Upload 14 files
Browse files- README.md +58 -0
- adapter_config.json +28 -0
- adapter_model.safetensors +3 -0
- all_results.json +8 -0
- running_log.txt +197 -0
- special_tokens_map.json +1 -0
- tokenization_chatglm.py +300 -0
- tokenizer.model +3 -0
- tokenizer_config.json +21 -0
- train_results.json +8 -0
- trainer_config.yaml +27 -0
- trainer_log.jsonl +10 -0
- trainer_state.json +93 -0
- training_args.bin +3 -0
README.md
ADDED
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: other
|
3 |
+
library_name: peft
|
4 |
+
tags:
|
5 |
+
- llama-factory
|
6 |
+
- lora
|
7 |
+
- generated_from_trainer
|
8 |
+
base_model: THUDM/chatglm3-6b-base
|
9 |
+
model-index:
|
10 |
+
- name: test1
|
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 |
+
# test1
|
18 |
+
|
19 |
+
This model is a fine-tuned version of [THUDM/chatglm3-6b-base](https://huggingface.co/THUDM/chatglm3-6b-base) on the im_the_fated_villain_chapters 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: 2e-05
|
39 |
+
- train_batch_size: 2
|
40 |
+
- eval_batch_size: 8
|
41 |
+
- seed: 42
|
42 |
+
- gradient_accumulation_steps: 8
|
43 |
+
- total_train_batch_size: 16
|
44 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
45 |
+
- lr_scheduler_type: cosine
|
46 |
+
- num_epochs: 1.0
|
47 |
+
|
48 |
+
### Training results
|
49 |
+
|
50 |
+
|
51 |
+
|
52 |
+
### Framework versions
|
53 |
+
|
54 |
+
- PEFT 0.10.0
|
55 |
+
- Transformers 4.40.1
|
56 |
+
- Pytorch 2.2.2
|
57 |
+
- Datasets 2.19.0
|
58 |
+
- Tokenizers 0.19.1
|
adapter_config.json
ADDED
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"alpha_pattern": {},
|
3 |
+
"auto_mapping": null,
|
4 |
+
"base_model_name_or_path": "THUDM/chatglm3-6b-base",
|
5 |
+
"bias": "none",
|
6 |
+
"fan_in_fan_out": false,
|
7 |
+
"inference_mode": true,
|
8 |
+
"init_lora_weights": true,
|
9 |
+
"layer_replication": null,
|
10 |
+
"layers_pattern": null,
|
11 |
+
"layers_to_transform": null,
|
12 |
+
"loftq_config": {},
|
13 |
+
"lora_alpha": 16,
|
14 |
+
"lora_dropout": 0,
|
15 |
+
"megatron_config": null,
|
16 |
+
"megatron_core": "megatron.core",
|
17 |
+
"modules_to_save": null,
|
18 |
+
"peft_type": "LORA",
|
19 |
+
"r": 8,
|
20 |
+
"rank_pattern": {},
|
21 |
+
"revision": null,
|
22 |
+
"target_modules": [
|
23 |
+
"query_key_value"
|
24 |
+
],
|
25 |
+
"task_type": "CAUSAL_LM",
|
26 |
+
"use_dora": false,
|
27 |
+
"use_rslora": false
|
28 |
+
}
|
adapter_model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6a254467229cc39818fc5091b8bc6b37f88a4ddd915a54c7b7a8c2a7a7d1aa87
|
3 |
+
size 7807744
|
all_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 0.9919137466307277,
|
3 |
+
"total_flos": 2.703782589549773e+16,
|
4 |
+
"train_loss": 5.550413712211277,
|
5 |
+
"train_runtime": 20332.1955,
|
6 |
+
"train_samples_per_second": 0.036,
|
7 |
+
"train_steps_per_second": 0.002
|
8 |
+
}
|
running_log.txt
ADDED
@@ -0,0 +1,197 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
05/23/2024 11:11:50 - INFO - transformers.tokenization_utils_base - loading file tokenizer.model from cache at /Users/hanyiye/.cache/huggingface/hub/models--THUDM--chatglm3-6b-base/snapshots/f91a1de587fdc692073367198e65369669a0b49d/tokenizer.model
|
2 |
+
|
3 |
+
05/23/2024 11:11:50 - INFO - transformers.tokenization_utils_base - loading file added_tokens.json from cache at None
|
4 |
+
|
5 |
+
05/23/2024 11:11:50 - INFO - transformers.tokenization_utils_base - loading file special_tokens_map.json from cache at None
|
6 |
+
|
7 |
+
05/23/2024 11:11:50 - INFO - transformers.tokenization_utils_base - loading file tokenizer_config.json from cache at /Users/hanyiye/.cache/huggingface/hub/models--THUDM--chatglm3-6b-base/snapshots/f91a1de587fdc692073367198e65369669a0b49d/tokenizer_config.json
|
8 |
+
|
9 |
+
05/23/2024 11:11:50 - INFO - transformers.tokenization_utils_base - loading file tokenizer.json from cache at None
|
10 |
+
|
11 |
+
05/23/2024 11:11:51 - INFO - llmtuner.data.loader - Loading dataset ImTheFatedVillainChaptersDataset.json...
|
12 |
+
|
13 |
+
05/23/2024 11:11:51 - WARNING - llmtuner.data.utils - Checksum failed: missing SHA-1 hash value in dataset_info.json.
|
14 |
+
|
15 |
+
05/23/2024 11:12:02 - INFO - transformers.configuration_utils - loading configuration file config.json from cache at /Users/hanyiye/.cache/huggingface/hub/models--THUDM--chatglm3-6b-base/snapshots/f91a1de587fdc692073367198e65369669a0b49d/config.json
|
16 |
+
|
17 |
+
05/23/2024 11:12:02 - INFO - transformers.configuration_utils - loading configuration file config.json from cache at /Users/hanyiye/.cache/huggingface/hub/models--THUDM--chatglm3-6b-base/snapshots/f91a1de587fdc692073367198e65369669a0b49d/config.json
|
18 |
+
|
19 |
+
05/23/2024 11:12:02 - INFO - transformers.configuration_utils - Model config ChatGLMConfig {
|
20 |
+
"_name_or_path": "THUDM/chatglm3-6b-base",
|
21 |
+
"add_bias_linear": false,
|
22 |
+
"add_qkv_bias": true,
|
23 |
+
"apply_query_key_layer_scaling": true,
|
24 |
+
"apply_residual_connection_post_layernorm": false,
|
25 |
+
"architectures": [
|
26 |
+
"ChatGLMModel"
|
27 |
+
],
|
28 |
+
"attention_dropout": 0.0,
|
29 |
+
"attention_softmax_in_fp32": true,
|
30 |
+
"auto_map": {
|
31 |
+
"AutoConfig": "THUDM/chatglm3-6b-base--configuration_chatglm.ChatGLMConfig",
|
32 |
+
"AutoModel": "THUDM/chatglm3-6b-base--modeling_chatglm.ChatGLMForConditionalGeneration",
|
33 |
+
"AutoModelForCausalLM": "THUDM/chatglm3-6b-base--modeling_chatglm.ChatGLMForConditionalGeneration",
|
34 |
+
"AutoModelForSeq2SeqLM": "THUDM/chatglm3-6b-base--modeling_chatglm.ChatGLMForConditionalGeneration",
|
35 |
+
"AutoModelForSequenceClassification": "THUDM/chatglm3-6b-base--modeling_chatglm.ChatGLMForSequenceClassification"
|
36 |
+
},
|
37 |
+
"bias_dropout_fusion": true,
|
38 |
+
"classifier_dropout": null,
|
39 |
+
"eos_token_id": 2,
|
40 |
+
"ffn_hidden_size": 13696,
|
41 |
+
"fp32_residual_connection": false,
|
42 |
+
"hidden_dropout": 0.0,
|
43 |
+
"hidden_size": 4096,
|
44 |
+
"kv_channels": 128,
|
45 |
+
"layernorm_epsilon": 1e-05,
|
46 |
+
"model_type": "chatglm",
|
47 |
+
"multi_query_attention": true,
|
48 |
+
"multi_query_group_num": 2,
|
49 |
+
"num_attention_heads": 32,
|
50 |
+
"num_layers": 28,
|
51 |
+
"original_rope": true,
|
52 |
+
"pad_token_id": 0,
|
53 |
+
"padded_vocab_size": 65024,
|
54 |
+
"post_layer_norm": true,
|
55 |
+
"pre_seq_len": null,
|
56 |
+
"prefix_projection": false,
|
57 |
+
"quantization_bit": 0,
|
58 |
+
"rmsnorm": true,
|
59 |
+
"seq_length": 32768,
|
60 |
+
"tie_word_embeddings": false,
|
61 |
+
"torch_dtype": "float16",
|
62 |
+
"transformers_version": "4.40.1",
|
63 |
+
"use_cache": true,
|
64 |
+
"vocab_size": 65024
|
65 |
+
}
|
66 |
+
|
67 |
+
|
68 |
+
05/23/2024 11:12:03 - INFO - transformers.modeling_utils - loading weights file pytorch_model.bin from cache at /Users/hanyiye/.cache/huggingface/hub/models--THUDM--chatglm3-6b-base/snapshots/f91a1de587fdc692073367198e65369669a0b49d/pytorch_model.bin.index.json
|
69 |
+
|
70 |
+
05/23/2024 11:30:58 - INFO - transformers.modeling_utils - Instantiating ChatGLMForConditionalGeneration model under default dtype torch.float32.
|
71 |
+
|
72 |
+
05/23/2024 11:30:58 - INFO - transformers.generation.configuration_utils - Generate config GenerationConfig {
|
73 |
+
"eos_token_id": 2,
|
74 |
+
"pad_token_id": 0
|
75 |
+
}
|
76 |
+
|
77 |
+
|
78 |
+
05/23/2024 11:31:20 - INFO - transformers.modeling_utils - All model checkpoint weights were used when initializing ChatGLMForConditionalGeneration.
|
79 |
+
|
80 |
+
|
81 |
+
05/23/2024 11:31:20 - INFO - transformers.modeling_utils - All the weights of ChatGLMForConditionalGeneration were initialized from the model checkpoint at THUDM/chatglm3-6b-base.
|
82 |
+
If your task is similar to the task the model of the checkpoint was trained on, you can already use ChatGLMForConditionalGeneration for predictions without further training.
|
83 |
+
|
84 |
+
05/23/2024 11:31:21 - INFO - transformers.modeling_utils - Generation config file not found, using a generation config created from the model config.
|
85 |
+
|
86 |
+
05/23/2024 11:31:21 - WARNING - llmtuner.model.utils.checkpointing - You are using the old GC format, some features (e.g. BAdam) will be invalid.
|
87 |
+
|
88 |
+
05/23/2024 11:31:21 - INFO - llmtuner.model.utils.checkpointing - Gradient checkpointing enabled.
|
89 |
+
|
90 |
+
05/23/2024 11:31:21 - INFO - llmtuner.model.utils.attention - Using vanilla Attention implementation.
|
91 |
+
|
92 |
+
05/23/2024 11:31:21 - INFO - llmtuner.model.adapter - Fine-tuning method: LoRA
|
93 |
+
|
94 |
+
05/23/2024 11:31:21 - INFO - llmtuner.model.loader - trainable params: 1949696 || all params: 6245533696 || trainable%: 0.0312
|
95 |
+
|
96 |
+
05/23/2024 11:31:21 - INFO - transformers.trainer - You have loaded a model on multiple GPUs. `is_model_parallel` attribute will be force-set to `True` to avoid any unexpected behavior such as device placement mismatching.
|
97 |
+
|
98 |
+
05/23/2024 11:31:21 - INFO - transformers.trainer - ***** Running training *****
|
99 |
+
|
100 |
+
05/23/2024 11:31:21 - INFO - transformers.trainer - Num examples = 741
|
101 |
+
|
102 |
+
05/23/2024 11:31:21 - INFO - transformers.trainer - Num Epochs = 1
|
103 |
+
|
104 |
+
05/23/2024 11:31:21 - INFO - transformers.trainer - Instantaneous batch size per device = 2
|
105 |
+
|
106 |
+
05/23/2024 11:31:21 - INFO - transformers.trainer - Total train batch size (w. parallel, distributed & accumulation) = 16
|
107 |
+
|
108 |
+
05/23/2024 11:31:21 - INFO - transformers.trainer - Gradient Accumulation steps = 8
|
109 |
+
|
110 |
+
05/23/2024 11:31:21 - INFO - transformers.trainer - Total optimization steps = 46
|
111 |
+
|
112 |
+
05/23/2024 11:31:21 - INFO - transformers.trainer - Number of trainable parameters = 1,949,696
|
113 |
+
|
114 |
+
05/23/2024 12:06:13 - INFO - llmtuner.extras.callbacks - {'loss': 1.8053, 'learning_rate': 1.9423e-05, 'epoch': 0.11}
|
115 |
+
|
116 |
+
05/23/2024 12:45:11 - INFO - llmtuner.extras.callbacks - {'loss': 1.7973, 'learning_rate': 1.7757e-05, 'epoch': 0.22}
|
117 |
+
|
118 |
+
05/23/2024 13:25:31 - INFO - llmtuner.extras.callbacks - {'loss': 1.7813, 'learning_rate': 1.5196e-05, 'epoch': 0.32}
|
119 |
+
|
120 |
+
05/23/2024 14:05:34 - INFO - llmtuner.extras.callbacks - {'loss': 1.8348, 'learning_rate': 1.2035e-05, 'epoch': 0.43}
|
121 |
+
|
122 |
+
05/23/2024 14:45:14 - INFO - llmtuner.extras.callbacks - {'loss': 1.7943, 'learning_rate': 8.6383e-06, 'epoch': 0.54}
|
123 |
+
|
124 |
+
05/23/2024 15:22:48 - INFO - llmtuner.extras.callbacks - {'loss': 42.0508, 'learning_rate': 5.3993e-06, 'epoch': 0.65}
|
125 |
+
|
126 |
+
05/23/2024 15:57:14 - INFO - llmtuner.extras.callbacks - {'loss': 0.0000, 'learning_rate': 2.6916e-06, 'epoch': 0.75}
|
127 |
+
|
128 |
+
05/23/2024 16:29:15 - INFO - llmtuner.extras.callbacks - {'loss': 0.0000, 'learning_rate': 8.2789e-07, 'epoch': 0.86}
|
129 |
+
|
130 |
+
05/23/2024 17:03:20 - INFO - llmtuner.extras.callbacks - {'loss': 0.0000, 'learning_rate': 2.3312e-08, 'epoch': 0.97}
|
131 |
+
|
132 |
+
05/23/2024 17:10:13 - INFO - transformers.trainer -
|
133 |
+
|
134 |
+
Training completed. Do not forget to share your model on huggingface.co/models =)
|
135 |
+
|
136 |
+
|
137 |
+
|
138 |
+
05/23/2024 17:10:14 - INFO - transformers.trainer - Saving model checkpoint to saves/ChatGLM3-6B-Base/lora/test1
|
139 |
+
|
140 |
+
05/23/2024 17:10:15 - INFO - transformers.configuration_utils - loading configuration file config.json from cache at /Users/hanyiye/.cache/huggingface/hub/models--THUDM--chatglm3-6b-base/snapshots/f91a1de587fdc692073367198e65369669a0b49d/config.json
|
141 |
+
|
142 |
+
05/23/2024 17:10:15 - INFO - transformers.configuration_utils - Model config ChatGLMConfig {
|
143 |
+
"_name_or_path": "THUDM/chatglm3-6b-base",
|
144 |
+
"add_bias_linear": false,
|
145 |
+
"add_qkv_bias": true,
|
146 |
+
"apply_query_key_layer_scaling": true,
|
147 |
+
"apply_residual_connection_post_layernorm": false,
|
148 |
+
"architectures": [
|
149 |
+
"ChatGLMModel"
|
150 |
+
],
|
151 |
+
"attention_dropout": 0.0,
|
152 |
+
"attention_softmax_in_fp32": true,
|
153 |
+
"auto_map": {
|
154 |
+
"AutoConfig": "THUDM/chatglm3-6b-base--configuration_chatglm.ChatGLMConfig",
|
155 |
+
"AutoModel": "THUDM/chatglm3-6b-base--modeling_chatglm.ChatGLMForConditionalGeneration",
|
156 |
+
"AutoModelForCausalLM": "THUDM/chatglm3-6b-base--modeling_chatglm.ChatGLMForConditionalGeneration",
|
157 |
+
"AutoModelForSeq2SeqLM": "THUDM/chatglm3-6b-base--modeling_chatglm.ChatGLMForConditionalGeneration",
|
158 |
+
"AutoModelForSequenceClassification": "THUDM/chatglm3-6b-base--modeling_chatglm.ChatGLMForSequenceClassification"
|
159 |
+
},
|
160 |
+
"bias_dropout_fusion": true,
|
161 |
+
"classifier_dropout": null,
|
162 |
+
"eos_token_id": 2,
|
163 |
+
"ffn_hidden_size": 13696,
|
164 |
+
"fp32_residual_connection": false,
|
165 |
+
"hidden_dropout": 0.0,
|
166 |
+
"hidden_size": 4096,
|
167 |
+
"kv_channels": 128,
|
168 |
+
"layernorm_epsilon": 1e-05,
|
169 |
+
"model_type": "chatglm",
|
170 |
+
"multi_query_attention": true,
|
171 |
+
"multi_query_group_num": 2,
|
172 |
+
"num_attention_heads": 32,
|
173 |
+
"num_layers": 28,
|
174 |
+
"original_rope": true,
|
175 |
+
"pad_token_id": 0,
|
176 |
+
"padded_vocab_size": 65024,
|
177 |
+
"post_layer_norm": true,
|
178 |
+
"pre_seq_len": null,
|
179 |
+
"prefix_projection": false,
|
180 |
+
"quantization_bit": 0,
|
181 |
+
"rmsnorm": true,
|
182 |
+
"seq_length": 32768,
|
183 |
+
"tie_word_embeddings": false,
|
184 |
+
"torch_dtype": "float16",
|
185 |
+
"transformers_version": "4.40.1",
|
186 |
+
"use_cache": true,
|
187 |
+
"vocab_size": 65024
|
188 |
+
}
|
189 |
+
|
190 |
+
|
191 |
+
05/23/2024 17:10:15 - INFO - transformers.tokenization_utils_base - tokenizer config file saved in saves/ChatGLM3-6B-Base/lora/test1/tokenizer_config.json
|
192 |
+
|
193 |
+
05/23/2024 17:10:15 - INFO - transformers.tokenization_utils_base - Special tokens file saved in saves/ChatGLM3-6B-Base/lora/test1/special_tokens_map.json
|
194 |
+
|
195 |
+
05/23/2024 17:10:15 - INFO - transformers.modelcard - Dropping the following result as it does not have all the necessary fields:
|
196 |
+
{'task': {'name': 'Causal Language Modeling', 'type': 'text-generation'}}
|
197 |
+
|
special_tokens_map.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{}
|
tokenization_chatglm.py
ADDED
@@ -0,0 +1,300 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import json
|
2 |
+
import os
|
3 |
+
import re
|
4 |
+
from typing import List, Optional, Union, Dict
|
5 |
+
from sentencepiece import SentencePieceProcessor
|
6 |
+
from transformers import PreTrainedTokenizer
|
7 |
+
from transformers.utils import logging, PaddingStrategy
|
8 |
+
from transformers.tokenization_utils_base import EncodedInput, BatchEncoding
|
9 |
+
|
10 |
+
|
11 |
+
class SPTokenizer:
|
12 |
+
def __init__(self, model_path: str):
|
13 |
+
# reload tokenizer
|
14 |
+
assert os.path.isfile(model_path), model_path
|
15 |
+
self.sp_model = SentencePieceProcessor(model_file=model_path)
|
16 |
+
|
17 |
+
# BOS / EOS token IDs
|
18 |
+
self.n_words: int = self.sp_model.vocab_size()
|
19 |
+
self.bos_id: int = self.sp_model.bos_id()
|
20 |
+
self.eos_id: int = self.sp_model.eos_id()
|
21 |
+
self.pad_id: int = self.sp_model.unk_id()
|
22 |
+
assert self.sp_model.vocab_size() == self.sp_model.get_piece_size()
|
23 |
+
|
24 |
+
role_special_tokens = ["<|system|>", "<|user|>", "<|assistant|>", "<|observation|>"]
|
25 |
+
special_tokens = ["[MASK]", "[gMASK]", "[sMASK]", "sop", "eop"] + role_special_tokens
|
26 |
+
self.special_tokens = {}
|
27 |
+
self.index_special_tokens = {}
|
28 |
+
for token in special_tokens:
|
29 |
+
self.special_tokens[token] = self.n_words
|
30 |
+
self.index_special_tokens[self.n_words] = token
|
31 |
+
self.n_words += 1
|
32 |
+
self.role_special_token_expression = "|".join([re.escape(token) for token in role_special_tokens])
|
33 |
+
|
34 |
+
def tokenize(self, s: str, encode_special_tokens=False):
|
35 |
+
if encode_special_tokens:
|
36 |
+
last_index = 0
|
37 |
+
t = []
|
38 |
+
for match in re.finditer(self.role_special_token_expression, s):
|
39 |
+
if last_index < match.start():
|
40 |
+
t.extend(self.sp_model.EncodeAsPieces(s[last_index:match.start()]))
|
41 |
+
t.append(s[match.start():match.end()])
|
42 |
+
last_index = match.end()
|
43 |
+
if last_index < len(s):
|
44 |
+
t.extend(self.sp_model.EncodeAsPieces(s[last_index:]))
|
45 |
+
return t
|
46 |
+
else:
|
47 |
+
return self.sp_model.EncodeAsPieces(s)
|
48 |
+
|
49 |
+
def encode(self, s: str, bos: bool = False, eos: bool = False) -> List[int]:
|
50 |
+
assert type(s) is str
|
51 |
+
t = self.sp_model.encode(s)
|
52 |
+
if bos:
|
53 |
+
t = [self.bos_id] + t
|
54 |
+
if eos:
|
55 |
+
t = t + [self.eos_id]
|
56 |
+
return t
|
57 |
+
|
58 |
+
def decode(self, t: List[int]) -> str:
|
59 |
+
text, buffer = "", []
|
60 |
+
for token in t:
|
61 |
+
if token in self.index_special_tokens:
|
62 |
+
if buffer:
|
63 |
+
text += self.sp_model.decode(buffer)
|
64 |
+
buffer = []
|
65 |
+
text += self.index_special_tokens[token]
|
66 |
+
else:
|
67 |
+
buffer.append(token)
|
68 |
+
if buffer:
|
69 |
+
text += self.sp_model.decode(buffer)
|
70 |
+
return text
|
71 |
+
|
72 |
+
def decode_tokens(self, tokens: List[str]) -> str:
|
73 |
+
text = self.sp_model.DecodePieces(tokens)
|
74 |
+
return text
|
75 |
+
|
76 |
+
def convert_token_to_id(self, token):
|
77 |
+
""" Converts a token (str) in an id using the vocab. """
|
78 |
+
if token in self.special_tokens:
|
79 |
+
return self.special_tokens[token]
|
80 |
+
return self.sp_model.PieceToId(token)
|
81 |
+
|
82 |
+
def convert_id_to_token(self, index):
|
83 |
+
"""Converts an index (integer) in a token (str) using the vocab."""
|
84 |
+
if index in self.index_special_tokens:
|
85 |
+
return self.index_special_tokens[index]
|
86 |
+
if index in [self.eos_id, self.bos_id, self.pad_id] or index < 0:
|
87 |
+
return ""
|
88 |
+
return self.sp_model.IdToPiece(index)
|
89 |
+
|
90 |
+
|
91 |
+
class ChatGLMTokenizer(PreTrainedTokenizer):
|
92 |
+
vocab_files_names = {"vocab_file": "tokenizer.model"}
|
93 |
+
|
94 |
+
model_input_names = ["input_ids", "attention_mask", "position_ids"]
|
95 |
+
|
96 |
+
def __init__(self, vocab_file, padding_side="left", clean_up_tokenization_spaces=False, encode_special_tokens=False,
|
97 |
+
**kwargs):
|
98 |
+
self.name = "GLMTokenizer"
|
99 |
+
|
100 |
+
self.vocab_file = vocab_file
|
101 |
+
self.tokenizer = SPTokenizer(vocab_file)
|
102 |
+
self.special_tokens = {
|
103 |
+
"<bos>": self.tokenizer.bos_id,
|
104 |
+
"<eos>": self.tokenizer.eos_id,
|
105 |
+
"<pad>": self.tokenizer.pad_id
|
106 |
+
}
|
107 |
+
self.encode_special_tokens = encode_special_tokens
|
108 |
+
super().__init__(padding_side=padding_side, clean_up_tokenization_spaces=clean_up_tokenization_spaces,
|
109 |
+
encode_special_tokens=encode_special_tokens,
|
110 |
+
**kwargs)
|
111 |
+
|
112 |
+
def get_command(self, token):
|
113 |
+
if token in self.special_tokens:
|
114 |
+
return self.special_tokens[token]
|
115 |
+
assert token in self.tokenizer.special_tokens, f"{token} is not a special token for {self.name}"
|
116 |
+
return self.tokenizer.special_tokens[token]
|
117 |
+
|
118 |
+
@property
|
119 |
+
def unk_token(self) -> str:
|
120 |
+
return "<unk>"
|
121 |
+
|
122 |
+
@property
|
123 |
+
def pad_token(self) -> str:
|
124 |
+
return "<unk>"
|
125 |
+
|
126 |
+
@property
|
127 |
+
def pad_token_id(self):
|
128 |
+
return self.get_command("<pad>")
|
129 |
+
|
130 |
+
@property
|
131 |
+
def eos_token(self) -> str:
|
132 |
+
return "</s>"
|
133 |
+
|
134 |
+
@property
|
135 |
+
def eos_token_id(self):
|
136 |
+
return self.get_command("<eos>")
|
137 |
+
|
138 |
+
@property
|
139 |
+
def vocab_size(self):
|
140 |
+
return self.tokenizer.n_words
|
141 |
+
|
142 |
+
def get_vocab(self):
|
143 |
+
""" Returns vocab as a dict """
|
144 |
+
vocab = {self._convert_id_to_token(i): i for i in range(self.vocab_size)}
|
145 |
+
vocab.update(self.added_tokens_encoder)
|
146 |
+
return vocab
|
147 |
+
|
148 |
+
def _tokenize(self, text, **kwargs):
|
149 |
+
return self.tokenizer.tokenize(text, encode_special_tokens=self.encode_special_tokens)
|
150 |
+
|
151 |
+
def _convert_token_to_id(self, token):
|
152 |
+
""" Converts a token (str) in an id using the vocab. """
|
153 |
+
return self.tokenizer.convert_token_to_id(token)
|
154 |
+
|
155 |
+
def _convert_id_to_token(self, index):
|
156 |
+
"""Converts an index (integer) in a token (str) using the vocab."""
|
157 |
+
return self.tokenizer.convert_id_to_token(index)
|
158 |
+
|
159 |
+
def convert_tokens_to_string(self, tokens: List[str]) -> str:
|
160 |
+
return self.tokenizer.decode_tokens(tokens)
|
161 |
+
|
162 |
+
def save_vocabulary(self, save_directory, filename_prefix=None):
|
163 |
+
"""
|
164 |
+
Save the vocabulary and special tokens file to a directory.
|
165 |
+
|
166 |
+
Args:
|
167 |
+
save_directory (`str`):
|
168 |
+
The directory in which to save the vocabulary.
|
169 |
+
filename_prefix (`str`, *optional*):
|
170 |
+
An optional prefix to add to the named of the saved files.
|
171 |
+
|
172 |
+
Returns:
|
173 |
+
`Tuple(str)`: Paths to the files saved.
|
174 |
+
"""
|
175 |
+
if os.path.isdir(save_directory):
|
176 |
+
vocab_file = os.path.join(
|
177 |
+
save_directory, self.vocab_files_names["vocab_file"]
|
178 |
+
)
|
179 |
+
else:
|
180 |
+
vocab_file = save_directory
|
181 |
+
|
182 |
+
with open(self.vocab_file, 'rb') as fin:
|
183 |
+
proto_str = fin.read()
|
184 |
+
|
185 |
+
with open(vocab_file, "wb") as writer:
|
186 |
+
writer.write(proto_str)
|
187 |
+
|
188 |
+
return (vocab_file,)
|
189 |
+
|
190 |
+
def get_prefix_tokens(self):
|
191 |
+
prefix_tokens = [self.get_command("[gMASK]"), self.get_command("sop")]
|
192 |
+
return prefix_tokens
|
193 |
+
|
194 |
+
def build_single_message(self, role, metadata, message):
|
195 |
+
assert role in ["system", "user", "assistant", "observation"], role
|
196 |
+
role_tokens = [self.get_command(f"<|{role}|>")] + self.tokenizer.encode(f"{metadata}\n")
|
197 |
+
message_tokens = self.tokenizer.encode(message)
|
198 |
+
tokens = role_tokens + message_tokens
|
199 |
+
return tokens
|
200 |
+
|
201 |
+
def build_chat_input(self, query, history=None, role="user"):
|
202 |
+
if history is None:
|
203 |
+
history = []
|
204 |
+
input_ids = []
|
205 |
+
for item in history:
|
206 |
+
content = item["content"]
|
207 |
+
if item["role"] == "system" and "tools" in item:
|
208 |
+
content = content + "\n" + json.dumps(item["tools"], indent=4, ensure_ascii=False)
|
209 |
+
input_ids.extend(self.build_single_message(item["role"], item.get("metadata", ""), content))
|
210 |
+
input_ids.extend(self.build_single_message(role, "", query))
|
211 |
+
input_ids.extend([self.get_command("<|assistant|>")])
|
212 |
+
return self.batch_encode_plus([input_ids], return_tensors="pt", is_split_into_words=True)
|
213 |
+
|
214 |
+
def build_inputs_with_special_tokens(
|
215 |
+
self, token_ids_0: List[int], token_ids_1: Optional[List[int]] = None
|
216 |
+
) -> List[int]:
|
217 |
+
"""
|
218 |
+
Build model inputs from a sequence or a pair of sequence for sequence classification tasks by concatenating and
|
219 |
+
adding special tokens. A BERT sequence has the following format:
|
220 |
+
|
221 |
+
- single sequence: `[CLS] X [SEP]`
|
222 |
+
- pair of sequences: `[CLS] A [SEP] B [SEP]`
|
223 |
+
|
224 |
+
Args:
|
225 |
+
token_ids_0 (`List[int]`):
|
226 |
+
List of IDs to which the special tokens will be added.
|
227 |
+
token_ids_1 (`List[int]`, *optional*):
|
228 |
+
Optional second list of IDs for sequence pairs.
|
229 |
+
|
230 |
+
Returns:
|
231 |
+
`List[int]`: List of [input IDs](../glossary#input-ids) with the appropriate special tokens.
|
232 |
+
"""
|
233 |
+
prefix_tokens = self.get_prefix_tokens()
|
234 |
+
token_ids_0 = prefix_tokens + token_ids_0
|
235 |
+
if token_ids_1 is not None:
|
236 |
+
token_ids_0 = token_ids_0 + token_ids_1 + [self.get_command("<eos>")]
|
237 |
+
return token_ids_0
|
238 |
+
|
239 |
+
def _pad(
|
240 |
+
self,
|
241 |
+
encoded_inputs: Union[Dict[str, EncodedInput], BatchEncoding],
|
242 |
+
max_length: Optional[int] = None,
|
243 |
+
padding_strategy: PaddingStrategy = PaddingStrategy.DO_NOT_PAD,
|
244 |
+
pad_to_multiple_of: Optional[int] = None,
|
245 |
+
return_attention_mask: Optional[bool] = None,
|
246 |
+
) -> dict:
|
247 |
+
"""
|
248 |
+
Pad encoded inputs (on left/right and up to predefined length or max length in the batch)
|
249 |
+
|
250 |
+
Args:
|
251 |
+
encoded_inputs:
|
252 |
+
Dictionary of tokenized inputs (`List[int]`) or batch of tokenized inputs (`List[List[int]]`).
|
253 |
+
max_length: maximum length of the returned list and optionally padding length (see below).
|
254 |
+
Will truncate by taking into account the special tokens.
|
255 |
+
padding_strategy: PaddingStrategy to use for padding.
|
256 |
+
|
257 |
+
- PaddingStrategy.LONGEST Pad to the longest sequence in the batch
|
258 |
+
- PaddingStrategy.MAX_LENGTH: Pad to the max length (default)
|
259 |
+
- PaddingStrategy.DO_NOT_PAD: Do not pad
|
260 |
+
The tokenizer padding sides are defined in self.padding_side:
|
261 |
+
|
262 |
+
- 'left': pads on the left of the sequences
|
263 |
+
- 'right': pads on the right of the sequences
|
264 |
+
pad_to_multiple_of: (optional) Integer if set will pad the sequence to a multiple of the provided value.
|
265 |
+
This is especially useful to enable the use of Tensor Core on NVIDIA hardware with compute capability
|
266 |
+
`>= 7.5` (Volta).
|
267 |
+
return_attention_mask:
|
268 |
+
(optional) Set to False to avoid returning attention mask (default: set to model specifics)
|
269 |
+
"""
|
270 |
+
# Load from model defaults
|
271 |
+
assert self.padding_side == "left"
|
272 |
+
|
273 |
+
required_input = encoded_inputs[self.model_input_names[0]]
|
274 |
+
seq_length = len(required_input)
|
275 |
+
|
276 |
+
if padding_strategy == PaddingStrategy.LONGEST:
|
277 |
+
max_length = len(required_input)
|
278 |
+
|
279 |
+
if max_length is not None and pad_to_multiple_of is not None and (max_length % pad_to_multiple_of != 0):
|
280 |
+
max_length = ((max_length // pad_to_multiple_of) + 1) * pad_to_multiple_of
|
281 |
+
|
282 |
+
needs_to_be_padded = padding_strategy != PaddingStrategy.DO_NOT_PAD and len(required_input) != max_length
|
283 |
+
|
284 |
+
# Initialize attention mask if not present.
|
285 |
+
if "attention_mask" not in encoded_inputs:
|
286 |
+
encoded_inputs["attention_mask"] = [1] * seq_length
|
287 |
+
|
288 |
+
if "position_ids" not in encoded_inputs:
|
289 |
+
encoded_inputs["position_ids"] = list(range(seq_length))
|
290 |
+
|
291 |
+
if needs_to_be_padded:
|
292 |
+
difference = max_length - len(required_input)
|
293 |
+
|
294 |
+
if "attention_mask" in encoded_inputs:
|
295 |
+
encoded_inputs["attention_mask"] = [0] * difference + encoded_inputs["attention_mask"]
|
296 |
+
if "position_ids" in encoded_inputs:
|
297 |
+
encoded_inputs["position_ids"] = [0] * difference + encoded_inputs["position_ids"]
|
298 |
+
encoded_inputs[self.model_input_names[0]] = [self.pad_token_id] * difference + required_input
|
299 |
+
|
300 |
+
return encoded_inputs
|
tokenizer.model
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e7dc4c393423b76e4373e5157ddc34803a0189ba96b21ddbb40269d31468a6f2
|
3 |
+
size 1018370
|
tokenizer_config.json
ADDED
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"added_tokens_decoder": {},
|
3 |
+
"auto_map": {
|
4 |
+
"AutoTokenizer": [
|
5 |
+
"tokenization_chatglm.ChatGLMTokenizer",
|
6 |
+
null
|
7 |
+
]
|
8 |
+
},
|
9 |
+
"chat_template": "{% if messages[0]['role'] == 'system' %}{% set system_message = messages[0]['content'] %}{% endif %}{% if system_message is defined %}{{ system_message + '\\n' }}{% endif %}{% for message in messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ 'Human: ' + content + '\\nAssistant: ' }}{% elif message['role'] == 'assistant' %}{{ content + '</s>' + '\\n' }}{% endif %}{% endfor %}",
|
10 |
+
"clean_up_tokenization_spaces": false,
|
11 |
+
"do_lower_case": false,
|
12 |
+
"encode_special_tokens": false,
|
13 |
+
"eos_token": "</s>",
|
14 |
+
"model_max_length": 1000000000000000019884624838656,
|
15 |
+
"pad_token": "<unk>",
|
16 |
+
"padding_side": "right",
|
17 |
+
"remove_space": false,
|
18 |
+
"split_special_tokens": false,
|
19 |
+
"tokenizer_class": "ChatGLMTokenizer",
|
20 |
+
"unk_token": "<unk>"
|
21 |
+
}
|
train_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 0.9919137466307277,
|
3 |
+
"total_flos": 2.703782589549773e+16,
|
4 |
+
"train_loss": 5.550413712211277,
|
5 |
+
"train_runtime": 20332.1955,
|
6 |
+
"train_samples_per_second": 0.036,
|
7 |
+
"train_steps_per_second": 0.002
|
8 |
+
}
|
trainer_config.yaml
ADDED
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
cutoff_len: 1024
|
2 |
+
dataset: im_the_fated_villain_chapters
|
3 |
+
dataset_dir: data
|
4 |
+
do_train: true
|
5 |
+
finetuning_type: lora
|
6 |
+
flash_attn: auto
|
7 |
+
gradient_accumulation_steps: 8
|
8 |
+
learning_rate: 2.0e-05
|
9 |
+
logging_steps: 5
|
10 |
+
lora_alpha: 16
|
11 |
+
lora_dropout: 0
|
12 |
+
lora_rank: 8
|
13 |
+
lora_target: query_key_value
|
14 |
+
lr_scheduler_type: cosine
|
15 |
+
max_grad_norm: 1.0
|
16 |
+
max_samples: 100000
|
17 |
+
model_name_or_path: THUDM/chatglm3-6b-base
|
18 |
+
num_train_epochs: 1.0
|
19 |
+
optim: adamw_torch
|
20 |
+
output_dir: saves/ChatGLM3-6B-Base/lora/test1
|
21 |
+
packing: false
|
22 |
+
per_device_train_batch_size: 2
|
23 |
+
report_to: none
|
24 |
+
save_steps: 100
|
25 |
+
stage: sft
|
26 |
+
template: default
|
27 |
+
warmup_steps: 0
|
trainer_log.jsonl
ADDED
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{"current_steps": 5, "total_steps": 46, "loss": 1.8053, "learning_rate": 1.9422609221188208e-05, "epoch": 0.1078167115902965, "percentage": 10.87, "elapsed_time": "0:34:51", "remaining_time": "4:45:50"}
|
2 |
+
{"current_steps": 10, "total_steps": 46, "loss": 1.7973, "learning_rate": 1.77571129070442e-05, "epoch": 0.215633423180593, "percentage": 21.74, "elapsed_time": "1:13:49", "remaining_time": "4:25:46"}
|
3 |
+
{"current_steps": 15, "total_steps": 46, "loss": 1.7813, "learning_rate": 1.5195839500354338e-05, "epoch": 0.32345013477088946, "percentage": 32.61, "elapsed_time": "1:54:09", "remaining_time": "3:55:55"}
|
4 |
+
{"current_steps": 20, "total_steps": 46, "loss": 1.8348, "learning_rate": 1.2034560130526341e-05, "epoch": 0.431266846361186, "percentage": 43.48, "elapsed_time": "2:34:12", "remaining_time": "3:20:27"}
|
5 |
+
{"current_steps": 25, "total_steps": 46, "loss": 1.7943, "learning_rate": 8.638333509037535e-06, "epoch": 0.5390835579514824, "percentage": 54.35, "elapsed_time": "3:13:52", "remaining_time": "2:42:51"}
|
6 |
+
{"current_steps": 30, "total_steps": 46, "loss": 42.0508, "learning_rate": 5.399349622688479e-06, "epoch": 0.6469002695417789, "percentage": 65.22, "elapsed_time": "3:51:26", "remaining_time": "2:03:26"}
|
7 |
+
{"current_steps": 35, "total_steps": 46, "loss": 0.0, "learning_rate": 2.691640357218759e-06, "epoch": 0.7547169811320755, "percentage": 76.09, "elapsed_time": "4:25:52", "remaining_time": "1:23:33"}
|
8 |
+
{"current_steps": 40, "total_steps": 46, "loss": 0.0, "learning_rate": 8.278869849454718e-07, "epoch": 0.862533692722372, "percentage": 86.96, "elapsed_time": "4:57:53", "remaining_time": "0:44:41"}
|
9 |
+
{"current_steps": 45, "total_steps": 46, "loss": 0.0, "learning_rate": 2.3312308094607382e-08, "epoch": 0.9703504043126685, "percentage": 97.83, "elapsed_time": "5:31:58", "remaining_time": "0:07:22"}
|
10 |
+
{"current_steps": 46, "total_steps": 46, "epoch": 0.9919137466307277, "percentage": 100.0, "elapsed_time": "5:38:52", "remaining_time": "0:00:00"}
|
trainer_state.json
ADDED
@@ -0,0 +1,93 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 0.9919137466307277,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 46,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.1078167115902965,
|
13 |
+
"grad_norm": 0.25050440430641174,
|
14 |
+
"learning_rate": 1.9422609221188208e-05,
|
15 |
+
"loss": 1.8053,
|
16 |
+
"step": 5
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 0.215633423180593,
|
20 |
+
"grad_norm": 0.2306346893310547,
|
21 |
+
"learning_rate": 1.77571129070442e-05,
|
22 |
+
"loss": 1.7973,
|
23 |
+
"step": 10
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.32345013477088946,
|
27 |
+
"grad_norm": 0.2684953510761261,
|
28 |
+
"learning_rate": 1.5195839500354338e-05,
|
29 |
+
"loss": 1.7813,
|
30 |
+
"step": 15
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"epoch": 0.431266846361186,
|
34 |
+
"grad_norm": 0.2725406587123871,
|
35 |
+
"learning_rate": 1.2034560130526341e-05,
|
36 |
+
"loss": 1.8348,
|
37 |
+
"step": 20
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.5390835579514824,
|
41 |
+
"grad_norm": 0.30252954363822937,
|
42 |
+
"learning_rate": 8.638333509037535e-06,
|
43 |
+
"loss": 1.7943,
|
44 |
+
"step": 25
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.6469002695417789,
|
48 |
+
"grad_norm": NaN,
|
49 |
+
"learning_rate": 5.399349622688479e-06,
|
50 |
+
"loss": 42.0508,
|
51 |
+
"step": 30
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.7547169811320755,
|
55 |
+
"grad_norm": NaN,
|
56 |
+
"learning_rate": 2.691640357218759e-06,
|
57 |
+
"loss": 0.0,
|
58 |
+
"step": 35
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"epoch": 0.862533692722372,
|
62 |
+
"grad_norm": NaN,
|
63 |
+
"learning_rate": 8.278869849454718e-07,
|
64 |
+
"loss": 0.0,
|
65 |
+
"step": 40
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.9703504043126685,
|
69 |
+
"grad_norm": NaN,
|
70 |
+
"learning_rate": 2.3312308094607382e-08,
|
71 |
+
"loss": 0.0,
|
72 |
+
"step": 45
|
73 |
+
},
|
74 |
+
{
|
75 |
+
"epoch": 0.9919137466307277,
|
76 |
+
"step": 46,
|
77 |
+
"total_flos": 2.703782589549773e+16,
|
78 |
+
"train_loss": 5.550413712211277,
|
79 |
+
"train_runtime": 20332.1955,
|
80 |
+
"train_samples_per_second": 0.036,
|
81 |
+
"train_steps_per_second": 0.002
|
82 |
+
}
|
83 |
+
],
|
84 |
+
"logging_steps": 5,
|
85 |
+
"max_steps": 46,
|
86 |
+
"num_input_tokens_seen": 0,
|
87 |
+
"num_train_epochs": 1,
|
88 |
+
"save_steps": 100,
|
89 |
+
"total_flos": 2.703782589549773e+16,
|
90 |
+
"train_batch_size": 2,
|
91 |
+
"trial_name": null,
|
92 |
+
"trial_params": null
|
93 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:da25c55235aa6432fae9f2038b23422a509e51cccc55e4a6c1403532d635b699
|
3 |
+
size 5112
|