+
"grad_norm": 0.19063629309146018,
|
| 1330 |
+
"learning_rate": 5e-06,
|
| 1331 |
+
"loss": 0.7533,
|
| 1332 |
+
"step": 1890
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 0.970873786407767,
|
| 1336 |
+
"grad_norm": 0.19471871824403422,
|
| 1337 |
+
"learning_rate": 5e-06,
|
| 1338 |
+
"loss": 0.7539,
|
| 1339 |
+
"step": 1900
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.9759836484414921,
|
| 1343 |
+
"grad_norm": 0.18403984668742995,
|
| 1344 |
+
"learning_rate": 5e-06,
|
| 1345 |
+
"loss": 0.766,
|
| 1346 |
+
"step": 1910
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.9810935104752172,
|
| 1350 |
+
"grad_norm": 0.19270062603489418,
|
| 1351 |
+
"learning_rate": 5e-06,
|
| 1352 |
+
"loss": 0.7661,
|
| 1353 |
+
"step": 1920
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.9862033725089423,
|
| 1357 |
+
"grad_norm": 0.19463685028697894,
|
| 1358 |
+
"learning_rate": 5e-06,
|
| 1359 |
+
"loss": 0.7591,
|
| 1360 |
+
"step": 1930
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 0.9913132345426674,
|
| 1364 |
+
"grad_norm": 0.18870267371498323,
|
| 1365 |
+
"learning_rate": 5e-06,
|
| 1366 |
+
"loss": 0.7619,
|
| 1367 |
+
"step": 1940
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.9964230965763924,
|
| 1371 |
+
"grad_norm": 0.19561509144751293,
|
| 1372 |
+
"learning_rate": 5e-06,
|
| 1373 |
+
"loss": 0.7727,
|
| 1374 |
+
"step": 1950
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 1.0015329586101176,
|
| 1378 |
+
"grad_norm": 0.1865614045173204,
|
| 1379 |
+
"learning_rate": 5e-06,
|
| 1380 |
+
"loss": 0.753,
|
| 1381 |
+
"step": 1960
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 1.0066428206438427,
|
| 1385 |
+
"grad_norm": 0.17671337095527262,
|
| 1386 |
+
"learning_rate": 5e-06,
|
| 1387 |
+
"loss": 0.7557,
|
| 1388 |
+
"step": 1970
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 1.0117526826775678,
|
| 1392 |
+
"grad_norm": 0.19010483409505236,
|
| 1393 |
+
"learning_rate": 5e-06,
|
| 1394 |
+
"loss": 0.7446,
|
| 1395 |
+
"step": 1980
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 1.016862544711293,
|
| 1399 |
+
"grad_norm": 0.17413564167981435,
|
| 1400 |
+
"learning_rate": 5e-06,
|
| 1401 |
+
"loss": 0.7389,
|
| 1402 |
+
"step": 1990
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 1.021972406745018,
|
| 1406 |
+
"grad_norm": 0.18070567046481728,
|
| 1407 |
+
"learning_rate": 5e-06,
|
| 1408 |
+
"loss": 0.7451,
|
| 1409 |
+
"step": 2000
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 1.0270822687787429,
|
| 1413 |
+
"grad_norm": 0.2047148011452083,
|
| 1414 |
+
"learning_rate": 5e-06,
|
| 1415 |
+
"loss": 0.7404,
|
| 1416 |
+
"step": 2010
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 1.032192130812468,
|
| 1420 |
+
"grad_norm": 0.19369702939207809,
|
| 1421 |
+
"learning_rate": 5e-06,
|
| 1422 |
+
"loss": 0.7432,
|
| 1423 |
+
"step": 2020
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 1.037301992846193,
|
| 1427 |
+
"grad_norm": 0.19067195646001925,
|
| 1428 |
+
"learning_rate": 5e-06,
|
| 1429 |
+
"loss": 0.7396,
|
| 1430 |
+
"step": 2030
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 1.0424118548799182,
|
| 1434 |
+
"grad_norm": 0.20517054385871566,
|
| 1435 |
+
"learning_rate": 5e-06,
|
| 1436 |
+
"loss": 0.7312,
|
| 1437 |
+
"step": 2040
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 1.0475217169136433,
|
| 1441 |
+
"grad_norm": 0.19412259781936536,
|
| 1442 |
+
"learning_rate": 5e-06,
|
| 1443 |
+
"loss": 0.7434,
|
| 1444 |
+
"step": 2050
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 1.0526315789473684,
|
| 1448 |
+
"grad_norm": 0.1873188104877111,
|
| 1449 |
+
"learning_rate": 5e-06,
|
| 1450 |
+
"loss": 0.7411,
|
| 1451 |
+
"step": 2060
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 1.0577414409810935,
|
| 1455 |
+
"grad_norm": 0.1897369700854769,
|
| 1456 |
+
"learning_rate": 5e-06,
|
| 1457 |
+
"loss": 0.7345,
|
| 1458 |
+
"step": 2070
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 1.0628513030148186,
|
| 1462 |
+
"grad_norm": 0.1853976118509793,
|
| 1463 |
+
"learning_rate": 5e-06,
|
| 1464 |
+
"loss": 0.7549,
|
| 1465 |
+
"step": 2080
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 1.0679611650485437,
|
| 1469 |
+
"grad_norm": 0.17933568978007214,
|
| 1470 |
+
"learning_rate": 5e-06,
|
| 1471 |
+
"loss": 0.7531,
|
| 1472 |
+
"step": 2090
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 1.0730710270822688,
|
| 1476 |
+
"grad_norm": 0.20312467005419368,
|
| 1477 |
+
"learning_rate": 5e-06,
|
| 1478 |
+
"loss": 0.7498,
|
| 1479 |
+
"step": 2100
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 1.0781808891159939,
|
| 1483 |
+
"grad_norm": 0.18428135458855252,
|
| 1484 |
+
"learning_rate": 5e-06,
|
| 1485 |
+
"loss": 0.7404,
|
| 1486 |
+
"step": 2110
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 1.083290751149719,
|
| 1490 |
+
"grad_norm": 0.1973755972738785,
|
| 1491 |
+
"learning_rate": 5e-06,
|
| 1492 |
+
"loss": 0.7329,
|
| 1493 |
+
"step": 2120
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 1.088400613183444,
|
| 1497 |
+
"grad_norm": 0.1917623224859124,
|
| 1498 |
+
"learning_rate": 5e-06,
|
| 1499 |
+
"loss": 0.7555,
|
| 1500 |
+
"step": 2130
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 1.0935104752171692,
|
| 1504 |
+
"grad_norm": 0.18406588696688597,
|
| 1505 |
+
"learning_rate": 5e-06,
|
| 1506 |
+
"loss": 0.7433,
|
| 1507 |
+
"step": 2140
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 1.0986203372508943,
|
| 1511 |
+
"grad_norm": 0.17921040061727433,
|
| 1512 |
+
"learning_rate": 5e-06,
|
| 1513 |
+
"loss": 0.7305,
|
| 1514 |
+
"step": 2150
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 1.1037301992846194,
|
| 1518 |
+
"grad_norm": 0.18963146030246644,
|
| 1519 |
+
"learning_rate": 5e-06,
|
| 1520 |
+
"loss": 0.7397,
|
| 1521 |
+
"step": 2160
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 1.1088400613183444,
|
| 1525 |
+
"grad_norm": 0.18712686418913257,
|
| 1526 |
+
"learning_rate": 5e-06,
|
| 1527 |
+
"loss": 0.7409,
|
| 1528 |
+
"step": 2170
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 1.1139499233520695,
|
| 1532 |
+
"grad_norm": 0.19771359798461643,
|
| 1533 |
+
"learning_rate": 5e-06,
|
| 1534 |
+
"loss": 0.7427,
|
| 1535 |
+
"step": 2180
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 1.1190597853857946,
|
| 1539 |
+
"grad_norm": 0.1840406013875161,
|
| 1540 |
+
"learning_rate": 5e-06,
|
| 1541 |
+
"loss": 0.7611,
|
| 1542 |
+
"step": 2190
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 1.1241696474195197,
|
| 1546 |
+
"grad_norm": 0.18827676786628125,
|
| 1547 |
+
"learning_rate": 5e-06,
|
| 1548 |
+
"loss": 0.7364,
|
| 1549 |
+
"step": 2200
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 1.1292795094532448,
|
| 1553 |
+
"grad_norm": 0.191153055094231,
|
| 1554 |
+
"learning_rate": 5e-06,
|
| 1555 |
+
"loss": 0.7539,
|
| 1556 |
+
"step": 2210
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 1.13438937148697,
|
| 1560 |
+
"grad_norm": 0.1822666967889064,
|
| 1561 |
+
"learning_rate": 5e-06,
|
| 1562 |
+
"loss": 0.7414,
|
| 1563 |
+
"step": 2220
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 1.139499233520695,
|
| 1567 |
+
"grad_norm": 0.19699245088881503,
|
| 1568 |
+
"learning_rate": 5e-06,
|
| 1569 |
+
"loss": 0.738,
|
| 1570 |
+
"step": 2230
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 1.14460909555442,
|
| 1574 |
+
"grad_norm": 0.1898722587971926,
|
| 1575 |
+
"learning_rate": 5e-06,
|
| 1576 |
+
"loss": 0.7461,
|
| 1577 |
+
"step": 2240
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 1.1497189575881452,
|
| 1581 |
+
"grad_norm": 0.18927249674651098,
|
| 1582 |
+
"learning_rate": 5e-06,
|
| 1583 |
+
"loss": 0.7429,
|
| 1584 |
+
"step": 2250
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 1.1548288196218701,
|
| 1588 |
+
"grad_norm": 0.18896514284627008,
|
| 1589 |
+
"learning_rate": 5e-06,
|
| 1590 |
+
"loss": 0.7549,
|
| 1591 |
+
"step": 2260
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 1.1599386816555952,
|
| 1595 |
+
"grad_norm": 0.18739970646008372,
|
| 1596 |
+
"learning_rate": 5e-06,
|
| 1597 |
+
"loss": 0.7499,
|
| 1598 |
+
"step": 2270
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 1.1650485436893203,
|
| 1602 |
+
"grad_norm": 0.1910100009129843,
|
| 1603 |
+
"learning_rate": 5e-06,
|
| 1604 |
+
"loss": 0.747,
|
| 1605 |
+
"step": 2280
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 1.1701584057230454,
|
| 1609 |
+
"grad_norm": 0.20198153170551428,
|
| 1610 |
+
"learning_rate": 5e-06,
|
| 1611 |
+
"loss": 0.7386,
|
| 1612 |
+
"step": 2290
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 1.1752682677567705,
|
| 1616 |
+
"grad_norm": 0.18720641288978465,
|
| 1617 |
+
"learning_rate": 5e-06,
|
| 1618 |
+
"loss": 0.7578,
|
| 1619 |
+
"step": 2300
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 1.1803781297904956,
|
| 1623 |
+
"grad_norm": 0.18961987449195758,
|
| 1624 |
+
"learning_rate": 5e-06,
|
| 1625 |
+
"loss": 0.7529,
|
| 1626 |
+
"step": 2310
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 1.1854879918242207,
|
| 1630 |
+
"grad_norm": 0.17712198248177036,
|
| 1631 |
+
"learning_rate": 5e-06,
|
| 1632 |
+
"loss": 0.7476,
|
| 1633 |
+
"step": 2320
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 1.1905978538579458,
|
| 1637 |
+
"grad_norm": 0.18490722732969878,
|
| 1638 |
+
"learning_rate": 5e-06,
|
| 1639 |
+
"loss": 0.7437,
|
| 1640 |
+
"step": 2330
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 1.195707715891671,
|
| 1644 |
+
"grad_norm": 0.18822524406653396,
|
| 1645 |
+
"learning_rate": 5e-06,
|
| 1646 |
+
"loss": 0.7469,
|
| 1647 |
+
"step": 2340
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 1.200817577925396,
|
| 1651 |
+
"grad_norm": 0.17979619340691932,
|
| 1652 |
+
"learning_rate": 5e-06,
|
| 1653 |
+
"loss": 0.7476,
|
| 1654 |
+
"step": 2350
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 1.205927439959121,
|
| 1658 |
+
"grad_norm": 0.19302201862857232,
|
| 1659 |
+
"learning_rate": 5e-06,
|
| 1660 |
+
"loss": 0.7519,
|
| 1661 |
+
"step": 2360
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 1.2110373019928462,
|
| 1665 |
+
"grad_norm": 0.17331596795239365,
|
| 1666 |
+
"learning_rate": 5e-06,
|
| 1667 |
+
"loss": 0.7421,
|
| 1668 |
+
"step": 2370
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 1.2161471640265713,
|
| 1672 |
+
"grad_norm": 0.1900974772777454,
|
| 1673 |
+
"learning_rate": 5e-06,
|
| 1674 |
+
"loss": 0.7491,
|
| 1675 |
+
"step": 2380
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 1.2212570260602964,
|
| 1679 |
+
"grad_norm": 0.18235079887869074,
|
| 1680 |
+
"learning_rate": 5e-06,
|
| 1681 |
+
"loss": 0.7398,
|
| 1682 |
+
"step": 2390
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 1.2263668880940215,
|
| 1686 |
+
"grad_norm": 0.1990024061618005,
|
| 1687 |
+
"learning_rate": 5e-06,
|
| 1688 |
+
"loss": 0.7367,
|
| 1689 |
+
"step": 2400
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 1.2314767501277466,
|
| 1693 |
+
"grad_norm": 0.19774880108315787,
|
| 1694 |
+
"learning_rate": 5e-06,
|
| 1695 |
+
"loss": 0.7509,
|
| 1696 |
+
"step": 2410
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 1.2365866121614717,
|
| 1700 |
+
"grad_norm": 0.18038613594979708,
|
| 1701 |
+
"learning_rate": 5e-06,
|
| 1702 |
+
"loss": 0.7397,
|
| 1703 |
+
"step": 2420
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 1.2416964741951968,
|
| 1707 |
+
"grad_norm": 0.19148490320343095,
|
| 1708 |
+
"learning_rate": 5e-06,
|
| 1709 |
+
"loss": 0.738,
|
| 1710 |
+
"step": 2430
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 1.246806336228922,
|
| 1714 |
+
"grad_norm": 0.1764579818726389,
|
| 1715 |
+
"learning_rate": 5e-06,
|
| 1716 |
+
"loss": 0.7344,
|
| 1717 |
+
"step": 2440
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 1.2519161982626468,
|
| 1721 |
+
"grad_norm": 0.19292171667184566,
|
| 1722 |
+
"learning_rate": 5e-06,
|
| 1723 |
+
"loss": 0.7386,
|
| 1724 |
+
"step": 2450
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 1.257026060296372,
|
| 1728 |
+
"grad_norm": 0.1830054013064937,
|
| 1729 |
+
"learning_rate": 5e-06,
|
| 1730 |
+
"loss": 0.7428,
|
| 1731 |
+
"step": 2460
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 1.262135922330097,
|
| 1735 |
+
"grad_norm": 0.1771188524320852,
|
| 1736 |
+
"learning_rate": 5e-06,
|
| 1737 |
+
"loss": 0.7349,
|
| 1738 |
+
"step": 2470
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 1.2672457843638223,
|
| 1742 |
+
"grad_norm": 0.19308200340380383,
|
| 1743 |
+
"learning_rate": 5e-06,
|
| 1744 |
+
"loss": 0.7428,
|
| 1745 |
+
"step": 2480
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 1.2723556463975472,
|
| 1749 |
+
"grad_norm": 0.17723395601019812,
|
| 1750 |
+
"learning_rate": 5e-06,
|
| 1751 |
+
"loss": 0.7271,
|
| 1752 |
+
"step": 2490
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 1.2774655084312725,
|
| 1756 |
+
"grad_norm": 0.21808963432632347,
|
| 1757 |
+
"learning_rate": 5e-06,
|
| 1758 |
+
"loss": 0.7551,
|
| 1759 |
+
"step": 2500
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 1.2825753704649974,
|
| 1763 |
+
"grad_norm": 0.1944324454402299,
|
| 1764 |
+
"learning_rate": 5e-06,
|
| 1765 |
+
"loss": 0.7388,
|
| 1766 |
+
"step": 2510
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 1.2876852324987225,
|
| 1770 |
+
"grad_norm": 0.1737167892046253,
|
| 1771 |
+
"learning_rate": 5e-06,
|
| 1772 |
+
"loss": 0.7383,
|
| 1773 |
+
"step": 2520
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 1.2927950945324476,
|
| 1777 |
+
"grad_norm": 0.19316243219818216,
|
| 1778 |
+
"learning_rate": 5e-06,
|
| 1779 |
+
"loss": 0.7447,
|
| 1780 |
+
"step": 2530
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 1.2979049565661727,
|
| 1784 |
+
"grad_norm": 0.18271045253382115,
|
| 1785 |
+
"learning_rate": 5e-06,
|
| 1786 |
+
"loss": 0.7439,
|
| 1787 |
+
"step": 2540
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 1.3030148185998978,
|
| 1791 |
+
"grad_norm": 0.19235060247622612,
|
| 1792 |
+
"learning_rate": 5e-06,
|
| 1793 |
+
"loss": 0.7362,
|
| 1794 |
+
"step": 2550
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 1.3081246806336229,
|
| 1798 |
+
"grad_norm": 0.17975422534798727,
|
| 1799 |
+
"learning_rate": 5e-06,
|
| 1800 |
+
"loss": 0.7355,
|
| 1801 |
+
"step": 2560
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 1.313234542667348,
|
| 1805 |
+
"grad_norm": 0.19133431284185412,
|
| 1806 |
+
"learning_rate": 5e-06,
|
| 1807 |
+
"loss": 0.7474,
|
| 1808 |
+
"step": 2570
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 1.318344404701073,
|
| 1812 |
+
"grad_norm": 0.18525107476229646,
|
| 1813 |
+
"learning_rate": 5e-06,
|
| 1814 |
+
"loss": 0.7362,
|
| 1815 |
+
"step": 2580
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 1.3234542667347982,
|
| 1819 |
+
"grad_norm": 0.17940598778920888,
|
| 1820 |
+
"learning_rate": 5e-06,
|
| 1821 |
+
"loss": 0.7456,
|
| 1822 |
+
"step": 2590
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 1.3285641287685233,
|
| 1826 |
+
"grad_norm": 0.18377019808836909,
|
| 1827 |
+
"learning_rate": 5e-06,
|
| 1828 |
+
"loss": 0.7371,
|
| 1829 |
+
"step": 2600
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 1.3336739908022484,
|
| 1833 |
+
"grad_norm": 0.19380006248186266,
|
| 1834 |
+
"learning_rate": 5e-06,
|
| 1835 |
+
"loss": 0.7382,
|
| 1836 |
+
"step": 2610
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 1.3387838528359735,
|
| 1840 |
+
"grad_norm": 0.19130341477919996,
|
| 1841 |
+
"learning_rate": 5e-06,
|
| 1842 |
+
"loss": 0.7249,
|
| 1843 |
+
"step": 2620
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 1.3438937148696986,
|
| 1847 |
+
"grad_norm": 0.19399937386692442,
|
| 1848 |
+
"learning_rate": 5e-06,
|
| 1849 |
+
"loss": 0.7461,
|
| 1850 |
+
"step": 2630
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 1.3490035769034237,
|
| 1854 |
+
"grad_norm": 0.19026606995853784,
|
| 1855 |
+
"learning_rate": 5e-06,
|
| 1856 |
+
"loss": 0.744,
|
| 1857 |
+
"step": 2640
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 1.3541134389371488,
|
| 1861 |
+
"grad_norm": 0.17865393066414276,
|
| 1862 |
+
"learning_rate": 5e-06,
|
| 1863 |
+
"loss": 0.7562,
|
| 1864 |
+
"step": 2650
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 1.3592233009708738,
|
| 1868 |
+
"grad_norm": 0.18268806699965215,
|
| 1869 |
+
"learning_rate": 5e-06,
|
| 1870 |
+
"loss": 0.7369,
|
| 1871 |
+
"step": 2660
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 1.364333163004599,
|
| 1875 |
+
"grad_norm": 0.1859894512893362,
|
| 1876 |
+
"learning_rate": 5e-06,
|
| 1877 |
+
"loss": 0.7303,
|
| 1878 |
+
"step": 2670
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 1.369443025038324,
|
| 1882 |
+
"grad_norm": 0.1858579729895718,
|
| 1883 |
+
"learning_rate": 5e-06,
|
| 1884 |
+
"loss": 0.7338,
|
| 1885 |
+
"step": 2680
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 1.3745528870720491,
|
| 1889 |
+
"grad_norm": 0.18844334675064925,
|
| 1890 |
+
"learning_rate": 5e-06,
|
| 1891 |
+
"loss": 0.7399,
|
| 1892 |
+
"step": 2690
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 1.379662749105774,
|
| 1896 |
+
"grad_norm": 0.19488780104365555,
|
| 1897 |
+
"learning_rate": 5e-06,
|
| 1898 |
+
"loss": 0.7297,
|
| 1899 |
+
"step": 2700
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 1.3847726111394993,
|
| 1903 |
+
"grad_norm": 0.2012609184785339,
|
| 1904 |
+
"learning_rate": 5e-06,
|
| 1905 |
+
"loss": 0.7438,
|
| 1906 |
+
"step": 2710
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 1.3898824731732242,
|
| 1910 |
+
"grad_norm": 0.19240664121181153,
|
| 1911 |
+
"learning_rate": 5e-06,
|
| 1912 |
+
"loss": 0.724,
|
| 1913 |
+
"step": 2720
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 1.3949923352069495,
|
| 1917 |
+
"grad_norm": 0.1989866337354731,
|
| 1918 |
+
"learning_rate": 5e-06,
|
| 1919 |
+
"loss": 0.7387,
|
| 1920 |
+
"step": 2730
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 1.4001021972406744,
|
| 1924 |
+
"grad_norm": 0.19173593318229185,
|
| 1925 |
+
"learning_rate": 5e-06,
|
| 1926 |
+
"loss": 0.7389,
|
| 1927 |
+
"step": 2740
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 1.4052120592743995,
|
| 1931 |
+
"grad_norm": 0.18490218706957628,
|
| 1932 |
+
"learning_rate": 5e-06,
|
| 1933 |
+
"loss": 0.7354,
|
| 1934 |
+
"step": 2750
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 1.4103219213081246,
|
| 1938 |
+
"grad_norm": 0.19596927441115194,
|
| 1939 |
+
"learning_rate": 5e-06,
|
| 1940 |
+
"loss": 0.7294,
|
| 1941 |
+
"step": 2760
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 1.4154317833418497,
|
| 1945 |
+
"grad_norm": 0.1884149575008095,
|
| 1946 |
+
"learning_rate": 5e-06,
|
| 1947 |
+
"loss": 0.7535,
|
| 1948 |
+
"step": 2770
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 1.4205416453755748,
|
| 1952 |
+
"grad_norm": 0.19011351454021505,
|
| 1953 |
+
"learning_rate": 5e-06,
|
| 1954 |
+
"loss": 0.7311,
|
| 1955 |
+
"step": 2780
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 1.4256515074093,
|
| 1959 |
+
"grad_norm": 0.18330086878582655,
|
| 1960 |
+
"learning_rate": 5e-06,
|
| 1961 |
+
"loss": 0.7238,
|
| 1962 |
+
"step": 2790
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 1.430761369443025,
|
| 1966 |
+
"grad_norm": 0.18245127655957494,
|
| 1967 |
+
"learning_rate": 5e-06,
|
| 1968 |
+
"loss": 0.7297,
|
| 1969 |
+
"step": 2800
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 1.43587123147675,
|
| 1973 |
+
"grad_norm": 0.1899683440493854,
|
| 1974 |
+
"learning_rate": 5e-06,
|
| 1975 |
+
"loss": 0.7388,
|
| 1976 |
+
"step": 2810
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 1.4409810935104752,
|
| 1980 |
+
"grad_norm": 0.17955923753560576,
|
| 1981 |
+
"learning_rate": 5e-06,
|
| 1982 |
+
"loss": 0.7382,
|
| 1983 |
+
"step": 2820
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 1.4460909555442003,
|
| 1987 |
+
"grad_norm": 0.19252118964036657,
|
| 1988 |
+
"learning_rate": 5e-06,
|
| 1989 |
+
"loss": 0.7428,
|
| 1990 |
+
"step": 2830
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 1.4512008175779254,
|
| 1994 |
+
"grad_norm": 0.1993095814580447,
|
| 1995 |
+
"learning_rate": 5e-06,
|
| 1996 |
+
"loss": 0.7453,
|
| 1997 |
+
"step": 2840
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 1.4563106796116505,
|
| 2001 |
+
"grad_norm": 0.19453340379561043,
|
| 2002 |
+
"learning_rate": 5e-06,
|
| 2003 |
+
"loss": 0.7316,
|
| 2004 |
+
"step": 2850
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 1.4614205416453756,
|
| 2008 |
+
"grad_norm": 0.18991487065634022,
|
| 2009 |
+
"learning_rate": 5e-06,
|
| 2010 |
+
"loss": 0.7386,
|
| 2011 |
+
"step": 2860
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 1.4665304036791007,
|
| 2015 |
+
"grad_norm": 0.18973483396294438,
|
| 2016 |
+
"learning_rate": 5e-06,
|
| 2017 |
+
"loss": 0.7466,
|
| 2018 |
+
"step": 2870
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 1.4716402657128258,
|
| 2022 |
+
"grad_norm": 0.1919078737205217,
|
| 2023 |
+
"learning_rate": 5e-06,
|
| 2024 |
+
"loss": 0.7495,
|
| 2025 |
+
"step": 2880
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 1.476750127746551,
|
| 2029 |
+
"grad_norm": 0.1839372662925669,
|
| 2030 |
+
"learning_rate": 5e-06,
|
| 2031 |
+
"loss": 0.7389,
|
| 2032 |
+
"step": 2890
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 1.481859989780276,
|
| 2036 |
+
"grad_norm": 0.19249510537236922,
|
| 2037 |
+
"learning_rate": 5e-06,
|
| 2038 |
+
"loss": 0.749,
|
| 2039 |
+
"step": 2900
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 1.486969851814001,
|
| 2043 |
+
"grad_norm": 0.179246890906481,
|
| 2044 |
+
"learning_rate": 5e-06,
|
| 2045 |
+
"loss": 0.7282,
|
| 2046 |
+
"step": 2910
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 1.4920797138477262,
|
| 2050 |
+
"grad_norm": 0.19514232006253573,
|
| 2051 |
+
"learning_rate": 5e-06,
|
| 2052 |
+
"loss": 0.7446,
|
| 2053 |
+
"step": 2920
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 1.497189575881451,
|
| 2057 |
+
"grad_norm": 0.18596165643891152,
|
| 2058 |
+
"learning_rate": 5e-06,
|
| 2059 |
+
"loss": 0.7334,
|
| 2060 |
+
"step": 2930
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 1.5022994379151764,
|
| 2064 |
+
"grad_norm": 0.18786595484474303,
|
| 2065 |
+
"learning_rate": 5e-06,
|
| 2066 |
+
"loss": 0.7279,
|
| 2067 |
+
"step": 2940
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 1.5074092999489013,
|
| 2071 |
+
"grad_norm": 0.1971738538736254,
|
| 2072 |
+
"learning_rate": 5e-06,
|
| 2073 |
+
"loss": 0.7461,
|
| 2074 |
+
"step": 2950
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 1.5125191619826266,
|
| 2078 |
+
"grad_norm": 0.18866040282412957,
|
| 2079 |
+
"learning_rate": 5e-06,
|
| 2080 |
+
"loss": 0.7474,
|
| 2081 |
+
"step": 2960
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 1.5176290240163515,
|
| 2085 |
+
"grad_norm": 0.19040908132898832,
|
| 2086 |
+
"learning_rate": 5e-06,
|
| 2087 |
+
"loss": 0.7445,
|
| 2088 |
+
"step": 2970
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 1.5227388860500768,
|
| 2092 |
+
"grad_norm": 0.1929810400978154,
|
| 2093 |
+
"learning_rate": 5e-06,
|
| 2094 |
+
"loss": 0.7362,
|
| 2095 |
+
"step": 2980
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 1.5278487480838017,
|
| 2099 |
+
"grad_norm": 0.19442375737357984,
|
| 2100 |
+
"learning_rate": 5e-06,
|
| 2101 |
+
"loss": 0.7305,
|
| 2102 |
+
"step": 2990
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 1.532958610117527,
|
| 2106 |
+
"grad_norm": 0.18546826858387847,
|
| 2107 |
+
"learning_rate": 5e-06,
|
| 2108 |
+
"loss": 0.7311,
|
| 2109 |
+
"step": 3000
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 1.5380684721512519,
|
| 2113 |
+
"grad_norm": 0.18542784404927515,
|
| 2114 |
+
"learning_rate": 5e-06,
|
| 2115 |
+
"loss": 0.7447,
|
| 2116 |
+
"step": 3010
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 1.543178334184977,
|
| 2120 |
+
"grad_norm": 0.2020846723209545,
|
| 2121 |
+
"learning_rate": 5e-06,
|
| 2122 |
+
"loss": 0.7113,
|
| 2123 |
+
"step": 3020
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 1.548288196218702,
|
| 2127 |
+
"grad_norm": 0.19026884893782828,
|
| 2128 |
+
"learning_rate": 5e-06,
|
| 2129 |
+
"loss": 0.7157,
|
| 2130 |
+
"step": 3030
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 1.5533980582524272,
|
| 2134 |
+
"grad_norm": 0.18111421662938304,
|
| 2135 |
+
"learning_rate": 5e-06,
|
| 2136 |
+
"loss": 0.7323,
|
| 2137 |
+
"step": 3040
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 1.5585079202861523,
|
| 2141 |
+
"grad_norm": 0.19367385202342016,
|
| 2142 |
+
"learning_rate": 5e-06,
|
| 2143 |
+
"loss": 0.7305,
|
| 2144 |
+
"step": 3050
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 1.5636177823198774,
|
| 2148 |
+
"grad_norm": 0.18590394121821466,
|
| 2149 |
+
"learning_rate": 5e-06,
|
| 2150 |
+
"loss": 0.7341,
|
| 2151 |
+
"step": 3060
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 1.5687276443536025,
|
| 2155 |
+
"grad_norm": 0.18488441186992707,
|
| 2156 |
+
"learning_rate": 5e-06,
|
| 2157 |
+
"loss": 0.7482,
|
| 2158 |
+
"step": 3070
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 1.5738375063873276,
|
| 2162 |
+
"grad_norm": 0.18226306867076514,
|
| 2163 |
+
"learning_rate": 5e-06,
|
| 2164 |
+
"loss": 0.7334,
|
| 2165 |
+
"step": 3080
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 1.5789473684210527,
|
| 2169 |
+
"grad_norm": 0.20053856155426641,
|
| 2170 |
+
"learning_rate": 5e-06,
|
| 2171 |
+
"loss": 0.7414,
|
| 2172 |
+
"step": 3090
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 1.5840572304547778,
|
| 2176 |
+
"grad_norm": 0.19672564131420983,
|
| 2177 |
+
"learning_rate": 5e-06,
|
| 2178 |
+
"loss": 0.7508,
|
| 2179 |
+
"step": 3100
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 1.5891670924885029,
|
| 2183 |
+
"grad_norm": 0.1790851772225089,
|
| 2184 |
+
"learning_rate": 5e-06,
|
| 2185 |
+
"loss": 0.7471,
|
| 2186 |
+
"step": 3110
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 1.594276954522228,
|
| 2190 |
+
"grad_norm": 0.1900047612676954,
|
| 2191 |
+
"learning_rate": 5e-06,
|
| 2192 |
+
"loss": 0.7421,
|
| 2193 |
+
"step": 3120
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 1.599386816555953,
|
| 2197 |
+
"grad_norm": 0.19746465955340986,
|
| 2198 |
+
"learning_rate": 5e-06,
|
| 2199 |
+
"loss": 0.7471,
|
| 2200 |
+
"step": 3130
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 1.604496678589678,
|
| 2204 |
+
"grad_norm": 0.186549540683221,
|
| 2205 |
+
"learning_rate": 5e-06,
|
| 2206 |
+
"loss": 0.7275,
|
| 2207 |
+
"step": 3140
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 1.6096065406234032,
|
| 2211 |
+
"grad_norm": 0.1876261054598287,
|
| 2212 |
+
"learning_rate": 5e-06,
|
| 2213 |
+
"loss": 0.7359,
|
| 2214 |
+
"step": 3150
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 1.6147164026571281,
|
| 2218 |
+
"grad_norm": 0.19082325492370317,
|
| 2219 |
+
"learning_rate": 5e-06,
|
| 2220 |
+
"loss": 0.7268,
|
| 2221 |
+
"step": 3160
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 1.6198262646908534,
|
| 2225 |
+
"grad_norm": 0.2016402119888201,
|
| 2226 |
+
"learning_rate": 5e-06,
|
| 2227 |
+
"loss": 0.7377,
|
| 2228 |
+
"step": 3170
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 1.6249361267245783,
|
| 2232 |
+
"grad_norm": 0.1888126317070555,
|
| 2233 |
+
"learning_rate": 5e-06,
|
| 2234 |
+
"loss": 0.7265,
|
| 2235 |
+
"step": 3180
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 1.6300459887583036,
|
| 2239 |
+
"grad_norm": 0.17743730583327474,
|
| 2240 |
+
"learning_rate": 5e-06,
|
| 2241 |
+
"loss": 0.7203,
|
| 2242 |
+
"step": 3190
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 1.6351558507920285,
|
| 2246 |
+
"grad_norm": 0.1826162903853255,
|
| 2247 |
+
"learning_rate": 5e-06,
|
| 2248 |
+
"loss": 0.7215,
|
| 2249 |
+
"step": 3200
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 1.6402657128257538,
|
| 2253 |
+
"grad_norm": 0.19419266754404552,
|
| 2254 |
+
"learning_rate": 5e-06,
|
| 2255 |
+
"loss": 0.7376,
|
| 2256 |
+
"step": 3210
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 1.6453755748594787,
|
| 2260 |
+
"grad_norm": 0.1956453565355767,
|
| 2261 |
+
"learning_rate": 5e-06,
|
| 2262 |
+
"loss": 0.7316,
|
| 2263 |
+
"step": 3220
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 1.650485436893204,
|
| 2267 |
+
"grad_norm": 0.19765143129125318,
|
| 2268 |
+
"learning_rate": 5e-06,
|
| 2269 |
+
"loss": 0.7374,
|
| 2270 |
+
"step": 3230
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 1.655595298926929,
|
| 2274 |
+
"grad_norm": 0.19127982430051427,
|
| 2275 |
+
"learning_rate": 5e-06,
|
| 2276 |
+
"loss": 0.7405,
|
| 2277 |
+
"step": 3240
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 1.660705160960654,
|
| 2281 |
+
"grad_norm": 0.1847472801458583,
|
| 2282 |
+
"learning_rate": 5e-06,
|
| 2283 |
+
"loss": 0.724,
|
| 2284 |
+
"step": 3250
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 1.6658150229943791,
|
| 2288 |
+
"grad_norm": 0.18698307703261788,
|
| 2289 |
+
"learning_rate": 5e-06,
|
| 2290 |
+
"loss": 0.7374,
|
| 2291 |
+
"step": 3260
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 1.6709248850281042,
|
| 2295 |
+
"grad_norm": 0.17533798523255767,
|
| 2296 |
+
"learning_rate": 5e-06,
|
| 2297 |
+
"loss": 0.7202,
|
| 2298 |
+
"step": 3270
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 1.6760347470618293,
|
| 2302 |
+
"grad_norm": 0.1806351825859557,
|
| 2303 |
+
"learning_rate": 5e-06,
|
| 2304 |
+
"loss": 0.7452,
|
| 2305 |
+
"step": 3280
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 1.6811446090955544,
|
| 2309 |
+
"grad_norm": 0.1767976805961292,
|
| 2310 |
+
"learning_rate": 5e-06,
|
| 2311 |
+
"loss": 0.7338,
|
| 2312 |
+
"step": 3290
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 1.6862544711292795,
|
| 2316 |
+
"grad_norm": 0.19498984484111873,
|
| 2317 |
+
"learning_rate": 5e-06,
|
| 2318 |
+
"loss": 0.7403,
|
| 2319 |
+
"step": 3300
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 1.6913643331630046,
|
| 2323 |
+
"grad_norm": 0.17701307669892918,
|
| 2324 |
+
"learning_rate": 5e-06,
|
| 2325 |
+
"loss": 0.7299,
|
| 2326 |
+
"step": 3310
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 1.6964741951967297,
|
| 2330 |
+
"grad_norm": 0.19220216566407472,
|
| 2331 |
+
"learning_rate": 5e-06,
|
| 2332 |
+
"loss": 0.7314,
|
| 2333 |
+
"step": 3320
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 1.7015840572304548,
|
| 2337 |
+
"grad_norm": 0.1829279730231264,
|
| 2338 |
+
"learning_rate": 5e-06,
|
| 2339 |
+
"loss": 0.743,
|
| 2340 |
+
"step": 3330
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 1.70669391926418,
|
| 2344 |
+
"grad_norm": 0.19526766653061225,
|
| 2345 |
+
"learning_rate": 5e-06,
|
| 2346 |
+
"loss": 0.7222,
|
| 2347 |
+
"step": 3340
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 1.711803781297905,
|
| 2351 |
+
"grad_norm": 0.19455609962672274,
|
| 2352 |
+
"learning_rate": 5e-06,
|
| 2353 |
+
"loss": 0.7253,
|
| 2354 |
+
"step": 3350
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 1.71691364333163,
|
| 2358 |
+
"grad_norm": 0.21002379536162816,
|
| 2359 |
+
"learning_rate": 5e-06,
|
| 2360 |
+
"loss": 0.7429,
|
| 2361 |
+
"step": 3360
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 1.722023505365355,
|
| 2365 |
+
"grad_norm": 0.1990882316461353,
|
| 2366 |
+
"learning_rate": 5e-06,
|
| 2367 |
+
"loss": 0.7443,
|
| 2368 |
+
"step": 3370
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 1.7271333673990803,
|
| 2372 |
+
"grad_norm": 0.17934672167038826,
|
| 2373 |
+
"learning_rate": 5e-06,
|
| 2374 |
+
"loss": 0.7497,
|
| 2375 |
+
"step": 3380
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 1.7322432294328052,
|
| 2379 |
+
"grad_norm": 0.19501165745940965,
|
| 2380 |
+
"learning_rate": 5e-06,
|
| 2381 |
+
"loss": 0.7425,
|
| 2382 |
+
"step": 3390
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 1.7373530914665305,
|
| 2386 |
+
"grad_norm": 0.19248650606756543,
|
| 2387 |
+
"learning_rate": 5e-06,
|
| 2388 |
+
"loss": 0.7297,
|
| 2389 |
+
"step": 3400
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 1.7424629535002554,
|
| 2393 |
+
"grad_norm": 0.17721599710417338,
|
| 2394 |
+
"learning_rate": 5e-06,
|
| 2395 |
+
"loss": 0.7251,
|
| 2396 |
+
"step": 3410
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 1.7475728155339807,
|
| 2400 |
+
"grad_norm": 0.18509365156353424,
|
| 2401 |
+
"learning_rate": 5e-06,
|
| 2402 |
+
"loss": 0.7221,
|
| 2403 |
+
"step": 3420
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 1.7526826775677056,
|
| 2407 |
+
"grad_norm": 0.18289284691122754,
|
| 2408 |
+
"learning_rate": 5e-06,
|
| 2409 |
+
"loss": 0.7327,
|
| 2410 |
+
"step": 3430
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 1.757792539601431,
|
| 2414 |
+
"grad_norm": 0.18756279151165123,
|
| 2415 |
+
"learning_rate": 5e-06,
|
| 2416 |
+
"loss": 0.732,
|
| 2417 |
+
"step": 3440
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 1.7629024016351558,
|
| 2421 |
+
"grad_norm": 0.17439303769229625,
|
| 2422 |
+
"learning_rate": 5e-06,
|
| 2423 |
+
"loss": 0.7344,
|
| 2424 |
+
"step": 3450
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 1.768012263668881,
|
| 2428 |
+
"grad_norm": 0.17783376482824478,
|
| 2429 |
+
"learning_rate": 5e-06,
|
| 2430 |
+
"loss": 0.7323,
|
| 2431 |
+
"step": 3460
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 1.773122125702606,
|
| 2435 |
+
"grad_norm": 0.19448194078586717,
|
| 2436 |
+
"learning_rate": 5e-06,
|
| 2437 |
+
"loss": 0.7292,
|
| 2438 |
+
"step": 3470
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 1.778231987736331,
|
| 2442 |
+
"grad_norm": 0.18000237860427712,
|
| 2443 |
+
"learning_rate": 5e-06,
|
| 2444 |
+
"loss": 0.7219,
|
| 2445 |
+
"step": 3480
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 1.7833418497700562,
|
| 2449 |
+
"grad_norm": 0.18519882940772772,
|
| 2450 |
+
"learning_rate": 5e-06,
|
| 2451 |
+
"loss": 0.7354,
|
| 2452 |
+
"step": 3490
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 1.7884517118037813,
|
| 2456 |
+
"grad_norm": 0.19301292549147336,
|
| 2457 |
+
"learning_rate": 5e-06,
|
| 2458 |
+
"loss": 0.7487,
|
| 2459 |
+
"step": 3500
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 1.7935615738375064,
|
| 2463 |
+
"grad_norm": 0.17758380897102066,
|
| 2464 |
+
"learning_rate": 5e-06,
|
| 2465 |
+
"loss": 0.7373,
|
| 2466 |
+
"step": 3510
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 1.7986714358712315,
|
| 2470 |
+
"grad_norm": 0.1794720757802905,
|
| 2471 |
+
"learning_rate": 5e-06,
|
| 2472 |
+
"loss": 0.7226,
|
| 2473 |
+
"step": 3520
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 1.8037812979049566,
|
| 2477 |
+
"grad_norm": 0.18100374933694008,
|
| 2478 |
+
"learning_rate": 5e-06,
|
| 2479 |
+
"loss": 0.7256,
|
| 2480 |
+
"step": 3530
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 1.8088911599386817,
|
| 2484 |
+
"grad_norm": 0.1954603145633284,
|
| 2485 |
+
"learning_rate": 5e-06,
|
| 2486 |
+
"loss": 0.7283,
|
| 2487 |
+
"step": 3540
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 1.8140010219724068,
|
| 2491 |
+
"grad_norm": 0.19558607958635285,
|
| 2492 |
+
"learning_rate": 5e-06,
|
| 2493 |
+
"loss": 0.7384,
|
| 2494 |
+
"step": 3550
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 1.8191108840061319,
|
| 2498 |
+
"grad_norm": 0.1772107537935853,
|
| 2499 |
+
"learning_rate": 5e-06,
|
| 2500 |
+
"loss": 0.7382,
|
| 2501 |
+
"step": 3560
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 1.824220746039857,
|
| 2505 |
+
"grad_norm": 0.17916901000763397,
|
| 2506 |
+
"learning_rate": 5e-06,
|
| 2507 |
+
"loss": 0.7367,
|
| 2508 |
+
"step": 3570
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 1.829330608073582,
|
| 2512 |
+
"grad_norm": 0.19083482072843658,
|
| 2513 |
+
"learning_rate": 5e-06,
|
| 2514 |
+
"loss": 0.7247,
|
| 2515 |
+
"step": 3580
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 1.8344404701073072,
|
| 2519 |
+
"grad_norm": 0.1770449813805881,
|
| 2520 |
+
"learning_rate": 5e-06,
|
| 2521 |
+
"loss": 0.7187,
|
| 2522 |
+
"step": 3590
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 1.839550332141032,
|
| 2526 |
+
"grad_norm": 0.18790158384523442,
|
| 2527 |
+
"learning_rate": 5e-06,
|
| 2528 |
+
"loss": 0.7403,
|
| 2529 |
+
"step": 3600
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 1.8446601941747574,
|
| 2533 |
+
"grad_norm": 0.17892362311216914,
|
| 2534 |
+
"learning_rate": 5e-06,
|
| 2535 |
+
"loss": 0.7368,
|
| 2536 |
+
"step": 3610
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 1.8497700562084822,
|
| 2540 |
+
"grad_norm": 0.1839659251785667,
|
| 2541 |
+
"learning_rate": 5e-06,
|
| 2542 |
+
"loss": 0.7255,
|
| 2543 |
+
"step": 3620
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 1.8548799182422075,
|
| 2547 |
+
"grad_norm": 0.19138496555502849,
|
| 2548 |
+
"learning_rate": 5e-06,
|
| 2549 |
+
"loss": 0.7453,
|
| 2550 |
+
"step": 3630
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 1.8599897802759324,
|
| 2554 |
+
"grad_norm": 0.18135734354491537,
|
| 2555 |
+
"learning_rate": 5e-06,
|
| 2556 |
+
"loss": 0.7166,
|
| 2557 |
+
"step": 3640
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 1.8650996423096577,
|
| 2561 |
+
"grad_norm": 0.1859082044449026,
|
| 2562 |
+
"learning_rate": 5e-06,
|
| 2563 |
+
"loss": 0.7285,
|
| 2564 |
+
"step": 3650
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 1.8702095043433826,
|
| 2568 |
+
"grad_norm": 0.1913280855307758,
|
| 2569 |
+
"learning_rate": 5e-06,
|
| 2570 |
+
"loss": 0.7279,
|
| 2571 |
+
"step": 3660
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 1.875319366377108,
|
| 2575 |
+
"grad_norm": 0.19148047998384163,
|
| 2576 |
+
"learning_rate": 5e-06,
|
| 2577 |
+
"loss": 0.7381,
|
| 2578 |
+
"step": 3670
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 1.8804292284108328,
|
| 2582 |
+
"grad_norm": 0.190776629149848,
|
| 2583 |
+
"learning_rate": 5e-06,
|
| 2584 |
+
"loss": 0.7347,
|
| 2585 |
+
"step": 3680
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 1.8855390904445581,
|
| 2589 |
+
"grad_norm": 0.18748890464701637,
|
| 2590 |
+
"learning_rate": 5e-06,
|
| 2591 |
+
"loss": 0.7214,
|
| 2592 |
+
"step": 3690
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 1.890648952478283,
|
| 2596 |
+
"grad_norm": 0.19691029617370956,
|
| 2597 |
+
"learning_rate": 5e-06,
|
| 2598 |
+
"loss": 0.7396,
|
| 2599 |
+
"step": 3700
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 1.8957588145120083,
|
| 2603 |
+
"grad_norm": 0.17143385370457725,
|
| 2604 |
+
"learning_rate": 5e-06,
|
| 2605 |
+
"loss": 0.7305,
|
| 2606 |
+
"step": 3710
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 1.9008686765457332,
|
| 2610 |
+
"grad_norm": 0.19115494111352774,
|
| 2611 |
+
"learning_rate": 5e-06,
|
| 2612 |
+
"loss": 0.7252,
|
| 2613 |
+
"step": 3720
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 1.9059785385794583,
|
| 2617 |
+
"grad_norm": 0.18589155104150923,
|
| 2618 |
+
"learning_rate": 5e-06,
|
| 2619 |
+
"loss": 0.728,
|
| 2620 |
+
"step": 3730
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 1.9110884006131834,
|
| 2624 |
+
"grad_norm": 0.1870078279938856,
|
| 2625 |
+
"learning_rate": 5e-06,
|
| 2626 |
+
"loss": 0.7259,
|
| 2627 |
+
"step": 3740
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 1.9161982626469085,
|
| 2631 |
+
"grad_norm": 0.17909224396912188,
|
| 2632 |
+
"learning_rate": 5e-06,
|
| 2633 |
+
"loss": 0.7345,
|
| 2634 |
+
"step": 3750
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 1.9213081246806336,
|
| 2638 |
+
"grad_norm": 0.19885632401705697,
|
| 2639 |
+
"learning_rate": 5e-06,
|
| 2640 |
+
"loss": 0.7289,
|
| 2641 |
+
"step": 3760
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 1.9264179867143587,
|
| 2645 |
+
"grad_norm": 0.18778552722329356,
|
| 2646 |
+
"learning_rate": 5e-06,
|
| 2647 |
+
"loss": 0.725,
|
| 2648 |
+
"step": 3770
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 1.9315278487480838,
|
| 2652 |
+
"grad_norm": 0.2019790249495369,
|
| 2653 |
+
"learning_rate": 5e-06,
|
| 2654 |
+
"loss": 0.7398,
|
| 2655 |
+
"step": 3780
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 1.936637710781809,
|
| 2659 |
+
"grad_norm": 0.19338701904495897,
|
| 2660 |
+
"learning_rate": 5e-06,
|
| 2661 |
+
"loss": 0.7265,
|
| 2662 |
+
"step": 3790
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 1.941747572815534,
|
| 2666 |
+
"grad_norm": 0.18703296264974872,
|
| 2667 |
+
"learning_rate": 5e-06,
|
| 2668 |
+
"loss": 0.7332,
|
| 2669 |
+
"step": 3800
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 1.946857434849259,
|
| 2673 |
+
"grad_norm": 0.1700175440342506,
|
| 2674 |
+
"learning_rate": 5e-06,
|
| 2675 |
+
"loss": 0.7205,
|
| 2676 |
+
"step": 3810
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 1.9519672968829842,
|
| 2680 |
+
"grad_norm": 0.18636496202992153,
|
| 2681 |
+
"learning_rate": 5e-06,
|
| 2682 |
+
"loss": 0.7154,
|
| 2683 |
+
"step": 3820
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 1.9570771589167093,
|
| 2687 |
+
"grad_norm": 0.1826391337083993,
|
| 2688 |
+
"learning_rate": 5e-06,
|
| 2689 |
+
"loss": 0.7356,
|
| 2690 |
+
"step": 3830
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 1.9621870209504344,
|
| 2694 |
+
"grad_norm": 0.17766191115765154,
|
| 2695 |
+
"learning_rate": 5e-06,
|
| 2696 |
+
"loss": 0.7369,
|
| 2697 |
+
"step": 3840
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 1.9672968829841593,
|
| 2701 |
+
"grad_norm": 0.18034528150782342,
|
| 2702 |
+
"learning_rate": 5e-06,
|
| 2703 |
+
"loss": 0.7288,
|
| 2704 |
+
"step": 3850
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 1.9724067450178846,
|
| 2708 |
+
"grad_norm": 0.18541894497456152,
|
| 2709 |
+
"learning_rate": 5e-06,
|
| 2710 |
+
"loss": 0.7296,
|
| 2711 |
+
"step": 3860
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 1.9775166070516095,
|
| 2715 |
+
"grad_norm": 0.19539634425789987,
|
| 2716 |
+
"learning_rate": 5e-06,
|
| 2717 |
+
"loss": 0.7374,
|
| 2718 |
+
"step": 3870
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 1.9826264690853348,
|
| 2722 |
+
"grad_norm": 0.1928837424204438,
|
| 2723 |
+
"learning_rate": 5e-06,
|
| 2724 |
+
"loss": 0.732,
|
| 2725 |
+
"step": 3880
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 1.9877363311190597,
|
| 2729 |
+
"grad_norm": 0.18813671735265705,
|
| 2730 |
+
"learning_rate": 5e-06,
|
| 2731 |
+
"loss": 0.7285,
|
| 2732 |
+
"step": 3890
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 1.992846193152785,
|
| 2736 |
+
"grad_norm": 0.19024591983517306,
|
| 2737 |
+
"learning_rate": 5e-06,
|
| 2738 |
+
"loss": 0.7386,
|
| 2739 |
+
"step": 3900
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 1.9979560551865099,
|
| 2743 |
+
"grad_norm": 0.1791330764130833,
|
| 2744 |
+
"learning_rate": 5e-06,
|
| 2745 |
+
"loss": 0.7324,
|
| 2746 |
+
"step": 3910
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 2.003065917220235,
|
| 2750 |
+
"grad_norm": 0.18457472513280188,
|
| 2751 |
+
"learning_rate": 5e-06,
|
| 2752 |
+
"loss": 0.7303,
|
| 2753 |
+
"step": 3920
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 2.00817577925396,
|
| 2757 |
+
"grad_norm": 0.18684788166920566,
|
| 2758 |
+
"learning_rate": 5e-06,
|
| 2759 |
+
"loss": 0.735,
|
| 2760 |
+
"step": 3930
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 2.0132856412876854,
|
| 2764 |
+
"grad_norm": 0.18385152341485855,
|
| 2765 |
+
"learning_rate": 5e-06,
|
| 2766 |
+
"loss": 0.7204,
|
| 2767 |
+
"step": 3940
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 2.0183955033214103,
|
| 2771 |
+
"grad_norm": 0.1893087134675762,
|
| 2772 |
+
"learning_rate": 5e-06,
|
| 2773 |
+
"loss": 0.7271,
|
| 2774 |
+
"step": 3950
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 2.0235053653551356,
|
| 2778 |
+
"grad_norm": 0.19001500071497598,
|
| 2779 |
+
"learning_rate": 5e-06,
|
| 2780 |
+
"loss": 0.7286,
|
| 2781 |
+
"step": 3960
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 2.0286152273888605,
|
| 2785 |
+
"grad_norm": 0.18883227887144083,
|
| 2786 |
+
"learning_rate": 5e-06,
|
| 2787 |
+
"loss": 0.7157,
|
| 2788 |
+
"step": 3970
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 2.033725089422586,
|
| 2792 |
+
"grad_norm": 0.17689498199700174,
|
| 2793 |
+
"learning_rate": 5e-06,
|
| 2794 |
+
"loss": 0.7063,
|
| 2795 |
+
"step": 3980
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 2.0388349514563107,
|
| 2799 |
+
"grad_norm": 0.19270059169949594,
|
| 2800 |
+
"learning_rate": 5e-06,
|
| 2801 |
+
"loss": 0.7295,
|
| 2802 |
+
"step": 3990
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 2.043944813490036,
|
| 2806 |
+
"grad_norm": 0.19597358123850073,
|
| 2807 |
+
"learning_rate": 5e-06,
|
| 2808 |
+
"loss": 0.7268,
|
| 2809 |
+
"step": 4000
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 2.049054675523761,
|
| 2813 |
+
"grad_norm": 0.19711968415004932,
|
| 2814 |
+
"learning_rate": 5e-06,
|
| 2815 |
+
"loss": 0.7229,
|
| 2816 |
+
"step": 4010
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 2.0541645375574857,
|
| 2820 |
+
"grad_norm": 0.18776496261358783,
|
| 2821 |
+
"learning_rate": 5e-06,
|
| 2822 |
+
"loss": 0.7158,
|
| 2823 |
+
"step": 4020
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 2.059274399591211,
|
| 2827 |
+
"grad_norm": 0.18818737294591004,
|
| 2828 |
+
"learning_rate": 5e-06,
|
| 2829 |
+
"loss": 0.7194,
|
| 2830 |
+
"step": 4030
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 2.064384261624936,
|
| 2834 |
+
"grad_norm": 0.19845798280803176,
|
| 2835 |
+
"learning_rate": 5e-06,
|
| 2836 |
+
"loss": 0.7265,
|
| 2837 |
+
"step": 4040
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 2.0694941236586613,
|
| 2841 |
+
"grad_norm": 0.18713228958457867,
|
| 2842 |
+
"learning_rate": 5e-06,
|
| 2843 |
+
"loss": 0.7095,
|
| 2844 |
+
"step": 4050
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 2.074603985692386,
|
| 2848 |
+
"grad_norm": 0.17787870833728897,
|
| 2849 |
+
"learning_rate": 5e-06,
|
| 2850 |
+
"loss": 0.7149,
|
| 2851 |
+
"step": 4060
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 2.0797138477261115,
|
| 2855 |
+
"grad_norm": 0.19472810880013228,
|
| 2856 |
+
"learning_rate": 5e-06,
|
| 2857 |
+
"loss": 0.7178,
|
| 2858 |
+
"step": 4070
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 2.0848237097598363,
|
| 2862 |
+
"grad_norm": 0.19429450371850024,
|
| 2863 |
+
"learning_rate": 5e-06,
|
| 2864 |
+
"loss": 0.7118,
|
| 2865 |
+
"step": 4080
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 2.0899335717935617,
|
| 2869 |
+
"grad_norm": 0.1941609760118733,
|
| 2870 |
+
"learning_rate": 5e-06,
|
| 2871 |
+
"loss": 0.7265,
|
| 2872 |
+
"step": 4090
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 2.0950434338272865,
|
| 2876 |
+
"grad_norm": 0.19290976310635458,
|
| 2877 |
+
"learning_rate": 5e-06,
|
| 2878 |
+
"loss": 0.7085,
|
| 2879 |
+
"step": 4100
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 2.100153295861012,
|
| 2883 |
+
"grad_norm": 0.1765241017205207,
|
| 2884 |
+
"learning_rate": 5e-06,
|
| 2885 |
+
"loss": 0.7301,
|
| 2886 |
+
"step": 4110
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 2.1052631578947367,
|
| 2890 |
+
"grad_norm": 0.17846133756954982,
|
| 2891 |
+
"learning_rate": 5e-06,
|
| 2892 |
+
"loss": 0.7105,
|
| 2893 |
+
"step": 4120
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 2.110373019928462,
|
| 2897 |
+
"grad_norm": 0.1990655813201847,
|
| 2898 |
+
"learning_rate": 5e-06,
|
| 2899 |
+
"loss": 0.7197,
|
| 2900 |
+
"step": 4130
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 2.115482881962187,
|
| 2904 |
+
"grad_norm": 0.18227727070573727,
|
| 2905 |
+
"learning_rate": 5e-06,
|
| 2906 |
+
"loss": 0.7225,
|
| 2907 |
+
"step": 4140
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 2.1205927439959122,
|
| 2911 |
+
"grad_norm": 0.1950861579503913,
|
| 2912 |
+
"learning_rate": 5e-06,
|
| 2913 |
+
"loss": 0.7151,
|
| 2914 |
+
"step": 4150
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 2.125702606029637,
|
| 2918 |
+
"grad_norm": 0.1852273630173904,
|
| 2919 |
+
"learning_rate": 5e-06,
|
| 2920 |
+
"loss": 0.7349,
|
| 2921 |
+
"step": 4160
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 2.1308124680633624,
|
| 2925 |
+
"grad_norm": 0.19463032921601597,
|
| 2926 |
+
"learning_rate": 5e-06,
|
| 2927 |
+
"loss": 0.7209,
|
| 2928 |
+
"step": 4170
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 2.1359223300970873,
|
| 2932 |
+
"grad_norm": 0.20049767813158514,
|
| 2933 |
+
"learning_rate": 5e-06,
|
| 2934 |
+
"loss": 0.7272,
|
| 2935 |
+
"step": 4180
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 2.1410321921308126,
|
| 2939 |
+
"grad_norm": 0.1964989410548727,
|
| 2940 |
+
"learning_rate": 5e-06,
|
| 2941 |
+
"loss": 0.7205,
|
| 2942 |
+
"step": 4190
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 2.1461420541645375,
|
| 2946 |
+
"grad_norm": 0.18656787454813695,
|
| 2947 |
+
"learning_rate": 5e-06,
|
| 2948 |
+
"loss": 0.73,
|
| 2949 |
+
"step": 4200
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 2.151251916198263,
|
| 2953 |
+
"grad_norm": 0.1869811709706087,
|
| 2954 |
+
"learning_rate": 5e-06,
|
| 2955 |
+
"loss": 0.7211,
|
| 2956 |
+
"step": 4210
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 2.1563617782319877,
|
| 2960 |
+
"grad_norm": 0.1822452717215839,
|
| 2961 |
+
"learning_rate": 5e-06,
|
| 2962 |
+
"loss": 0.7149,
|
| 2963 |
+
"step": 4220
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 2.161471640265713,
|
| 2967 |
+
"grad_norm": 0.18992052776084958,
|
| 2968 |
+
"learning_rate": 5e-06,
|
| 2969 |
+
"loss": 0.7343,
|
| 2970 |
+
"step": 4230
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 2.166581502299438,
|
| 2974 |
+
"grad_norm": 0.18438525645647527,
|
| 2975 |
+
"learning_rate": 5e-06,
|
| 2976 |
+
"loss": 0.7094,
|
| 2977 |
+
"step": 4240
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 2.171691364333163,
|
| 2981 |
+
"grad_norm": 0.19529306952253045,
|
| 2982 |
+
"learning_rate": 5e-06,
|
| 2983 |
+
"loss": 0.7309,
|
| 2984 |
+
"step": 4250
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 2.176801226366888,
|
| 2988 |
+
"grad_norm": 0.19540047885297857,
|
| 2989 |
+
"learning_rate": 5e-06,
|
| 2990 |
+
"loss": 0.7128,
|
| 2991 |
+
"step": 4260
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 2.181911088400613,
|
| 2995 |
+
"grad_norm": 0.18411641667621176,
|
| 2996 |
+
"learning_rate": 5e-06,
|
| 2997 |
+
"loss": 0.7225,
|
| 2998 |
+
"step": 4270
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 2.1870209504343383,
|
| 3002 |
+
"grad_norm": 0.19493918166759708,
|
| 3003 |
+
"learning_rate": 5e-06,
|
| 3004 |
+
"loss": 0.723,
|
| 3005 |
+
"step": 4280
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 2.192130812468063,
|
| 3009 |
+
"grad_norm": 0.1843201705851165,
|
| 3010 |
+
"learning_rate": 5e-06,
|
| 3011 |
+
"loss": 0.7133,
|
| 3012 |
+
"step": 4290
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 2.1972406745017885,
|
| 3016 |
+
"grad_norm": 0.1945434906423164,
|
| 3017 |
+
"learning_rate": 5e-06,
|
| 3018 |
+
"loss": 0.7196,
|
| 3019 |
+
"step": 4300
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 2.2023505365355134,
|
| 3023 |
+
"grad_norm": 0.19008929161664118,
|
| 3024 |
+
"learning_rate": 5e-06,
|
| 3025 |
+
"loss": 0.7224,
|
| 3026 |
+
"step": 4310
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 2.2074603985692387,
|
| 3030 |
+
"grad_norm": 0.18391558057092405,
|
| 3031 |
+
"learning_rate": 5e-06,
|
| 3032 |
+
"loss": 0.723,
|
| 3033 |
+
"step": 4320
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 2.2125702606029636,
|
| 3037 |
+
"grad_norm": 0.19348061164149957,
|
| 3038 |
+
"learning_rate": 5e-06,
|
| 3039 |
+
"loss": 0.7331,
|
| 3040 |
+
"step": 4330
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 2.217680122636689,
|
| 3044 |
+
"grad_norm": 0.18607054173383442,
|
| 3045 |
+
"learning_rate": 5e-06,
|
| 3046 |
+
"loss": 0.7244,
|
| 3047 |
+
"step": 4340
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 2.2227899846704138,
|
| 3051 |
+
"grad_norm": 0.18618780601208676,
|
| 3052 |
+
"learning_rate": 5e-06,
|
| 3053 |
+
"loss": 0.7272,
|
| 3054 |
+
"step": 4350
|
| 3055 |
+
},
|
| 3056 |
+
{
|
| 3057 |
+
"epoch": 2.227899846704139,
|
| 3058 |
+
"grad_norm": 0.17823884880532717,
|
| 3059 |
+
"learning_rate": 5e-06,
|
| 3060 |
+
"loss": 0.7091,
|
| 3061 |
+
"step": 4360
|
| 3062 |
+
},
|
| 3063 |
+
{
|
| 3064 |
+
"epoch": 2.233009708737864,
|
| 3065 |
+
"grad_norm": 0.18561422862650115,
|
| 3066 |
+
"learning_rate": 5e-06,
|
| 3067 |
+
"loss": 0.7205,
|
| 3068 |
+
"step": 4370
|
| 3069 |
+
},
|
| 3070 |
+
{
|
| 3071 |
+
"epoch": 2.2381195707715893,
|
| 3072 |
+
"grad_norm": 0.17953482905796772,
|
| 3073 |
+
"learning_rate": 5e-06,
|
| 3074 |
+
"loss": 0.729,
|
| 3075 |
+
"step": 4380
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 2.243229432805314,
|
| 3079 |
+
"grad_norm": 0.18191363924178688,
|
| 3080 |
+
"learning_rate": 5e-06,
|
| 3081 |
+
"loss": 0.7111,
|
| 3082 |
+
"step": 4390
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 2.2483392948390395,
|
| 3086 |
+
"grad_norm": 0.180435799688224,
|
| 3087 |
+
"learning_rate": 5e-06,
|
| 3088 |
+
"loss": 0.7107,
|
| 3089 |
+
"step": 4400
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 2.2534491568727644,
|
| 3093 |
+
"grad_norm": 0.18960270587770217,
|
| 3094 |
+
"learning_rate": 5e-06,
|
| 3095 |
+
"loss": 0.725,
|
| 3096 |
+
"step": 4410
|
| 3097 |
+
},
|
| 3098 |
+
{
|
| 3099 |
+
"epoch": 2.2585590189064897,
|
| 3100 |
+
"grad_norm": 0.17769968679729267,
|
| 3101 |
+
"learning_rate": 5e-06,
|
| 3102 |
+
"loss": 0.7143,
|
| 3103 |
+
"step": 4420
|
| 3104 |
+
},
|
| 3105 |
+
{
|
| 3106 |
+
"epoch": 2.2636688809402146,
|
| 3107 |
+
"grad_norm": 0.19163278997385685,
|
| 3108 |
+
"learning_rate": 5e-06,
|
| 3109 |
+
"loss": 0.7182,
|
| 3110 |
+
"step": 4430
|
| 3111 |
+
},
|
| 3112 |
+
{
|
| 3113 |
+
"epoch": 2.26877874297394,
|
| 3114 |
+
"grad_norm": 0.1940788858796441,
|
| 3115 |
+
"learning_rate": 5e-06,
|
| 3116 |
+
"loss": 0.7366,
|
| 3117 |
+
"step": 4440
|
| 3118 |
+
},
|
| 3119 |
+
{
|
| 3120 |
+
"epoch": 2.2738886050076648,
|
| 3121 |
+
"grad_norm": 0.1776800674678045,
|
| 3122 |
+
"learning_rate": 5e-06,
|
| 3123 |
+
"loss": 0.7069,
|
| 3124 |
+
"step": 4450
|
| 3125 |
+
},
|
| 3126 |
+
{
|
| 3127 |
+
"epoch": 2.27899846704139,
|
| 3128 |
+
"grad_norm": 0.17843289642160187,
|
| 3129 |
+
"learning_rate": 5e-06,
|
| 3130 |
+
"loss": 0.724,
|
| 3131 |
+
"step": 4460
|
| 3132 |
+
},
|
| 3133 |
+
{
|
| 3134 |
+
"epoch": 2.284108329075115,
|
| 3135 |
+
"grad_norm": 0.18057741833938729,
|
| 3136 |
+
"learning_rate": 5e-06,
|
| 3137 |
+
"loss": 0.7137,
|
| 3138 |
+
"step": 4470
|
| 3139 |
+
},
|
| 3140 |
+
{
|
| 3141 |
+
"epoch": 2.28921819110884,
|
| 3142 |
+
"grad_norm": 0.1866133357180047,
|
| 3143 |
+
"learning_rate": 5e-06,
|
| 3144 |
+
"loss": 0.7166,
|
| 3145 |
+
"step": 4480
|
| 3146 |
+
},
|
| 3147 |
+
{
|
| 3148 |
+
"epoch": 2.294328053142565,
|
| 3149 |
+
"grad_norm": 0.18193348825294622,
|
| 3150 |
+
"learning_rate": 5e-06,
|
| 3151 |
+
"loss": 0.7244,
|
| 3152 |
+
"step": 4490
|
| 3153 |
+
},
|
| 3154 |
+
{
|
| 3155 |
+
"epoch": 2.2994379151762905,
|
| 3156 |
+
"grad_norm": 0.1929190761683958,
|
| 3157 |
+
"learning_rate": 5e-06,
|
| 3158 |
+
"loss": 0.7277,
|
| 3159 |
+
"step": 4500
|
| 3160 |
+
},
|
| 3161 |
+
{
|
| 3162 |
+
"epoch": 2.3045477772100154,
|
| 3163 |
+
"grad_norm": 0.19392177452359835,
|
| 3164 |
+
"learning_rate": 5e-06,
|
| 3165 |
+
"loss": 0.7198,
|
| 3166 |
+
"step": 4510
|
| 3167 |
+
},
|
| 3168 |
+
{
|
| 3169 |
+
"epoch": 2.3096576392437402,
|
| 3170 |
+
"grad_norm": 0.18143577291270357,
|
| 3171 |
+
"learning_rate": 5e-06,
|
| 3172 |
+
"loss": 0.728,
|
| 3173 |
+
"step": 4520
|
| 3174 |
+
},
|
| 3175 |
+
{
|
| 3176 |
+
"epoch": 2.3147675012774656,
|
| 3177 |
+
"grad_norm": 0.19443872804506757,
|
| 3178 |
+
"learning_rate": 5e-06,
|
| 3179 |
+
"loss": 0.7272,
|
| 3180 |
+
"step": 4530
|
| 3181 |
+
},
|
| 3182 |
+
{
|
| 3183 |
+
"epoch": 2.3198773633111904,
|
| 3184 |
+
"grad_norm": 0.1851328945489432,
|
| 3185 |
+
"learning_rate": 5e-06,
|
| 3186 |
+
"loss": 0.7234,
|
| 3187 |
+
"step": 4540
|
| 3188 |
+
},
|
| 3189 |
+
{
|
| 3190 |
+
"epoch": 2.3249872253449158,
|
| 3191 |
+
"grad_norm": 0.19038110214278162,
|
| 3192 |
+
"learning_rate": 5e-06,
|
| 3193 |
+
"loss": 0.713,
|
| 3194 |
+
"step": 4550
|
| 3195 |
+
},
|
| 3196 |
+
{
|
| 3197 |
+
"epoch": 2.3300970873786406,
|
| 3198 |
+
"grad_norm": 0.1874741106018047,
|
| 3199 |
+
"learning_rate": 5e-06,
|
| 3200 |
+
"loss": 0.7161,
|
| 3201 |
+
"step": 4560
|
| 3202 |
+
},
|
| 3203 |
+
{
|
| 3204 |
+
"epoch": 2.335206949412366,
|
| 3205 |
+
"grad_norm": 0.19581977792981697,
|
| 3206 |
+
"learning_rate": 5e-06,
|
| 3207 |
+
"loss": 0.7162,
|
| 3208 |
+
"step": 4570
|
| 3209 |
+
},
|
| 3210 |
+
{
|
| 3211 |
+
"epoch": 2.340316811446091,
|
| 3212 |
+
"grad_norm": 0.17541497625191085,
|
| 3213 |
+
"learning_rate": 5e-06,
|
| 3214 |
+
"loss": 0.723,
|
| 3215 |
+
"step": 4580
|
| 3216 |
+
},
|
| 3217 |
+
{
|
| 3218 |
+
"epoch": 2.345426673479816,
|
| 3219 |
+
"grad_norm": 0.20023862697490177,
|
| 3220 |
+
"learning_rate": 5e-06,
|
| 3221 |
+
"loss": 0.7185,
|
| 3222 |
+
"step": 4590
|
| 3223 |
+
},
|
| 3224 |
+
{
|
| 3225 |
+
"epoch": 2.350536535513541,
|
| 3226 |
+
"grad_norm": 0.1959974413991938,
|
| 3227 |
+
"learning_rate": 5e-06,
|
| 3228 |
+
"loss": 0.7088,
|
| 3229 |
+
"step": 4600
|
| 3230 |
+
},
|
| 3231 |
+
{
|
| 3232 |
+
"epoch": 2.3556463975472663,
|
| 3233 |
+
"grad_norm": 0.18473631827351808,
|
| 3234 |
+
"learning_rate": 5e-06,
|
| 3235 |
+
"loss": 0.719,
|
| 3236 |
+
"step": 4610
|
| 3237 |
+
},
|
| 3238 |
+
{
|
| 3239 |
+
"epoch": 2.3607562595809912,
|
| 3240 |
+
"grad_norm": 0.19338189496040809,
|
| 3241 |
+
"learning_rate": 5e-06,
|
| 3242 |
+
"loss": 0.7239,
|
| 3243 |
+
"step": 4620
|
| 3244 |
+
},
|
| 3245 |
+
{
|
| 3246 |
+
"epoch": 2.3658661216147165,
|
| 3247 |
+
"grad_norm": 0.19066355345819264,
|
| 3248 |
+
"learning_rate": 5e-06,
|
| 3249 |
+
"loss": 0.7266,
|
| 3250 |
+
"step": 4630
|
| 3251 |
+
},
|
| 3252 |
+
{
|
| 3253 |
+
"epoch": 2.3709759836484414,
|
| 3254 |
+
"grad_norm": 0.1883524242314222,
|
| 3255 |
+
"learning_rate": 5e-06,
|
| 3256 |
+
"loss": 0.7186,
|
| 3257 |
+
"step": 4640
|
| 3258 |
+
},
|
| 3259 |
+
{
|
| 3260 |
+
"epoch": 2.3760858456821667,
|
| 3261 |
+
"grad_norm": 0.18502939250064523,
|
| 3262 |
+
"learning_rate": 5e-06,
|
| 3263 |
+
"loss": 0.7199,
|
| 3264 |
+
"step": 4650
|
| 3265 |
+
},
|
| 3266 |
+
{
|
| 3267 |
+
"epoch": 2.3811957077158916,
|
| 3268 |
+
"grad_norm": 0.1792178984338457,
|
| 3269 |
+
"learning_rate": 5e-06,
|
| 3270 |
+
"loss": 0.714,
|
| 3271 |
+
"step": 4660
|
| 3272 |
+
},
|
| 3273 |
+
{
|
| 3274 |
+
"epoch": 2.386305569749617,
|
| 3275 |
+
"grad_norm": 0.18226944439325007,
|
| 3276 |
+
"learning_rate": 5e-06,
|
| 3277 |
+
"loss": 0.7041,
|
| 3278 |
+
"step": 4670
|
| 3279 |
+
},
|
| 3280 |
+
{
|
| 3281 |
+
"epoch": 2.391415431783342,
|
| 3282 |
+
"grad_norm": 0.18015473190902764,
|
| 3283 |
+
"learning_rate": 5e-06,
|
| 3284 |
+
"loss": 0.716,
|
| 3285 |
+
"step": 4680
|
| 3286 |
+
},
|
| 3287 |
+
{
|
| 3288 |
+
"epoch": 2.396525293817067,
|
| 3289 |
+
"grad_norm": 0.18115429600775665,
|
| 3290 |
+
"learning_rate": 5e-06,
|
| 3291 |
+
"loss": 0.7086,
|
| 3292 |
+
"step": 4690
|
| 3293 |
+
},
|
| 3294 |
+
{
|
| 3295 |
+
"epoch": 2.401635155850792,
|
| 3296 |
+
"grad_norm": 0.2043473041616948,
|
| 3297 |
+
"learning_rate": 5e-06,
|
| 3298 |
+
"loss": 0.7165,
|
| 3299 |
+
"step": 4700
|
| 3300 |
+
},
|
| 3301 |
+
{
|
| 3302 |
+
"epoch": 2.406745017884517,
|
| 3303 |
+
"grad_norm": 0.20268566073488395,
|
| 3304 |
+
"learning_rate": 5e-06,
|
| 3305 |
+
"loss": 0.7153,
|
| 3306 |
+
"step": 4710
|
| 3307 |
+
},
|
| 3308 |
+
{
|
| 3309 |
+
"epoch": 2.411854879918242,
|
| 3310 |
+
"grad_norm": 0.19488489934465453,
|
| 3311 |
+
"learning_rate": 5e-06,
|
| 3312 |
+
"loss": 0.7261,
|
| 3313 |
+
"step": 4720
|
| 3314 |
+
},
|
| 3315 |
+
{
|
| 3316 |
+
"epoch": 2.4169647419519675,
|
| 3317 |
+
"grad_norm": 0.1869426176451986,
|
| 3318 |
+
"learning_rate": 5e-06,
|
| 3319 |
+
"loss": 0.6977,
|
| 3320 |
+
"step": 4730
|
| 3321 |
+
},
|
| 3322 |
+
{
|
| 3323 |
+
"epoch": 2.4220746039856924,
|
| 3324 |
+
"grad_norm": 0.19817580694708986,
|
| 3325 |
+
"learning_rate": 5e-06,
|
| 3326 |
+
"loss": 0.71,
|
| 3327 |
+
"step": 4740
|
| 3328 |
+
},
|
| 3329 |
+
{
|
| 3330 |
+
"epoch": 2.4271844660194173,
|
| 3331 |
+
"grad_norm": 0.2099570721696579,
|
| 3332 |
+
"learning_rate": 5e-06,
|
| 3333 |
+
"loss": 0.7204,
|
| 3334 |
+
"step": 4750
|
| 3335 |
+
},
|
| 3336 |
+
{
|
| 3337 |
+
"epoch": 2.4322943280531426,
|
| 3338 |
+
"grad_norm": 0.1964657272176851,
|
| 3339 |
+
"learning_rate": 5e-06,
|
| 3340 |
+
"loss": 0.7104,
|
| 3341 |
+
"step": 4760
|
| 3342 |
+
},
|
| 3343 |
+
{
|
| 3344 |
+
"epoch": 2.4374041900868675,
|
| 3345 |
+
"grad_norm": 0.1886905609835038,
|
| 3346 |
+
"learning_rate": 5e-06,
|
| 3347 |
+
"loss": 0.7085,
|
| 3348 |
+
"step": 4770
|
| 3349 |
+
},
|
| 3350 |
+
{
|
| 3351 |
+
"epoch": 2.442514052120593,
|
| 3352 |
+
"grad_norm": 0.18220683381465547,
|
| 3353 |
+
"learning_rate": 5e-06,
|
| 3354 |
+
"loss": 0.722,
|
| 3355 |
+
"step": 4780
|
| 3356 |
+
},
|
| 3357 |
+
{
|
| 3358 |
+
"epoch": 2.4476239141543177,
|
| 3359 |
+
"grad_norm": 0.18487998134947345,
|
| 3360 |
+
"learning_rate": 5e-06,
|
| 3361 |
+
"loss": 0.7112,
|
| 3362 |
+
"step": 4790
|
| 3363 |
+
},
|
| 3364 |
+
{
|
| 3365 |
+
"epoch": 2.452733776188043,
|
| 3366 |
+
"grad_norm": 0.16830917755794478,
|
| 3367 |
+
"learning_rate": 5e-06,
|
| 3368 |
+
"loss": 0.7157,
|
| 3369 |
+
"step": 4800
|
| 3370 |
+
},
|
| 3371 |
+
{
|
| 3372 |
+
"epoch": 2.457843638221768,
|
| 3373 |
+
"grad_norm": 0.18354244943334422,
|
| 3374 |
+
"learning_rate": 5e-06,
|
| 3375 |
+
"loss": 0.7152,
|
| 3376 |
+
"step": 4810
|
| 3377 |
+
},
|
| 3378 |
+
{
|
| 3379 |
+
"epoch": 2.462953500255493,
|
| 3380 |
+
"grad_norm": 0.18986840496033355,
|
| 3381 |
+
"learning_rate": 5e-06,
|
| 3382 |
+
"loss": 0.7161,
|
| 3383 |
+
"step": 4820
|
| 3384 |
+
},
|
| 3385 |
+
{
|
| 3386 |
+
"epoch": 2.468063362289218,
|
| 3387 |
+
"grad_norm": 0.1826318712937385,
|
| 3388 |
+
"learning_rate": 5e-06,
|
| 3389 |
+
"loss": 0.7295,
|
| 3390 |
+
"step": 4830
|
| 3391 |
+
},
|
| 3392 |
+
{
|
| 3393 |
+
"epoch": 2.4731732243229434,
|
| 3394 |
+
"grad_norm": 0.18527008923688568,
|
| 3395 |
+
"learning_rate": 5e-06,
|
| 3396 |
+
"loss": 0.7115,
|
| 3397 |
+
"step": 4840
|
| 3398 |
+
},
|
| 3399 |
+
{
|
| 3400 |
+
"epoch": 2.4782830863566683,
|
| 3401 |
+
"grad_norm": 0.18224411387144135,
|
| 3402 |
+
"learning_rate": 5e-06,
|
| 3403 |
+
"loss": 0.7173,
|
| 3404 |
+
"step": 4850
|
| 3405 |
+
},
|
| 3406 |
+
{
|
| 3407 |
+
"epoch": 2.4833929483903936,
|
| 3408 |
+
"grad_norm": 0.17811035407666295,
|
| 3409 |
+
"learning_rate": 5e-06,
|
| 3410 |
+
"loss": 0.7095,
|
| 3411 |
+
"step": 4860
|
| 3412 |
+
},
|
| 3413 |
+
{
|
| 3414 |
+
"epoch": 2.4885028104241185,
|
| 3415 |
+
"grad_norm": 0.18261169308001168,
|
| 3416 |
+
"learning_rate": 5e-06,
|
| 3417 |
+
"loss": 0.7162,
|
| 3418 |
+
"step": 4870
|
| 3419 |
+
},
|
| 3420 |
+
{
|
| 3421 |
+
"epoch": 2.493612672457844,
|
| 3422 |
+
"grad_norm": 0.2002887514635378,
|
| 3423 |
+
"learning_rate": 5e-06,
|
| 3424 |
+
"loss": 0.7208,
|
| 3425 |
+
"step": 4880
|
| 3426 |
+
},
|
| 3427 |
+
{
|
| 3428 |
+
"epoch": 2.4987225344915687,
|
| 3429 |
+
"grad_norm": 0.1869680657646229,
|
| 3430 |
+
"learning_rate": 5e-06,
|
| 3431 |
+
"loss": 0.7081,
|
| 3432 |
+
"step": 4890
|
| 3433 |
+
},
|
| 3434 |
+
{
|
| 3435 |
+
"epoch": 2.5038323965252935,
|
| 3436 |
+
"grad_norm": 0.18675570886820952,
|
| 3437 |
+
"learning_rate": 5e-06,
|
| 3438 |
+
"loss": 0.7117,
|
| 3439 |
+
"step": 4900
|
| 3440 |
+
},
|
| 3441 |
+
{
|
| 3442 |
+
"epoch": 2.508942258559019,
|
| 3443 |
+
"grad_norm": 0.19944201285579877,
|
| 3444 |
+
"learning_rate": 5e-06,
|
| 3445 |
+
"loss": 0.7198,
|
| 3446 |
+
"step": 4910
|
| 3447 |
+
},
|
| 3448 |
+
{
|
| 3449 |
+
"epoch": 2.514052120592744,
|
| 3450 |
+
"grad_norm": 0.18956975903685294,
|
| 3451 |
+
"learning_rate": 5e-06,
|
| 3452 |
+
"loss": 0.7179,
|
| 3453 |
+
"step": 4920
|
| 3454 |
+
},
|
| 3455 |
+
{
|
| 3456 |
+
"epoch": 2.519161982626469,
|
| 3457 |
+
"grad_norm": 0.1942120087148544,
|
| 3458 |
+
"learning_rate": 5e-06,
|
| 3459 |
+
"loss": 0.7216,
|
| 3460 |
+
"step": 4930
|
| 3461 |
+
},
|
| 3462 |
+
{
|
| 3463 |
+
"epoch": 2.524271844660194,
|
| 3464 |
+
"grad_norm": 0.18623164151592378,
|
| 3465 |
+
"learning_rate": 5e-06,
|
| 3466 |
+
"loss": 0.6997,
|
| 3467 |
+
"step": 4940
|
| 3468 |
+
},
|
| 3469 |
+
{
|
| 3470 |
+
"epoch": 2.5293817066939193,
|
| 3471 |
+
"grad_norm": 0.19081196748322382,
|
| 3472 |
+
"learning_rate": 5e-06,
|
| 3473 |
+
"loss": 0.7091,
|
| 3474 |
+
"step": 4950
|
| 3475 |
+
},
|
| 3476 |
+
{
|
| 3477 |
+
"epoch": 2.5344915687276446,
|
| 3478 |
+
"grad_norm": 0.191889152108264,
|
| 3479 |
+
"learning_rate": 5e-06,
|
| 3480 |
+
"loss": 0.7122,
|
| 3481 |
+
"step": 4960
|
| 3482 |
+
},
|
| 3483 |
+
{
|
| 3484 |
+
"epoch": 2.5396014307613695,
|
| 3485 |
+
"grad_norm": 0.19018109192168367,
|
| 3486 |
+
"learning_rate": 5e-06,
|
| 3487 |
+
"loss": 0.7195,
|
| 3488 |
+
"step": 4970
|
| 3489 |
+
},
|
| 3490 |
+
{
|
| 3491 |
+
"epoch": 2.5447112927950943,
|
| 3492 |
+
"grad_norm": 0.18029975730571046,
|
| 3493 |
+
"learning_rate": 5e-06,
|
| 3494 |
+
"loss": 0.7067,
|
| 3495 |
+
"step": 4980
|
| 3496 |
+
},
|
| 3497 |
+
{
|
| 3498 |
+
"epoch": 2.5498211548288197,
|
| 3499 |
+
"grad_norm": 0.19081905131548996,
|
| 3500 |
+
"learning_rate": 5e-06,
|
| 3501 |
+
"loss": 0.7184,
|
| 3502 |
+
"step": 4990
|
| 3503 |
+
},
|
| 3504 |
+
{
|
| 3505 |
+
"epoch": 2.554931016862545,
|
| 3506 |
+
"grad_norm": 0.18184068561404432,
|
| 3507 |
+
"learning_rate": 5e-06,
|
| 3508 |
+
"loss": 0.711,
|
| 3509 |
+
"step": 5000
|
| 3510 |
+
},
|
| 3511 |
+
{
|
| 3512 |
+
"epoch": 2.56004087889627,
|
| 3513 |
+
"grad_norm": 0.18141265501857734,
|
| 3514 |
+
"learning_rate": 5e-06,
|
| 3515 |
+
"loss": 0.7298,
|
| 3516 |
+
"step": 5010
|
| 3517 |
+
},
|
| 3518 |
+
{
|
| 3519 |
+
"epoch": 2.5651507409299947,
|
| 3520 |
+
"grad_norm": 0.19059042586600472,
|
| 3521 |
+
"learning_rate": 5e-06,
|
| 3522 |
+
"loss": 0.7114,
|
| 3523 |
+
"step": 5020
|
| 3524 |
+
},
|
| 3525 |
+
{
|
| 3526 |
+
"epoch": 2.57026060296372,
|
| 3527 |
+
"grad_norm": 0.18752747662093455,
|
| 3528 |
+
"learning_rate": 5e-06,
|
| 3529 |
+
"loss": 0.7139,
|
| 3530 |
+
"step": 5030
|
| 3531 |
+
},
|
| 3532 |
+
{
|
| 3533 |
+
"epoch": 2.575370464997445,
|
| 3534 |
+
"grad_norm": 0.18531393933159526,
|
| 3535 |
+
"learning_rate": 5e-06,
|
| 3536 |
+
"loss": 0.7152,
|
| 3537 |
+
"step": 5040
|
| 3538 |
+
},
|
| 3539 |
+
{
|
| 3540 |
+
"epoch": 2.5804803270311703,
|
| 3541 |
+
"grad_norm": 0.2020545252464496,
|
| 3542 |
+
"learning_rate": 5e-06,
|
| 3543 |
+
"loss": 0.724,
|
| 3544 |
+
"step": 5050
|
| 3545 |
+
},
|
| 3546 |
+
{
|
| 3547 |
+
"epoch": 2.585590189064895,
|
| 3548 |
+
"grad_norm": 0.1817661308445167,
|
| 3549 |
+
"learning_rate": 5e-06,
|
| 3550 |
+
"loss": 0.7099,
|
| 3551 |
+
"step": 5060
|
| 3552 |
+
},
|
| 3553 |
+
{
|
| 3554 |
+
"epoch": 2.5907000510986204,
|
| 3555 |
+
"grad_norm": 0.1874710204305083,
|
| 3556 |
+
"learning_rate": 5e-06,
|
| 3557 |
+
"loss": 0.727,
|
| 3558 |
+
"step": 5070
|
| 3559 |
+
},
|
| 3560 |
+
{
|
| 3561 |
+
"epoch": 2.5958099131323453,
|
| 3562 |
+
"grad_norm": 0.19697140144619885,
|
| 3563 |
+
"learning_rate": 5e-06,
|
| 3564 |
+
"loss": 0.7184,
|
| 3565 |
+
"step": 5080
|
| 3566 |
+
},
|
| 3567 |
+
{
|
| 3568 |
+
"epoch": 2.6009197751660706,
|
| 3569 |
+
"grad_norm": 0.1889733854232041,
|
| 3570 |
+
"learning_rate": 5e-06,
|
| 3571 |
+
"loss": 0.7169,
|
| 3572 |
+
"step": 5090
|
| 3573 |
+
},
|
| 3574 |
+
{
|
| 3575 |
+
"epoch": 2.6060296371997955,
|
| 3576 |
+
"grad_norm": 0.20000474912796498,
|
| 3577 |
+
"learning_rate": 5e-06,
|
| 3578 |
+
"loss": 0.7089,
|
| 3579 |
+
"step": 5100
|
| 3580 |
+
},
|
| 3581 |
+
{
|
| 3582 |
+
"epoch": 2.611139499233521,
|
| 3583 |
+
"grad_norm": 0.18020813610110156,
|
| 3584 |
+
"learning_rate": 5e-06,
|
| 3585 |
+
"loss": 0.7264,
|
| 3586 |
+
"step": 5110
|
| 3587 |
+
},
|
| 3588 |
+
{
|
| 3589 |
+
"epoch": 2.6162493612672457,
|
| 3590 |
+
"grad_norm": 0.17876620180082428,
|
| 3591 |
+
"learning_rate": 5e-06,
|
| 3592 |
+
"loss": 0.7007,
|
| 3593 |
+
"step": 5120
|
| 3594 |
+
},
|
| 3595 |
+
{
|
| 3596 |
+
"epoch": 2.6213592233009706,
|
| 3597 |
+
"grad_norm": 0.186937820916383,
|
| 3598 |
+
"learning_rate": 5e-06,
|
| 3599 |
+
"loss": 0.7219,
|
| 3600 |
+
"step": 5130
|
| 3601 |
+
},
|
| 3602 |
+
{
|
| 3603 |
+
"epoch": 2.626469085334696,
|
| 3604 |
+
"grad_norm": 0.19293145357331443,
|
| 3605 |
+
"learning_rate": 5e-06,
|
| 3606 |
+
"loss": 0.7116,
|
| 3607 |
+
"step": 5140
|
| 3608 |
+
},
|
| 3609 |
+
{
|
| 3610 |
+
"epoch": 2.6315789473684212,
|
| 3611 |
+
"grad_norm": 0.18779972078705487,
|
| 3612 |
+
"learning_rate": 5e-06,
|
| 3613 |
+
"loss": 0.7147,
|
| 3614 |
+
"step": 5150
|
| 3615 |
+
},
|
| 3616 |
+
{
|
| 3617 |
+
"epoch": 2.636688809402146,
|
| 3618 |
+
"grad_norm": 0.2004320087337195,
|
| 3619 |
+
"learning_rate": 5e-06,
|
| 3620 |
+
"loss": 0.7046,
|
| 3621 |
+
"step": 5160
|
| 3622 |
+
},
|
| 3623 |
+
{
|
| 3624 |
+
"epoch": 2.641798671435871,
|
| 3625 |
+
"grad_norm": 0.18155864298582336,
|
| 3626 |
+
"learning_rate": 5e-06,
|
| 3627 |
+
"loss": 0.7024,
|
| 3628 |
+
"step": 5170
|
| 3629 |
+
},
|
| 3630 |
+
{
|
| 3631 |
+
"epoch": 2.6469085334695963,
|
| 3632 |
+
"grad_norm": 0.18766505517066065,
|
| 3633 |
+
"learning_rate": 5e-06,
|
| 3634 |
+
"loss": 0.7092,
|
| 3635 |
+
"step": 5180
|
| 3636 |
+
},
|
| 3637 |
+
{
|
| 3638 |
+
"epoch": 2.6520183955033216,
|
| 3639 |
+
"grad_norm": 0.18241808648535346,
|
| 3640 |
+
"learning_rate": 5e-06,
|
| 3641 |
+
"loss": 0.7231,
|
| 3642 |
+
"step": 5190
|
| 3643 |
+
},
|
| 3644 |
+
{
|
| 3645 |
+
"epoch": 2.6571282575370465,
|
| 3646 |
+
"grad_norm": 0.1956290063678602,
|
| 3647 |
+
"learning_rate": 5e-06,
|
| 3648 |
+
"loss": 0.7071,
|
| 3649 |
+
"step": 5200
|
| 3650 |
+
},
|
| 3651 |
+
{
|
| 3652 |
+
"epoch": 2.6622381195707714,
|
| 3653 |
+
"grad_norm": 0.19687281837098963,
|
| 3654 |
+
"learning_rate": 5e-06,
|
| 3655 |
+
"loss": 0.7237,
|
| 3656 |
+
"step": 5210
|
| 3657 |
+
},
|
| 3658 |
+
{
|
| 3659 |
+
"epoch": 2.6673479816044967,
|
| 3660 |
+
"grad_norm": 0.18253071868952309,
|
| 3661 |
+
"learning_rate": 5e-06,
|
| 3662 |
+
"loss": 0.7085,
|
| 3663 |
+
"step": 5220
|
| 3664 |
+
},
|
| 3665 |
+
{
|
| 3666 |
+
"epoch": 2.672457843638222,
|
| 3667 |
+
"grad_norm": 0.18783632566318903,
|
| 3668 |
+
"learning_rate": 5e-06,
|
| 3669 |
+
"loss": 0.717,
|
| 3670 |
+
"step": 5230
|
| 3671 |
+
},
|
| 3672 |
+
{
|
| 3673 |
+
"epoch": 2.677567705671947,
|
| 3674 |
+
"grad_norm": 0.1856712541693123,
|
| 3675 |
+
"learning_rate": 5e-06,
|
| 3676 |
+
"loss": 0.7121,
|
| 3677 |
+
"step": 5240
|
| 3678 |
+
},
|
| 3679 |
+
{
|
| 3680 |
+
"epoch": 2.682677567705672,
|
| 3681 |
+
"grad_norm": 0.17913630024093868,
|
| 3682 |
+
"learning_rate": 5e-06,
|
| 3683 |
+
"loss": 0.7141,
|
| 3684 |
+
"step": 5250
|
| 3685 |
+
},
|
| 3686 |
+
{
|
| 3687 |
+
"epoch": 2.687787429739397,
|
| 3688 |
+
"grad_norm": 0.1814607328558563,
|
| 3689 |
+
"learning_rate": 5e-06,
|
| 3690 |
+
"loss": 0.7143,
|
| 3691 |
+
"step": 5260
|
| 3692 |
+
},
|
| 3693 |
+
{
|
| 3694 |
+
"epoch": 2.692897291773122,
|
| 3695 |
+
"grad_norm": 0.20516860190923697,
|
| 3696 |
+
"learning_rate": 5e-06,
|
| 3697 |
+
"loss": 0.7153,
|
| 3698 |
+
"step": 5270
|
| 3699 |
+
},
|
| 3700 |
+
{
|
| 3701 |
+
"epoch": 2.6980071538068473,
|
| 3702 |
+
"grad_norm": 0.18691014833866346,
|
| 3703 |
+
"learning_rate": 5e-06,
|
| 3704 |
+
"loss": 0.7021,
|
| 3705 |
+
"step": 5280
|
| 3706 |
+
},
|
| 3707 |
+
{
|
| 3708 |
+
"epoch": 2.703117015840572,
|
| 3709 |
+
"grad_norm": 0.19243627032023453,
|
| 3710 |
+
"learning_rate": 5e-06,
|
| 3711 |
+
"loss": 0.7198,
|
| 3712 |
+
"step": 5290
|
| 3713 |
+
},
|
| 3714 |
+
{
|
| 3715 |
+
"epoch": 2.7082268778742975,
|
| 3716 |
+
"grad_norm": 0.1782810408332171,
|
| 3717 |
+
"learning_rate": 5e-06,
|
| 3718 |
+
"loss": 0.7144,
|
| 3719 |
+
"step": 5300
|
| 3720 |
+
},
|
| 3721 |
+
{
|
| 3722 |
+
"epoch": 2.7133367399080224,
|
| 3723 |
+
"grad_norm": 0.18246834780709842,
|
| 3724 |
+
"learning_rate": 5e-06,
|
| 3725 |
+
"loss": 0.7073,
|
| 3726 |
+
"step": 5310
|
| 3727 |
+
},
|
| 3728 |
+
{
|
| 3729 |
+
"epoch": 2.7184466019417477,
|
| 3730 |
+
"grad_norm": 0.19188458416675994,
|
| 3731 |
+
"learning_rate": 5e-06,
|
| 3732 |
+
"loss": 0.7033,
|
| 3733 |
+
"step": 5320
|
| 3734 |
+
},
|
| 3735 |
+
{
|
| 3736 |
+
"epoch": 2.7235564639754726,
|
| 3737 |
+
"grad_norm": 0.19431224425803312,
|
| 3738 |
+
"learning_rate": 5e-06,
|
| 3739 |
+
"loss": 0.7121,
|
| 3740 |
+
"step": 5330
|
| 3741 |
+
},
|
| 3742 |
+
{
|
| 3743 |
+
"epoch": 2.728666326009198,
|
| 3744 |
+
"grad_norm": 0.19472680477435397,
|
| 3745 |
+
"learning_rate": 5e-06,
|
| 3746 |
+
"loss": 0.7204,
|
| 3747 |
+
"step": 5340
|
| 3748 |
+
},
|
| 3749 |
+
{
|
| 3750 |
+
"epoch": 2.7337761880429228,
|
| 3751 |
+
"grad_norm": 0.17416333272267362,
|
| 3752 |
+
"learning_rate": 5e-06,
|
| 3753 |
+
"loss": 0.722,
|
| 3754 |
+
"step": 5350
|
| 3755 |
+
},
|
| 3756 |
+
{
|
| 3757 |
+
"epoch": 2.738886050076648,
|
| 3758 |
+
"grad_norm": 0.17997245175531543,
|
| 3759 |
+
"learning_rate": 5e-06,
|
| 3760 |
+
"loss": 0.7209,
|
| 3761 |
+
"step": 5360
|
| 3762 |
+
},
|
| 3763 |
+
{
|
| 3764 |
+
"epoch": 2.743995912110373,
|
| 3765 |
+
"grad_norm": 0.1818254672253798,
|
| 3766 |
+
"learning_rate": 5e-06,
|
| 3767 |
+
"loss": 0.7064,
|
| 3768 |
+
"step": 5370
|
| 3769 |
+
},
|
| 3770 |
+
{
|
| 3771 |
+
"epoch": 2.7491057741440983,
|
| 3772 |
+
"grad_norm": 0.19143052614657435,
|
| 3773 |
+
"learning_rate": 5e-06,
|
| 3774 |
+
"loss": 0.7251,
|
| 3775 |
+
"step": 5380
|
| 3776 |
+
},
|
| 3777 |
+
{
|
| 3778 |
+
"epoch": 2.754215636177823,
|
| 3779 |
+
"grad_norm": 0.1849449249230007,
|
| 3780 |
+
"learning_rate": 5e-06,
|
| 3781 |
+
"loss": 0.7134,
|
| 3782 |
+
"step": 5390
|
| 3783 |
+
},
|
| 3784 |
+
{
|
| 3785 |
+
"epoch": 2.759325498211548,
|
| 3786 |
+
"grad_norm": 0.18960800644635575,
|
| 3787 |
+
"learning_rate": 5e-06,
|
| 3788 |
+
"loss": 0.7132,
|
| 3789 |
+
"step": 5400
|
| 3790 |
+
},
|
| 3791 |
+
{
|
| 3792 |
+
"epoch": 2.7644353602452734,
|
| 3793 |
+
"grad_norm": 0.18450711551330742,
|
| 3794 |
+
"learning_rate": 5e-06,
|
| 3795 |
+
"loss": 0.7218,
|
| 3796 |
+
"step": 5410
|
| 3797 |
+
},
|
| 3798 |
+
{
|
| 3799 |
+
"epoch": 2.7695452222789987,
|
| 3800 |
+
"grad_norm": 0.1775798100347251,
|
| 3801 |
+
"learning_rate": 5e-06,
|
| 3802 |
+
"loss": 0.7246,
|
| 3803 |
+
"step": 5420
|
| 3804 |
+
},
|
| 3805 |
+
{
|
| 3806 |
+
"epoch": 2.7746550843127236,
|
| 3807 |
+
"grad_norm": 0.19385869953809406,
|
| 3808 |
+
"learning_rate": 5e-06,
|
| 3809 |
+
"loss": 0.7154,
|
| 3810 |
+
"step": 5430
|
| 3811 |
+
},
|
| 3812 |
+
{
|
| 3813 |
+
"epoch": 2.7797649463464484,
|
| 3814 |
+
"grad_norm": 0.18947315882216123,
|
| 3815 |
+
"learning_rate": 5e-06,
|
| 3816 |
+
"loss": 0.7146,
|
| 3817 |
+
"step": 5440
|
| 3818 |
+
},
|
| 3819 |
+
{
|
| 3820 |
+
"epoch": 2.7848748083801738,
|
| 3821 |
+
"grad_norm": 0.19874490028355687,
|
| 3822 |
+
"learning_rate": 5e-06,
|
| 3823 |
+
"loss": 0.7098,
|
| 3824 |
+
"step": 5450
|
| 3825 |
+
},
|
| 3826 |
+
{
|
| 3827 |
+
"epoch": 2.789984670413899,
|
| 3828 |
+
"grad_norm": 0.19216650910265412,
|
| 3829 |
+
"learning_rate": 5e-06,
|
| 3830 |
+
"loss": 0.7157,
|
| 3831 |
+
"step": 5460
|
| 3832 |
+
},
|
| 3833 |
+
{
|
| 3834 |
+
"epoch": 2.795094532447624,
|
| 3835 |
+
"grad_norm": 0.1884532290700664,
|
| 3836 |
+
"learning_rate": 5e-06,
|
| 3837 |
+
"loss": 0.73,
|
| 3838 |
+
"step": 5470
|
| 3839 |
+
},
|
| 3840 |
+
{
|
| 3841 |
+
"epoch": 2.800204394481349,
|
| 3842 |
+
"grad_norm": 0.178944069456414,
|
| 3843 |
+
"learning_rate": 5e-06,
|
| 3844 |
+
"loss": 0.7023,
|
| 3845 |
+
"step": 5480
|
| 3846 |
+
},
|
| 3847 |
+
{
|
| 3848 |
+
"epoch": 2.805314256515074,
|
| 3849 |
+
"grad_norm": 0.19897724638804584,
|
| 3850 |
+
"learning_rate": 5e-06,
|
| 3851 |
+
"loss": 0.7266,
|
| 3852 |
+
"step": 5490
|
| 3853 |
+
},
|
| 3854 |
+
{
|
| 3855 |
+
"epoch": 2.810424118548799,
|
| 3856 |
+
"grad_norm": 0.19676207860709408,
|
| 3857 |
+
"learning_rate": 5e-06,
|
| 3858 |
+
"loss": 0.7042,
|
| 3859 |
+
"step": 5500
|
| 3860 |
+
},
|
| 3861 |
+
{
|
| 3862 |
+
"epoch": 2.8155339805825244,
|
| 3863 |
+
"grad_norm": 0.17311366117436,
|
| 3864 |
+
"learning_rate": 5e-06,
|
| 3865 |
+
"loss": 0.7059,
|
| 3866 |
+
"step": 5510
|
| 3867 |
+
},
|
| 3868 |
+
{
|
| 3869 |
+
"epoch": 2.8206438426162492,
|
| 3870 |
+
"grad_norm": 0.1862348663291006,
|
| 3871 |
+
"learning_rate": 5e-06,
|
| 3872 |
+
"loss": 0.7012,
|
| 3873 |
+
"step": 5520
|
| 3874 |
+
},
|
| 3875 |
+
{
|
| 3876 |
+
"epoch": 2.8257537046499746,
|
| 3877 |
+
"grad_norm": 0.19532852701482903,
|
| 3878 |
+
"learning_rate": 5e-06,
|
| 3879 |
+
"loss": 0.7105,
|
| 3880 |
+
"step": 5530
|
| 3881 |
+
},
|
| 3882 |
+
{
|
| 3883 |
+
"epoch": 2.8308635666836994,
|
| 3884 |
+
"grad_norm": 0.18410412264007187,
|
| 3885 |
+
"learning_rate": 5e-06,
|
| 3886 |
+
"loss": 0.7163,
|
| 3887 |
+
"step": 5540
|
| 3888 |
+
},
|
| 3889 |
+
{
|
| 3890 |
+
"epoch": 2.8359734287174247,
|
| 3891 |
+
"grad_norm": 0.21025902958863738,
|
| 3892 |
+
"learning_rate": 5e-06,
|
| 3893 |
+
"loss": 0.7359,
|
| 3894 |
+
"step": 5550
|
| 3895 |
+
},
|
| 3896 |
+
{
|
| 3897 |
+
"epoch": 2.8410832907511496,
|
| 3898 |
+
"grad_norm": 0.17818540055082727,
|
| 3899 |
+
"learning_rate": 5e-06,
|
| 3900 |
+
"loss": 0.7111,
|
| 3901 |
+
"step": 5560
|
| 3902 |
+
},
|
| 3903 |
+
{
|
| 3904 |
+
"epoch": 2.846193152784875,
|
| 3905 |
+
"grad_norm": 0.1842948913149892,
|
| 3906 |
+
"learning_rate": 5e-06,
|
| 3907 |
+
"loss": 0.7167,
|
| 3908 |
+
"step": 5570
|
| 3909 |
+
},
|
| 3910 |
+
{
|
| 3911 |
+
"epoch": 2.8513030148186,
|
| 3912 |
+
"grad_norm": 0.18686037650976978,
|
| 3913 |
+
"learning_rate": 5e-06,
|
| 3914 |
+
"loss": 0.7138,
|
| 3915 |
+
"step": 5580
|
| 3916 |
+
},
|
| 3917 |
+
{
|
| 3918 |
+
"epoch": 2.856412876852325,
|
| 3919 |
+
"grad_norm": 0.203403946822995,
|
| 3920 |
+
"learning_rate": 5e-06,
|
| 3921 |
+
"loss": 0.6996,
|
| 3922 |
+
"step": 5590
|
| 3923 |
+
},
|
| 3924 |
+
{
|
| 3925 |
+
"epoch": 2.86152273888605,
|
| 3926 |
+
"grad_norm": 0.19304193259535427,
|
| 3927 |
+
"learning_rate": 5e-06,
|
| 3928 |
+
"loss": 0.7184,
|
| 3929 |
+
"step": 5600
|
| 3930 |
+
},
|
| 3931 |
+
{
|
| 3932 |
+
"epoch": 2.8666326009197753,
|
| 3933 |
+
"grad_norm": 0.17636221017029402,
|
| 3934 |
+
"learning_rate": 5e-06,
|
| 3935 |
+
"loss": 0.6989,
|
| 3936 |
+
"step": 5610
|
| 3937 |
+
},
|
| 3938 |
+
{
|
| 3939 |
+
"epoch": 2.8717424629535,
|
| 3940 |
+
"grad_norm": 0.2037376472703345,
|
| 3941 |
+
"learning_rate": 5e-06,
|
| 3942 |
+
"loss": 0.7179,
|
| 3943 |
+
"step": 5620
|
| 3944 |
+
},
|
| 3945 |
+
{
|
| 3946 |
+
"epoch": 2.876852324987225,
|
| 3947 |
+
"grad_norm": 0.19200332931870842,
|
| 3948 |
+
"learning_rate": 5e-06,
|
| 3949 |
+
"loss": 0.7139,
|
| 3950 |
+
"step": 5630
|
| 3951 |
+
},
|
| 3952 |
+
{
|
| 3953 |
+
"epoch": 2.8819621870209504,
|
| 3954 |
+
"grad_norm": 0.18491595431761487,
|
| 3955 |
+
"learning_rate": 5e-06,
|
| 3956 |
+
"loss": 0.7128,
|
| 3957 |
+
"step": 5640
|
| 3958 |
+
},
|
| 3959 |
+
{
|
| 3960 |
+
"epoch": 2.8870720490546757,
|
| 3961 |
+
"grad_norm": 0.1838785582117868,
|
| 3962 |
+
"learning_rate": 5e-06,
|
| 3963 |
+
"loss": 0.6955,
|
| 3964 |
+
"step": 5650
|
| 3965 |
+
},
|
| 3966 |
+
{
|
| 3967 |
+
"epoch": 2.8921819110884006,
|
| 3968 |
+
"grad_norm": 0.19428058651152774,
|
| 3969 |
+
"learning_rate": 5e-06,
|
| 3970 |
+
"loss": 0.7102,
|
| 3971 |
+
"step": 5660
|
| 3972 |
+
},
|
| 3973 |
+
{
|
| 3974 |
+
"epoch": 2.8972917731221255,
|
| 3975 |
+
"grad_norm": 0.17936626194327465,
|
| 3976 |
+
"learning_rate": 5e-06,
|
| 3977 |
+
"loss": 0.7078,
|
| 3978 |
+
"step": 5670
|
| 3979 |
+
},
|
| 3980 |
+
{
|
| 3981 |
+
"epoch": 2.902401635155851,
|
| 3982 |
+
"grad_norm": 0.17631468596491473,
|
| 3983 |
+
"learning_rate": 5e-06,
|
| 3984 |
+
"loss": 0.7162,
|
| 3985 |
+
"step": 5680
|
| 3986 |
+
},
|
| 3987 |
+
{
|
| 3988 |
+
"epoch": 2.907511497189576,
|
| 3989 |
+
"grad_norm": 0.18415745158446212,
|
| 3990 |
+
"learning_rate": 5e-06,
|
| 3991 |
+
"loss": 0.7233,
|
| 3992 |
+
"step": 5690
|
| 3993 |
+
},
|
| 3994 |
+
{
|
| 3995 |
+
"epoch": 2.912621359223301,
|
| 3996 |
+
"grad_norm": 0.17279113129182289,
|
| 3997 |
+
"learning_rate": 5e-06,
|
| 3998 |
+
"loss": 0.6992,
|
| 3999 |
+
"step": 5700
|
| 4000 |
+
},
|
| 4001 |
+
{
|
| 4002 |
+
"epoch": 2.917731221257026,
|
| 4003 |
+
"grad_norm": 0.19005862366918605,
|
| 4004 |
+
"learning_rate": 5e-06,
|
| 4005 |
+
"loss": 0.7015,
|
| 4006 |
+
"step": 5710
|
| 4007 |
+
},
|
| 4008 |
+
{
|
| 4009 |
+
"epoch": 2.922841083290751,
|
| 4010 |
+
"grad_norm": 0.1865410089900959,
|
| 4011 |
+
"learning_rate": 5e-06,
|
| 4012 |
+
"loss": 0.7043,
|
| 4013 |
+
"step": 5720
|
| 4014 |
+
},
|
| 4015 |
+
{
|
| 4016 |
+
"epoch": 2.927950945324476,
|
| 4017 |
+
"grad_norm": 0.19166084388796084,
|
| 4018 |
+
"learning_rate": 5e-06,
|
| 4019 |
+
"loss": 0.7212,
|
| 4020 |
+
"step": 5730
|
| 4021 |
+
},
|
| 4022 |
+
{
|
| 4023 |
+
"epoch": 2.9330608073582014,
|
| 4024 |
+
"grad_norm": 0.19183886093470404,
|
| 4025 |
+
"learning_rate": 5e-06,
|
| 4026 |
+
"loss": 0.7117,
|
| 4027 |
+
"step": 5740
|
| 4028 |
+
},
|
| 4029 |
+
{
|
| 4030 |
+
"epoch": 2.9381706693919263,
|
| 4031 |
+
"grad_norm": 0.18542747500934698,
|
| 4032 |
+
"learning_rate": 5e-06,
|
| 4033 |
+
"loss": 0.7031,
|
| 4034 |
+
"step": 5750
|
| 4035 |
+
},
|
| 4036 |
+
{
|
| 4037 |
+
"epoch": 2.9432805314256516,
|
| 4038 |
+
"grad_norm": 0.19800100325472195,
|
| 4039 |
+
"learning_rate": 5e-06,
|
| 4040 |
+
"loss": 0.7003,
|
| 4041 |
+
"step": 5760
|
| 4042 |
+
},
|
| 4043 |
+
{
|
| 4044 |
+
"epoch": 2.9483903934593765,
|
| 4045 |
+
"grad_norm": 0.19068782203027565,
|
| 4046 |
+
"learning_rate": 5e-06,
|
| 4047 |
+
"loss": 0.7157,
|
| 4048 |
+
"step": 5770
|
| 4049 |
+
},
|
| 4050 |
+
{
|
| 4051 |
+
"epoch": 2.953500255493102,
|
| 4052 |
+
"grad_norm": 0.1893011994562663,
|
| 4053 |
+
"learning_rate": 5e-06,
|
| 4054 |
+
"loss": 0.7049,
|
| 4055 |
+
"step": 5780
|
| 4056 |
+
},
|
| 4057 |
+
{
|
| 4058 |
+
"epoch": 2.9586101175268267,
|
| 4059 |
+
"grad_norm": 0.17894632421886825,
|
| 4060 |
+
"learning_rate": 5e-06,
|
| 4061 |
+
"loss": 0.721,
|
| 4062 |
+
"step": 5790
|
| 4063 |
+
},
|
| 4064 |
+
{
|
| 4065 |
+
"epoch": 2.963719979560552,
|
| 4066 |
+
"grad_norm": 0.18749644947518349,
|
| 4067 |
+
"learning_rate": 5e-06,
|
| 4068 |
+
"loss": 0.7245,
|
| 4069 |
+
"step": 5800
|
| 4070 |
+
},
|
| 4071 |
+
{
|
| 4072 |
+
"epoch": 2.968829841594277,
|
| 4073 |
+
"grad_norm": 0.18920042895955116,
|
| 4074 |
+
"learning_rate": 5e-06,
|
| 4075 |
+
"loss": 0.7086,
|
| 4076 |
+
"step": 5810
|
| 4077 |
+
},
|
| 4078 |
+
{
|
| 4079 |
+
"epoch": 2.973939703628002,
|
| 4080 |
+
"grad_norm": 0.18597523756708192,
|
| 4081 |
+
"learning_rate": 5e-06,
|
| 4082 |
+
"loss": 0.7143,
|
| 4083 |
+
"step": 5820
|
| 4084 |
+
},
|
| 4085 |
+
{
|
| 4086 |
+
"epoch": 2.979049565661727,
|
| 4087 |
+
"grad_norm": 0.18741696462849414,
|
| 4088 |
+
"learning_rate": 5e-06,
|
| 4089 |
+
"loss": 0.7195,
|
| 4090 |
+
"step": 5830
|
| 4091 |
+
},
|
| 4092 |
+
{
|
| 4093 |
+
"epoch": 2.9841594276954524,
|
| 4094 |
+
"grad_norm": 0.18985455877786367,
|
| 4095 |
+
"learning_rate": 5e-06,
|
| 4096 |
+
"loss": 0.7091,
|
| 4097 |
+
"step": 5840
|
| 4098 |
+
},
|
| 4099 |
+
{
|
| 4100 |
+
"epoch": 2.9892692897291773,
|
| 4101 |
+
"grad_norm": 0.18003618910403507,
|
| 4102 |
+
"learning_rate": 5e-06,
|
| 4103 |
+
"loss": 0.7111,
|
| 4104 |
+
"step": 5850
|
| 4105 |
+
},
|
| 4106 |
+
{
|
| 4107 |
+
"epoch": 2.994379151762902,
|
| 4108 |
+
"grad_norm": 0.19511416723954572,
|
| 4109 |
+
"learning_rate": 5e-06,
|
| 4110 |
+
"loss": 0.7105,
|
| 4111 |
+
"step": 5860
|
| 4112 |
+
},
|
| 4113 |
+
{
|
| 4114 |
+
"epoch": 2.9994890137966275,
|
| 4115 |
+
"grad_norm": 0.18652462173874393,
|
| 4116 |
+
"learning_rate": 5e-06,
|
| 4117 |
+
"loss": 0.7294,
|
| 4118 |
+
"step": 5870
|
| 4119 |
+
},
|
| 4120 |
+
{
|
| 4121 |
+
"epoch": 3.0,
|
| 4122 |
+
"step": 5871,
|
| 4123 |
+
"total_flos": 2506240622592000.0,
|
| 4124 |
+
"train_loss": 0.754048796695819,
|
| 4125 |
+
"train_runtime": 27383.3094,
|
| 4126 |
+
"train_samples_per_second": 109.726,
|
| 4127 |
+
"train_steps_per_second": 0.214
|
| 4128 |
+
}
|
| 4129 |
+
],
|
| 4130 |
+
"logging_steps": 10,
|
| 4131 |
+
"max_steps": 5871,
|
| 4132 |
+
"num_input_tokens_seen": 0,
|
| 4133 |
+
"num_train_epochs": 3,
|
| 4134 |
+
"save_steps": 500,
|
| 4135 |
+
"stateful_callbacks": {
|
| 4136 |
+
"TrainerControl": {
|
| 4137 |
+
"args": {
|
| 4138 |
+
"should_epoch_stop": false,
|
| 4139 |
+
"should_evaluate": false,
|
| 4140 |
+
"should_log": false,
|
| 4141 |
+
"should_save": true,
|
| 4142 |
+
"should_training_stop": true
|
| 4143 |
+
},
|
| 4144 |
+
"attributes": {}
|
| 4145 |
+
}
|
| 4146 |
+
},
|
| 4147 |
+
"total_flos": 2506240622592000.0,
|
| 4148 |
+
"train_batch_size": 16,
|
| 4149 |
+
"trial_name": null,
|
| 4150 |
+
"trial_params": null
|
| 4151 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b8409659c1c8c4fb3eec19b3b0fa2ddc052cbb20117ccdda61994cd3fc0bf761
|
| 3 |
+
size 7480
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|