Upload folder using huggingface_hub
Browse files- checkpoint-1000/config.json +41 -0
- checkpoint-1000/generation_config.json +7 -0
- checkpoint-1000/model.safetensors +3 -0
- checkpoint-1000/optimizer.pt +3 -0
- checkpoint-1000/rng_state.pth +3 -0
- checkpoint-1000/scheduler.pt +3 -0
- checkpoint-1000/trainer_state.json +741 -0
- checkpoint-1000/training_args.bin +3 -0
- checkpoint-1500/config.json +41 -0
- checkpoint-1500/generation_config.json +7 -0
- checkpoint-1500/model.safetensors +3 -0
- checkpoint-1500/optimizer.pt +3 -0
- checkpoint-1500/rng_state.pth +3 -0
- checkpoint-1500/scheduler.pt +3 -0
- checkpoint-1500/trainer_state.json +1099 -0
- checkpoint-1500/training_args.bin +3 -0
- checkpoint-1539/config.json +41 -0
- checkpoint-1539/generation_config.json +7 -0
- checkpoint-1539/model.safetensors +3 -0
- checkpoint-1539/optimizer.pt +3 -0
- checkpoint-1539/rng_state.pth +3 -0
- checkpoint-1539/scheduler.pt +3 -0
- checkpoint-1539/trainer_state.json +1120 -0
- checkpoint-1539/training_args.bin +3 -0
- checkpoint-500/config.json +41 -0
- checkpoint-500/generation_config.json +7 -0
- checkpoint-500/model.safetensors +3 -0
- checkpoint-500/optimizer.pt +3 -0
- checkpoint-500/rng_state.pth +3 -0
- checkpoint-500/scheduler.pt +3 -0
- checkpoint-500/trainer_state.json +383 -0
- checkpoint-500/training_args.bin +3 -0
- runs/Feb17_10-24-44_8cfcbcdf5939/events.out.tfevents.1739787888.8cfcbcdf5939.1018.0 +3 -0
- runs/Feb17_10-27-50_8cfcbcdf5939/events.out.tfevents.1739788072.8cfcbcdf5939.1018.1 +3 -0
checkpoint-1000/config.json
ADDED
|
@@ -0,0 +1,41 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "ai-forever/rugpt3small_based_on_gpt2",
|
| 3 |
+
"activation_function": "gelu_new",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.1,
|
| 8 |
+
"bos_token_id": 1,
|
| 9 |
+
"embd_pdrop": 0.1,
|
| 10 |
+
"eos_token_id": 2,
|
| 11 |
+
"gradient_checkpointing": false,
|
| 12 |
+
"id2label": {
|
| 13 |
+
"0": "LABEL_0"
|
| 14 |
+
},
|
| 15 |
+
"initializer_range": 0.02,
|
| 16 |
+
"label2id": {
|
| 17 |
+
"LABEL_0": 0
|
| 18 |
+
},
|
| 19 |
+
"layer_norm_epsilon": 1e-05,
|
| 20 |
+
"model_type": "gpt2",
|
| 21 |
+
"n_ctx": 2048,
|
| 22 |
+
"n_embd": 768,
|
| 23 |
+
"n_head": 12,
|
| 24 |
+
"n_inner": null,
|
| 25 |
+
"n_layer": 12,
|
| 26 |
+
"n_positions": 2048,
|
| 27 |
+
"pad_token_id": 0,
|
| 28 |
+
"reorder_and_upcast_attn": false,
|
| 29 |
+
"resid_pdrop": 0.1,
|
| 30 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 31 |
+
"scale_attn_weights": true,
|
| 32 |
+
"summary_activation": null,
|
| 33 |
+
"summary_first_dropout": 0.1,
|
| 34 |
+
"summary_proj_to_labels": true,
|
| 35 |
+
"summary_type": "cls_index",
|
| 36 |
+
"summary_use_proj": true,
|
| 37 |
+
"torch_dtype": "float32",
|
| 38 |
+
"transformers_version": "4.48.3",
|
| 39 |
+
"use_cache": true,
|
| 40 |
+
"vocab_size": 50264
|
| 41 |
+
}
|
checkpoint-1000/generation_config.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"pad_token_id": 0,
|
| 6 |
+
"transformers_version": "4.48.3"
|
| 7 |
+
}
|
checkpoint-1000/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:51c01879d30c6d50480072552a6c1c3d31532452873c1227c1804a463d87c6fd
|
| 3 |
+
size 500941440
|
checkpoint-1000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:da3d5d08e6e6a224a2101739dc4bbd447a00477d33295311caa1535ba67d535f
|
| 3 |
+
size 1001976762
|
checkpoint-1000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6fcc0a104ada20327c90041c7bd51a2f270f9f17a5d2df8bdfc22c5a7749b350
|
| 3 |
+
size 14244
|
checkpoint-1000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:237ac36dfcc4bfdcf176a1c31f3509c37a224c58d3ae5450f7c45403b73609aa
|
| 3 |
+
size 1064
|
checkpoint-1000/trainer_state.json
ADDED
|
@@ -0,0 +1,741 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.949317738791423,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.01949317738791423,
|
| 13 |
+
"grad_norm": 9.292141914367676,
|
| 14 |
+
"learning_rate": 4.967511371020143e-05,
|
| 15 |
+
"loss": 6.2548,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.03898635477582846,
|
| 20 |
+
"grad_norm": 7.766495227813721,
|
| 21 |
+
"learning_rate": 4.935022742040286e-05,
|
| 22 |
+
"loss": 5.0175,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.05847953216374269,
|
| 27 |
+
"grad_norm": 6.4375128746032715,
|
| 28 |
+
"learning_rate": 4.902534113060429e-05,
|
| 29 |
+
"loss": 4.59,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.07797270955165692,
|
| 34 |
+
"grad_norm": 7.395759105682373,
|
| 35 |
+
"learning_rate": 4.8700454840805724e-05,
|
| 36 |
+
"loss": 4.2891,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.09746588693957114,
|
| 41 |
+
"grad_norm": 5.542599678039551,
|
| 42 |
+
"learning_rate": 4.837556855100715e-05,
|
| 43 |
+
"loss": 4.1904,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.11695906432748537,
|
| 48 |
+
"grad_norm": 8.749999046325684,
|
| 49 |
+
"learning_rate": 4.805068226120858e-05,
|
| 50 |
+
"loss": 4.1512,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.1364522417153996,
|
| 55 |
+
"grad_norm": 6.6702446937561035,
|
| 56 |
+
"learning_rate": 4.772579597141001e-05,
|
| 57 |
+
"loss": 3.9695,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.15594541910331383,
|
| 62 |
+
"grad_norm": 13.8934326171875,
|
| 63 |
+
"learning_rate": 4.740090968161144e-05,
|
| 64 |
+
"loss": 3.9657,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.17543859649122806,
|
| 69 |
+
"grad_norm": 15.161931037902832,
|
| 70 |
+
"learning_rate": 4.707602339181287e-05,
|
| 71 |
+
"loss": 3.9312,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.1949317738791423,
|
| 76 |
+
"grad_norm": 8.096311569213867,
|
| 77 |
+
"learning_rate": 4.67511371020143e-05,
|
| 78 |
+
"loss": 3.9407,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.21442495126705652,
|
| 83 |
+
"grad_norm": 6.077565670013428,
|
| 84 |
+
"learning_rate": 4.642625081221573e-05,
|
| 85 |
+
"loss": 3.7438,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.23391812865497075,
|
| 90 |
+
"grad_norm": 5.909548282623291,
|
| 91 |
+
"learning_rate": 4.610136452241716e-05,
|
| 92 |
+
"loss": 3.8495,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.253411306042885,
|
| 97 |
+
"grad_norm": 6.919253349304199,
|
| 98 |
+
"learning_rate": 4.577647823261859e-05,
|
| 99 |
+
"loss": 3.8178,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.2729044834307992,
|
| 104 |
+
"grad_norm": 5.3416218757629395,
|
| 105 |
+
"learning_rate": 4.545159194282001e-05,
|
| 106 |
+
"loss": 3.6388,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.29239766081871343,
|
| 111 |
+
"grad_norm": 6.484707832336426,
|
| 112 |
+
"learning_rate": 4.512670565302144e-05,
|
| 113 |
+
"loss": 3.8361,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.31189083820662766,
|
| 118 |
+
"grad_norm": 6.068100929260254,
|
| 119 |
+
"learning_rate": 4.480181936322287e-05,
|
| 120 |
+
"loss": 3.7164,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.3313840155945419,
|
| 125 |
+
"grad_norm": 6.298447608947754,
|
| 126 |
+
"learning_rate": 4.4476933073424304e-05,
|
| 127 |
+
"loss": 3.6589,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.3508771929824561,
|
| 132 |
+
"grad_norm": 4.752441883087158,
|
| 133 |
+
"learning_rate": 4.4152046783625734e-05,
|
| 134 |
+
"loss": 3.7328,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.37037037037037035,
|
| 139 |
+
"grad_norm": 5.050961971282959,
|
| 140 |
+
"learning_rate": 4.3827160493827164e-05,
|
| 141 |
+
"loss": 3.7162,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.3898635477582846,
|
| 146 |
+
"grad_norm": 4.73190975189209,
|
| 147 |
+
"learning_rate": 4.3502274204028594e-05,
|
| 148 |
+
"loss": 3.6748,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.4093567251461988,
|
| 153 |
+
"grad_norm": 10.872115135192871,
|
| 154 |
+
"learning_rate": 4.3177387914230025e-05,
|
| 155 |
+
"loss": 3.7618,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.42884990253411304,
|
| 160 |
+
"grad_norm": 6.119602680206299,
|
| 161 |
+
"learning_rate": 4.285250162443145e-05,
|
| 162 |
+
"loss": 3.747,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.44834307992202727,
|
| 167 |
+
"grad_norm": 5.332015514373779,
|
| 168 |
+
"learning_rate": 4.252761533463288e-05,
|
| 169 |
+
"loss": 3.6214,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.4678362573099415,
|
| 174 |
+
"grad_norm": 5.08021354675293,
|
| 175 |
+
"learning_rate": 4.220272904483431e-05,
|
| 176 |
+
"loss": 3.5122,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.4873294346978557,
|
| 181 |
+
"grad_norm": 4.609364986419678,
|
| 182 |
+
"learning_rate": 4.187784275503574e-05,
|
| 183 |
+
"loss": 3.5281,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.50682261208577,
|
| 188 |
+
"grad_norm": 24.73488998413086,
|
| 189 |
+
"learning_rate": 4.155295646523717e-05,
|
| 190 |
+
"loss": 3.7142,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.5263157894736842,
|
| 195 |
+
"grad_norm": 5.869173526763916,
|
| 196 |
+
"learning_rate": 4.12280701754386e-05,
|
| 197 |
+
"loss": 3.577,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.5458089668615984,
|
| 202 |
+
"grad_norm": 5.6849470138549805,
|
| 203 |
+
"learning_rate": 4.090318388564003e-05,
|
| 204 |
+
"loss": 3.6146,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.5653021442495126,
|
| 209 |
+
"grad_norm": 5.919517517089844,
|
| 210 |
+
"learning_rate": 4.057829759584146e-05,
|
| 211 |
+
"loss": 3.5417,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.5847953216374269,
|
| 216 |
+
"grad_norm": 5.177787780761719,
|
| 217 |
+
"learning_rate": 4.025341130604289e-05,
|
| 218 |
+
"loss": 3.5453,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.6042884990253411,
|
| 223 |
+
"grad_norm": 5.569265842437744,
|
| 224 |
+
"learning_rate": 3.9928525016244314e-05,
|
| 225 |
+
"loss": 3.5203,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.6237816764132553,
|
| 230 |
+
"grad_norm": 6.63994026184082,
|
| 231 |
+
"learning_rate": 3.9603638726445744e-05,
|
| 232 |
+
"loss": 3.5504,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.6432748538011696,
|
| 237 |
+
"grad_norm": 5.785546779632568,
|
| 238 |
+
"learning_rate": 3.9278752436647174e-05,
|
| 239 |
+
"loss": 3.5587,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.6627680311890838,
|
| 244 |
+
"grad_norm": 5.489815711975098,
|
| 245 |
+
"learning_rate": 3.8953866146848605e-05,
|
| 246 |
+
"loss": 3.5612,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.682261208576998,
|
| 251 |
+
"grad_norm": 4.650988578796387,
|
| 252 |
+
"learning_rate": 3.8628979857050035e-05,
|
| 253 |
+
"loss": 3.5062,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.7017543859649122,
|
| 258 |
+
"grad_norm": 5.7823872566223145,
|
| 259 |
+
"learning_rate": 3.8304093567251465e-05,
|
| 260 |
+
"loss": 3.5819,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.7212475633528265,
|
| 265 |
+
"grad_norm": 7.364371299743652,
|
| 266 |
+
"learning_rate": 3.7979207277452896e-05,
|
| 267 |
+
"loss": 3.5845,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.7407407407407407,
|
| 272 |
+
"grad_norm": 6.099256992340088,
|
| 273 |
+
"learning_rate": 3.7654320987654326e-05,
|
| 274 |
+
"loss": 3.5787,
|
| 275 |
+
"step": 380
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.7602339181286549,
|
| 279 |
+
"grad_norm": 5.115094184875488,
|
| 280 |
+
"learning_rate": 3.7329434697855756e-05,
|
| 281 |
+
"loss": 3.5901,
|
| 282 |
+
"step": 390
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.7797270955165692,
|
| 286 |
+
"grad_norm": 5.4442830085754395,
|
| 287 |
+
"learning_rate": 3.700454840805718e-05,
|
| 288 |
+
"loss": 3.4565,
|
| 289 |
+
"step": 400
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.7992202729044834,
|
| 293 |
+
"grad_norm": 6.6567816734313965,
|
| 294 |
+
"learning_rate": 3.667966211825861e-05,
|
| 295 |
+
"loss": 3.5481,
|
| 296 |
+
"step": 410
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.8187134502923976,
|
| 300 |
+
"grad_norm": 5.831213474273682,
|
| 301 |
+
"learning_rate": 3.635477582846004e-05,
|
| 302 |
+
"loss": 3.4274,
|
| 303 |
+
"step": 420
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.8382066276803118,
|
| 307 |
+
"grad_norm": 5.754021167755127,
|
| 308 |
+
"learning_rate": 3.602988953866147e-05,
|
| 309 |
+
"loss": 3.423,
|
| 310 |
+
"step": 430
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.8576998050682261,
|
| 314 |
+
"grad_norm": 5.590385437011719,
|
| 315 |
+
"learning_rate": 3.57050032488629e-05,
|
| 316 |
+
"loss": 3.6208,
|
| 317 |
+
"step": 440
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.8771929824561403,
|
| 321 |
+
"grad_norm": 4.159304141998291,
|
| 322 |
+
"learning_rate": 3.538011695906433e-05,
|
| 323 |
+
"loss": 3.4107,
|
| 324 |
+
"step": 450
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.8966861598440545,
|
| 328 |
+
"grad_norm": 5.699474334716797,
|
| 329 |
+
"learning_rate": 3.505523066926576e-05,
|
| 330 |
+
"loss": 3.3759,
|
| 331 |
+
"step": 460
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.9161793372319688,
|
| 335 |
+
"grad_norm": 5.0284929275512695,
|
| 336 |
+
"learning_rate": 3.473034437946719e-05,
|
| 337 |
+
"loss": 3.4719,
|
| 338 |
+
"step": 470
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.935672514619883,
|
| 342 |
+
"grad_norm": 4.884040832519531,
|
| 343 |
+
"learning_rate": 3.440545808966862e-05,
|
| 344 |
+
"loss": 3.5534,
|
| 345 |
+
"step": 480
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.9551656920077972,
|
| 349 |
+
"grad_norm": 5.428916931152344,
|
| 350 |
+
"learning_rate": 3.4080571799870045e-05,
|
| 351 |
+
"loss": 3.3919,
|
| 352 |
+
"step": 490
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.9746588693957114,
|
| 356 |
+
"grad_norm": 5.549111843109131,
|
| 357 |
+
"learning_rate": 3.3755685510071476e-05,
|
| 358 |
+
"loss": 3.2799,
|
| 359 |
+
"step": 500
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.9941520467836257,
|
| 363 |
+
"grad_norm": 5.919485092163086,
|
| 364 |
+
"learning_rate": 3.3430799220272906e-05,
|
| 365 |
+
"loss": 3.6097,
|
| 366 |
+
"step": 510
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 1.0,
|
| 370 |
+
"eval_loss": 3.2824156284332275,
|
| 371 |
+
"eval_runtime": 21.5834,
|
| 372 |
+
"eval_samples_per_second": 21.127,
|
| 373 |
+
"eval_steps_per_second": 2.641,
|
| 374 |
+
"step": 513
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 1.01364522417154,
|
| 378 |
+
"grad_norm": 4.9640631675720215,
|
| 379 |
+
"learning_rate": 3.3105912930474336e-05,
|
| 380 |
+
"loss": 3.2341,
|
| 381 |
+
"step": 520
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 1.0331384015594542,
|
| 385 |
+
"grad_norm": 5.887562274932861,
|
| 386 |
+
"learning_rate": 3.2781026640675766e-05,
|
| 387 |
+
"loss": 3.0186,
|
| 388 |
+
"step": 530
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 1.0526315789473684,
|
| 392 |
+
"grad_norm": 7.111718654632568,
|
| 393 |
+
"learning_rate": 3.24561403508772e-05,
|
| 394 |
+
"loss": 3.0381,
|
| 395 |
+
"step": 540
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 1.0721247563352827,
|
| 399 |
+
"grad_norm": 6.662885665893555,
|
| 400 |
+
"learning_rate": 3.213125406107863e-05,
|
| 401 |
+
"loss": 3.0068,
|
| 402 |
+
"step": 550
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 1.0916179337231968,
|
| 406 |
+
"grad_norm": 5.402489185333252,
|
| 407 |
+
"learning_rate": 3.180636777128006e-05,
|
| 408 |
+
"loss": 3.0492,
|
| 409 |
+
"step": 560
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 1.1111111111111112,
|
| 413 |
+
"grad_norm": 6.615454196929932,
|
| 414 |
+
"learning_rate": 3.148148148148148e-05,
|
| 415 |
+
"loss": 2.9823,
|
| 416 |
+
"step": 570
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 1.1306042884990253,
|
| 420 |
+
"grad_norm": 6.344214916229248,
|
| 421 |
+
"learning_rate": 3.115659519168291e-05,
|
| 422 |
+
"loss": 2.863,
|
| 423 |
+
"step": 580
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 1.1500974658869396,
|
| 427 |
+
"grad_norm": 6.1259589195251465,
|
| 428 |
+
"learning_rate": 3.083170890188434e-05,
|
| 429 |
+
"loss": 2.9253,
|
| 430 |
+
"step": 590
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 1.1695906432748537,
|
| 434 |
+
"grad_norm": 5.922159671783447,
|
| 435 |
+
"learning_rate": 3.050682261208577e-05,
|
| 436 |
+
"loss": 3.1142,
|
| 437 |
+
"step": 600
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 1.189083820662768,
|
| 441 |
+
"grad_norm": 4.965208530426025,
|
| 442 |
+
"learning_rate": 3.0181936322287202e-05,
|
| 443 |
+
"loss": 3.0236,
|
| 444 |
+
"step": 610
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 1.2085769980506822,
|
| 448 |
+
"grad_norm": 4.726465225219727,
|
| 449 |
+
"learning_rate": 2.9857050032488632e-05,
|
| 450 |
+
"loss": 3.114,
|
| 451 |
+
"step": 620
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 1.2280701754385965,
|
| 455 |
+
"grad_norm": 5.802026748657227,
|
| 456 |
+
"learning_rate": 2.9532163742690062e-05,
|
| 457 |
+
"loss": 2.9944,
|
| 458 |
+
"step": 630
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 1.2475633528265107,
|
| 462 |
+
"grad_norm": 5.344657897949219,
|
| 463 |
+
"learning_rate": 2.920727745289149e-05,
|
| 464 |
+
"loss": 3.024,
|
| 465 |
+
"step": 640
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 1.267056530214425,
|
| 469 |
+
"grad_norm": 6.377889633178711,
|
| 470 |
+
"learning_rate": 2.888239116309292e-05,
|
| 471 |
+
"loss": 3.1357,
|
| 472 |
+
"step": 650
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 1.286549707602339,
|
| 476 |
+
"grad_norm": 5.2032694816589355,
|
| 477 |
+
"learning_rate": 2.855750487329435e-05,
|
| 478 |
+
"loss": 3.0854,
|
| 479 |
+
"step": 660
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 1.3060428849902534,
|
| 483 |
+
"grad_norm": 5.197242736816406,
|
| 484 |
+
"learning_rate": 2.823261858349578e-05,
|
| 485 |
+
"loss": 3.1208,
|
| 486 |
+
"step": 670
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 1.3255360623781676,
|
| 490 |
+
"grad_norm": 5.073383808135986,
|
| 491 |
+
"learning_rate": 2.7907732293697207e-05,
|
| 492 |
+
"loss": 2.9982,
|
| 493 |
+
"step": 680
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 1.345029239766082,
|
| 497 |
+
"grad_norm": 5.227969646453857,
|
| 498 |
+
"learning_rate": 2.7582846003898637e-05,
|
| 499 |
+
"loss": 3.0834,
|
| 500 |
+
"step": 690
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 1.364522417153996,
|
| 504 |
+
"grad_norm": 5.831137657165527,
|
| 505 |
+
"learning_rate": 2.7257959714100068e-05,
|
| 506 |
+
"loss": 3.0019,
|
| 507 |
+
"step": 700
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 1.3840155945419104,
|
| 511 |
+
"grad_norm": 5.306942939758301,
|
| 512 |
+
"learning_rate": 2.6933073424301498e-05,
|
| 513 |
+
"loss": 2.9489,
|
| 514 |
+
"step": 710
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 1.4035087719298245,
|
| 518 |
+
"grad_norm": 5.346749305725098,
|
| 519 |
+
"learning_rate": 2.6608187134502928e-05,
|
| 520 |
+
"loss": 3.0281,
|
| 521 |
+
"step": 720
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 1.4230019493177388,
|
| 525 |
+
"grad_norm": 4.3056817054748535,
|
| 526 |
+
"learning_rate": 2.6283300844704355e-05,
|
| 527 |
+
"loss": 2.8552,
|
| 528 |
+
"step": 730
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 1.442495126705653,
|
| 532 |
+
"grad_norm": 8.335524559020996,
|
| 533 |
+
"learning_rate": 2.5958414554905785e-05,
|
| 534 |
+
"loss": 2.9833,
|
| 535 |
+
"step": 740
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 1.4619883040935673,
|
| 539 |
+
"grad_norm": 4.510199069976807,
|
| 540 |
+
"learning_rate": 2.5633528265107216e-05,
|
| 541 |
+
"loss": 3.01,
|
| 542 |
+
"step": 750
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 1.4814814814814814,
|
| 546 |
+
"grad_norm": 4.329857349395752,
|
| 547 |
+
"learning_rate": 2.5308641975308646e-05,
|
| 548 |
+
"loss": 3.0841,
|
| 549 |
+
"step": 760
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 1.5009746588693957,
|
| 553 |
+
"grad_norm": 4.561777114868164,
|
| 554 |
+
"learning_rate": 2.4983755685510073e-05,
|
| 555 |
+
"loss": 2.9037,
|
| 556 |
+
"step": 770
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 1.52046783625731,
|
| 560 |
+
"grad_norm": 4.895258903503418,
|
| 561 |
+
"learning_rate": 2.46588693957115e-05,
|
| 562 |
+
"loss": 2.9441,
|
| 563 |
+
"step": 780
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 1.5399610136452242,
|
| 567 |
+
"grad_norm": 4.836617946624756,
|
| 568 |
+
"learning_rate": 2.433398310591293e-05,
|
| 569 |
+
"loss": 2.9684,
|
| 570 |
+
"step": 790
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 1.5594541910331383,
|
| 574 |
+
"grad_norm": 5.523194789886475,
|
| 575 |
+
"learning_rate": 2.400909681611436e-05,
|
| 576 |
+
"loss": 3.0216,
|
| 577 |
+
"step": 800
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 1.5789473684210527,
|
| 581 |
+
"grad_norm": 6.470391750335693,
|
| 582 |
+
"learning_rate": 2.368421052631579e-05,
|
| 583 |
+
"loss": 2.9688,
|
| 584 |
+
"step": 810
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 1.598440545808967,
|
| 588 |
+
"grad_norm": 5.374258995056152,
|
| 589 |
+
"learning_rate": 2.3359324236517217e-05,
|
| 590 |
+
"loss": 2.934,
|
| 591 |
+
"step": 820
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 1.6179337231968811,
|
| 595 |
+
"grad_norm": 4.702390670776367,
|
| 596 |
+
"learning_rate": 2.3034437946718648e-05,
|
| 597 |
+
"loss": 2.8229,
|
| 598 |
+
"step": 830
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 1.6374269005847952,
|
| 602 |
+
"grad_norm": 4.931224822998047,
|
| 603 |
+
"learning_rate": 2.2709551656920078e-05,
|
| 604 |
+
"loss": 2.9748,
|
| 605 |
+
"step": 840
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 1.6569200779727096,
|
| 609 |
+
"grad_norm": 5.478194713592529,
|
| 610 |
+
"learning_rate": 2.2384665367121508e-05,
|
| 611 |
+
"loss": 2.8791,
|
| 612 |
+
"step": 850
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 1.676413255360624,
|
| 616 |
+
"grad_norm": 4.936557769775391,
|
| 617 |
+
"learning_rate": 2.205977907732294e-05,
|
| 618 |
+
"loss": 2.9022,
|
| 619 |
+
"step": 860
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 1.695906432748538,
|
| 623 |
+
"grad_norm": 5.581218242645264,
|
| 624 |
+
"learning_rate": 2.1734892787524365e-05,
|
| 625 |
+
"loss": 3.0293,
|
| 626 |
+
"step": 870
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 1.7153996101364521,
|
| 630 |
+
"grad_norm": 6.841440677642822,
|
| 631 |
+
"learning_rate": 2.1410006497725796e-05,
|
| 632 |
+
"loss": 2.9343,
|
| 633 |
+
"step": 880
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 1.7348927875243665,
|
| 637 |
+
"grad_norm": 5.218410968780518,
|
| 638 |
+
"learning_rate": 2.1085120207927226e-05,
|
| 639 |
+
"loss": 3.0231,
|
| 640 |
+
"step": 890
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 1.7543859649122808,
|
| 644 |
+
"grad_norm": 5.5271077156066895,
|
| 645 |
+
"learning_rate": 2.0760233918128656e-05,
|
| 646 |
+
"loss": 2.8563,
|
| 647 |
+
"step": 900
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 1.773879142300195,
|
| 651 |
+
"grad_norm": 5.889744758605957,
|
| 652 |
+
"learning_rate": 2.0435347628330083e-05,
|
| 653 |
+
"loss": 3.0004,
|
| 654 |
+
"step": 910
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 1.793372319688109,
|
| 658 |
+
"grad_norm": 5.243134498596191,
|
| 659 |
+
"learning_rate": 2.0110461338531513e-05,
|
| 660 |
+
"loss": 2.7516,
|
| 661 |
+
"step": 920
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 1.8128654970760234,
|
| 665 |
+
"grad_norm": 5.98087739944458,
|
| 666 |
+
"learning_rate": 1.9785575048732944e-05,
|
| 667 |
+
"loss": 3.0529,
|
| 668 |
+
"step": 930
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 1.8323586744639377,
|
| 672 |
+
"grad_norm": 4.677685737609863,
|
| 673 |
+
"learning_rate": 1.9460688758934374e-05,
|
| 674 |
+
"loss": 2.8891,
|
| 675 |
+
"step": 940
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 1.8518518518518519,
|
| 679 |
+
"grad_norm": 6.311639308929443,
|
| 680 |
+
"learning_rate": 1.91358024691358e-05,
|
| 681 |
+
"loss": 2.9617,
|
| 682 |
+
"step": 950
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 1.871345029239766,
|
| 686 |
+
"grad_norm": 4.64074182510376,
|
| 687 |
+
"learning_rate": 1.881091617933723e-05,
|
| 688 |
+
"loss": 2.9111,
|
| 689 |
+
"step": 960
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 1.8908382066276803,
|
| 693 |
+
"grad_norm": 4.606050491333008,
|
| 694 |
+
"learning_rate": 1.848602988953866e-05,
|
| 695 |
+
"loss": 2.8621,
|
| 696 |
+
"step": 970
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 1.9103313840155947,
|
| 700 |
+
"grad_norm": 4.810323238372803,
|
| 701 |
+
"learning_rate": 1.816114359974009e-05,
|
| 702 |
+
"loss": 3.017,
|
| 703 |
+
"step": 980
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 1.9298245614035088,
|
| 707 |
+
"grad_norm": 5.766260147094727,
|
| 708 |
+
"learning_rate": 1.7836257309941522e-05,
|
| 709 |
+
"loss": 2.8715,
|
| 710 |
+
"step": 990
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 1.949317738791423,
|
| 714 |
+
"grad_norm": 5.515121936798096,
|
| 715 |
+
"learning_rate": 1.751137102014295e-05,
|
| 716 |
+
"loss": 2.9338,
|
| 717 |
+
"step": 1000
|
| 718 |
+
}
|
| 719 |
+
],
|
| 720 |
+
"logging_steps": 10,
|
| 721 |
+
"max_steps": 1539,
|
| 722 |
+
"num_input_tokens_seen": 0,
|
| 723 |
+
"num_train_epochs": 3,
|
| 724 |
+
"save_steps": 500,
|
| 725 |
+
"stateful_callbacks": {
|
| 726 |
+
"TrainerControl": {
|
| 727 |
+
"args": {
|
| 728 |
+
"should_epoch_stop": false,
|
| 729 |
+
"should_evaluate": false,
|
| 730 |
+
"should_log": false,
|
| 731 |
+
"should_save": true,
|
| 732 |
+
"should_training_stop": false
|
| 733 |
+
},
|
| 734 |
+
"attributes": {}
|
| 735 |
+
}
|
| 736 |
+
},
|
| 737 |
+
"total_flos": 2089552379904000.0,
|
| 738 |
+
"train_batch_size": 8,
|
| 739 |
+
"trial_name": null,
|
| 740 |
+
"trial_params": null
|
| 741 |
+
}
|
checkpoint-1000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6b0e13a27dc9eb93f3cc3f0f69eeb907a00fef0cf3de646ea9710a3354ece317
|
| 3 |
+
size 5304
|
checkpoint-1500/config.json
ADDED
|
@@ -0,0 +1,41 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "ai-forever/rugpt3small_based_on_gpt2",
|
| 3 |
+
"activation_function": "gelu_new",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.1,
|
| 8 |
+
"bos_token_id": 1,
|
| 9 |
+
"embd_pdrop": 0.1,
|
| 10 |
+
"eos_token_id": 2,
|
| 11 |
+
"gradient_checkpointing": false,
|
| 12 |
+
"id2label": {
|
| 13 |
+
"0": "LABEL_0"
|
| 14 |
+
},
|
| 15 |
+
"initializer_range": 0.02,
|
| 16 |
+
"label2id": {
|
| 17 |
+
"LABEL_0": 0
|
| 18 |
+
},
|
| 19 |
+
"layer_norm_epsilon": 1e-05,
|
| 20 |
+
"model_type": "gpt2",
|
| 21 |
+
"n_ctx": 2048,
|
| 22 |
+
"n_embd": 768,
|
| 23 |
+
"n_head": 12,
|
| 24 |
+
"n_inner": null,
|
| 25 |
+
"n_layer": 12,
|
| 26 |
+
"n_positions": 2048,
|
| 27 |
+
"pad_token_id": 0,
|
| 28 |
+
"reorder_and_upcast_attn": false,
|
| 29 |
+
"resid_pdrop": 0.1,
|
| 30 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 31 |
+
"scale_attn_weights": true,
|
| 32 |
+
"summary_activation": null,
|
| 33 |
+
"summary_first_dropout": 0.1,
|
| 34 |
+
"summary_proj_to_labels": true,
|
| 35 |
+
"summary_type": "cls_index",
|
| 36 |
+
"summary_use_proj": true,
|
| 37 |
+
"torch_dtype": "float32",
|
| 38 |
+
"transformers_version": "4.48.3",
|
| 39 |
+
"use_cache": true,
|
| 40 |
+
"vocab_size": 50264
|
| 41 |
+
}
|
checkpoint-1500/generation_config.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"pad_token_id": 0,
|
| 6 |
+
"transformers_version": "4.48.3"
|
| 7 |
+
}
|
checkpoint-1500/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:99e9343bca3268339c57064038db4cddad508f40b87240a223a047abe0eadd34
|
| 3 |
+
size 500941440
|
checkpoint-1500/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:247e18d6852717ffbf831afa5f29e4e0a492907de654b222679e0afb92d815b4
|
| 3 |
+
size 1001976762
|
checkpoint-1500/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:86696a331a922812ccbcc98c52c578af703f9473565b2011b6bcf72f251c2098
|
| 3 |
+
size 14244
|
checkpoint-1500/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1e64c3f3106a44e240ff059dc11a82abb56687dcdc1c4fd8ab1c2c02f2ebf035
|
| 3 |
+
size 1064
|
checkpoint-1500/trainer_state.json
ADDED
|
@@ -0,0 +1,1099 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.9239766081871346,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1500,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.01949317738791423,
|
| 13 |
+
"grad_norm": 9.292141914367676,
|
| 14 |
+
"learning_rate": 4.967511371020143e-05,
|
| 15 |
+
"loss": 6.2548,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.03898635477582846,
|
| 20 |
+
"grad_norm": 7.766495227813721,
|
| 21 |
+
"learning_rate": 4.935022742040286e-05,
|
| 22 |
+
"loss": 5.0175,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.05847953216374269,
|
| 27 |
+
"grad_norm": 6.4375128746032715,
|
| 28 |
+
"learning_rate": 4.902534113060429e-05,
|
| 29 |
+
"loss": 4.59,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.07797270955165692,
|
| 34 |
+
"grad_norm": 7.395759105682373,
|
| 35 |
+
"learning_rate": 4.8700454840805724e-05,
|
| 36 |
+
"loss": 4.2891,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.09746588693957114,
|
| 41 |
+
"grad_norm": 5.542599678039551,
|
| 42 |
+
"learning_rate": 4.837556855100715e-05,
|
| 43 |
+
"loss": 4.1904,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.11695906432748537,
|
| 48 |
+
"grad_norm": 8.749999046325684,
|
| 49 |
+
"learning_rate": 4.805068226120858e-05,
|
| 50 |
+
"loss": 4.1512,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.1364522417153996,
|
| 55 |
+
"grad_norm": 6.6702446937561035,
|
| 56 |
+
"learning_rate": 4.772579597141001e-05,
|
| 57 |
+
"loss": 3.9695,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.15594541910331383,
|
| 62 |
+
"grad_norm": 13.8934326171875,
|
| 63 |
+
"learning_rate": 4.740090968161144e-05,
|
| 64 |
+
"loss": 3.9657,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.17543859649122806,
|
| 69 |
+
"grad_norm": 15.161931037902832,
|
| 70 |
+
"learning_rate": 4.707602339181287e-05,
|
| 71 |
+
"loss": 3.9312,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.1949317738791423,
|
| 76 |
+
"grad_norm": 8.096311569213867,
|
| 77 |
+
"learning_rate": 4.67511371020143e-05,
|
| 78 |
+
"loss": 3.9407,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.21442495126705652,
|
| 83 |
+
"grad_norm": 6.077565670013428,
|
| 84 |
+
"learning_rate": 4.642625081221573e-05,
|
| 85 |
+
"loss": 3.7438,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.23391812865497075,
|
| 90 |
+
"grad_norm": 5.909548282623291,
|
| 91 |
+
"learning_rate": 4.610136452241716e-05,
|
| 92 |
+
"loss": 3.8495,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.253411306042885,
|
| 97 |
+
"grad_norm": 6.919253349304199,
|
| 98 |
+
"learning_rate": 4.577647823261859e-05,
|
| 99 |
+
"loss": 3.8178,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.2729044834307992,
|
| 104 |
+
"grad_norm": 5.3416218757629395,
|
| 105 |
+
"learning_rate": 4.545159194282001e-05,
|
| 106 |
+
"loss": 3.6388,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.29239766081871343,
|
| 111 |
+
"grad_norm": 6.484707832336426,
|
| 112 |
+
"learning_rate": 4.512670565302144e-05,
|
| 113 |
+
"loss": 3.8361,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.31189083820662766,
|
| 118 |
+
"grad_norm": 6.068100929260254,
|
| 119 |
+
"learning_rate": 4.480181936322287e-05,
|
| 120 |
+
"loss": 3.7164,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.3313840155945419,
|
| 125 |
+
"grad_norm": 6.298447608947754,
|
| 126 |
+
"learning_rate": 4.4476933073424304e-05,
|
| 127 |
+
"loss": 3.6589,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.3508771929824561,
|
| 132 |
+
"grad_norm": 4.752441883087158,
|
| 133 |
+
"learning_rate": 4.4152046783625734e-05,
|
| 134 |
+
"loss": 3.7328,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.37037037037037035,
|
| 139 |
+
"grad_norm": 5.050961971282959,
|
| 140 |
+
"learning_rate": 4.3827160493827164e-05,
|
| 141 |
+
"loss": 3.7162,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.3898635477582846,
|
| 146 |
+
"grad_norm": 4.73190975189209,
|
| 147 |
+
"learning_rate": 4.3502274204028594e-05,
|
| 148 |
+
"loss": 3.6748,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.4093567251461988,
|
| 153 |
+
"grad_norm": 10.872115135192871,
|
| 154 |
+
"learning_rate": 4.3177387914230025e-05,
|
| 155 |
+
"loss": 3.7618,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.42884990253411304,
|
| 160 |
+
"grad_norm": 6.119602680206299,
|
| 161 |
+
"learning_rate": 4.285250162443145e-05,
|
| 162 |
+
"loss": 3.747,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.44834307992202727,
|
| 167 |
+
"grad_norm": 5.332015514373779,
|
| 168 |
+
"learning_rate": 4.252761533463288e-05,
|
| 169 |
+
"loss": 3.6214,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.4678362573099415,
|
| 174 |
+
"grad_norm": 5.08021354675293,
|
| 175 |
+
"learning_rate": 4.220272904483431e-05,
|
| 176 |
+
"loss": 3.5122,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.4873294346978557,
|
| 181 |
+
"grad_norm": 4.609364986419678,
|
| 182 |
+
"learning_rate": 4.187784275503574e-05,
|
| 183 |
+
"loss": 3.5281,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.50682261208577,
|
| 188 |
+
"grad_norm": 24.73488998413086,
|
| 189 |
+
"learning_rate": 4.155295646523717e-05,
|
| 190 |
+
"loss": 3.7142,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.5263157894736842,
|
| 195 |
+
"grad_norm": 5.869173526763916,
|
| 196 |
+
"learning_rate": 4.12280701754386e-05,
|
| 197 |
+
"loss": 3.577,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.5458089668615984,
|
| 202 |
+
"grad_norm": 5.6849470138549805,
|
| 203 |
+
"learning_rate": 4.090318388564003e-05,
|
| 204 |
+
"loss": 3.6146,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.5653021442495126,
|
| 209 |
+
"grad_norm": 5.919517517089844,
|
| 210 |
+
"learning_rate": 4.057829759584146e-05,
|
| 211 |
+
"loss": 3.5417,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.5847953216374269,
|
| 216 |
+
"grad_norm": 5.177787780761719,
|
| 217 |
+
"learning_rate": 4.025341130604289e-05,
|
| 218 |
+
"loss": 3.5453,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.6042884990253411,
|
| 223 |
+
"grad_norm": 5.569265842437744,
|
| 224 |
+
"learning_rate": 3.9928525016244314e-05,
|
| 225 |
+
"loss": 3.5203,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.6237816764132553,
|
| 230 |
+
"grad_norm": 6.63994026184082,
|
| 231 |
+
"learning_rate": 3.9603638726445744e-05,
|
| 232 |
+
"loss": 3.5504,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.6432748538011696,
|
| 237 |
+
"grad_norm": 5.785546779632568,
|
| 238 |
+
"learning_rate": 3.9278752436647174e-05,
|
| 239 |
+
"loss": 3.5587,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.6627680311890838,
|
| 244 |
+
"grad_norm": 5.489815711975098,
|
| 245 |
+
"learning_rate": 3.8953866146848605e-05,
|
| 246 |
+
"loss": 3.5612,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.682261208576998,
|
| 251 |
+
"grad_norm": 4.650988578796387,
|
| 252 |
+
"learning_rate": 3.8628979857050035e-05,
|
| 253 |
+
"loss": 3.5062,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.7017543859649122,
|
| 258 |
+
"grad_norm": 5.7823872566223145,
|
| 259 |
+
"learning_rate": 3.8304093567251465e-05,
|
| 260 |
+
"loss": 3.5819,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.7212475633528265,
|
| 265 |
+
"grad_norm": 7.364371299743652,
|
| 266 |
+
"learning_rate": 3.7979207277452896e-05,
|
| 267 |
+
"loss": 3.5845,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.7407407407407407,
|
| 272 |
+
"grad_norm": 6.099256992340088,
|
| 273 |
+
"learning_rate": 3.7654320987654326e-05,
|
| 274 |
+
"loss": 3.5787,
|
| 275 |
+
"step": 380
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.7602339181286549,
|
| 279 |
+
"grad_norm": 5.115094184875488,
|
| 280 |
+
"learning_rate": 3.7329434697855756e-05,
|
| 281 |
+
"loss": 3.5901,
|
| 282 |
+
"step": 390
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.7797270955165692,
|
| 286 |
+
"grad_norm": 5.4442830085754395,
|
| 287 |
+
"learning_rate": 3.700454840805718e-05,
|
| 288 |
+
"loss": 3.4565,
|
| 289 |
+
"step": 400
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.7992202729044834,
|
| 293 |
+
"grad_norm": 6.6567816734313965,
|
| 294 |
+
"learning_rate": 3.667966211825861e-05,
|
| 295 |
+
"loss": 3.5481,
|
| 296 |
+
"step": 410
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.8187134502923976,
|
| 300 |
+
"grad_norm": 5.831213474273682,
|
| 301 |
+
"learning_rate": 3.635477582846004e-05,
|
| 302 |
+
"loss": 3.4274,
|
| 303 |
+
"step": 420
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.8382066276803118,
|
| 307 |
+
"grad_norm": 5.754021167755127,
|
| 308 |
+
"learning_rate": 3.602988953866147e-05,
|
| 309 |
+
"loss": 3.423,
|
| 310 |
+
"step": 430
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.8576998050682261,
|
| 314 |
+
"grad_norm": 5.590385437011719,
|
| 315 |
+
"learning_rate": 3.57050032488629e-05,
|
| 316 |
+
"loss": 3.6208,
|
| 317 |
+
"step": 440
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.8771929824561403,
|
| 321 |
+
"grad_norm": 4.159304141998291,
|
| 322 |
+
"learning_rate": 3.538011695906433e-05,
|
| 323 |
+
"loss": 3.4107,
|
| 324 |
+
"step": 450
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.8966861598440545,
|
| 328 |
+
"grad_norm": 5.699474334716797,
|
| 329 |
+
"learning_rate": 3.505523066926576e-05,
|
| 330 |
+
"loss": 3.3759,
|
| 331 |
+
"step": 460
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.9161793372319688,
|
| 335 |
+
"grad_norm": 5.0284929275512695,
|
| 336 |
+
"learning_rate": 3.473034437946719e-05,
|
| 337 |
+
"loss": 3.4719,
|
| 338 |
+
"step": 470
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.935672514619883,
|
| 342 |
+
"grad_norm": 4.884040832519531,
|
| 343 |
+
"learning_rate": 3.440545808966862e-05,
|
| 344 |
+
"loss": 3.5534,
|
| 345 |
+
"step": 480
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.9551656920077972,
|
| 349 |
+
"grad_norm": 5.428916931152344,
|
| 350 |
+
"learning_rate": 3.4080571799870045e-05,
|
| 351 |
+
"loss": 3.3919,
|
| 352 |
+
"step": 490
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.9746588693957114,
|
| 356 |
+
"grad_norm": 5.549111843109131,
|
| 357 |
+
"learning_rate": 3.3755685510071476e-05,
|
| 358 |
+
"loss": 3.2799,
|
| 359 |
+
"step": 500
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.9941520467836257,
|
| 363 |
+
"grad_norm": 5.919485092163086,
|
| 364 |
+
"learning_rate": 3.3430799220272906e-05,
|
| 365 |
+
"loss": 3.6097,
|
| 366 |
+
"step": 510
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 1.0,
|
| 370 |
+
"eval_loss": 3.2824156284332275,
|
| 371 |
+
"eval_runtime": 21.5834,
|
| 372 |
+
"eval_samples_per_second": 21.127,
|
| 373 |
+
"eval_steps_per_second": 2.641,
|
| 374 |
+
"step": 513
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 1.01364522417154,
|
| 378 |
+
"grad_norm": 4.9640631675720215,
|
| 379 |
+
"learning_rate": 3.3105912930474336e-05,
|
| 380 |
+
"loss": 3.2341,
|
| 381 |
+
"step": 520
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 1.0331384015594542,
|
| 385 |
+
"grad_norm": 5.887562274932861,
|
| 386 |
+
"learning_rate": 3.2781026640675766e-05,
|
| 387 |
+
"loss": 3.0186,
|
| 388 |
+
"step": 530
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 1.0526315789473684,
|
| 392 |
+
"grad_norm": 7.111718654632568,
|
| 393 |
+
"learning_rate": 3.24561403508772e-05,
|
| 394 |
+
"loss": 3.0381,
|
| 395 |
+
"step": 540
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 1.0721247563352827,
|
| 399 |
+
"grad_norm": 6.662885665893555,
|
| 400 |
+
"learning_rate": 3.213125406107863e-05,
|
| 401 |
+
"loss": 3.0068,
|
| 402 |
+
"step": 550
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 1.0916179337231968,
|
| 406 |
+
"grad_norm": 5.402489185333252,
|
| 407 |
+
"learning_rate": 3.180636777128006e-05,
|
| 408 |
+
"loss": 3.0492,
|
| 409 |
+
"step": 560
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 1.1111111111111112,
|
| 413 |
+
"grad_norm": 6.615454196929932,
|
| 414 |
+
"learning_rate": 3.148148148148148e-05,
|
| 415 |
+
"loss": 2.9823,
|
| 416 |
+
"step": 570
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 1.1306042884990253,
|
| 420 |
+
"grad_norm": 6.344214916229248,
|
| 421 |
+
"learning_rate": 3.115659519168291e-05,
|
| 422 |
+
"loss": 2.863,
|
| 423 |
+
"step": 580
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 1.1500974658869396,
|
| 427 |
+
"grad_norm": 6.1259589195251465,
|
| 428 |
+
"learning_rate": 3.083170890188434e-05,
|
| 429 |
+
"loss": 2.9253,
|
| 430 |
+
"step": 590
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 1.1695906432748537,
|
| 434 |
+
"grad_norm": 5.922159671783447,
|
| 435 |
+
"learning_rate": 3.050682261208577e-05,
|
| 436 |
+
"loss": 3.1142,
|
| 437 |
+
"step": 600
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 1.189083820662768,
|
| 441 |
+
"grad_norm": 4.965208530426025,
|
| 442 |
+
"learning_rate": 3.0181936322287202e-05,
|
| 443 |
+
"loss": 3.0236,
|
| 444 |
+
"step": 610
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 1.2085769980506822,
|
| 448 |
+
"grad_norm": 4.726465225219727,
|
| 449 |
+
"learning_rate": 2.9857050032488632e-05,
|
| 450 |
+
"loss": 3.114,
|
| 451 |
+
"step": 620
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 1.2280701754385965,
|
| 455 |
+
"grad_norm": 5.802026748657227,
|
| 456 |
+
"learning_rate": 2.9532163742690062e-05,
|
| 457 |
+
"loss": 2.9944,
|
| 458 |
+
"step": 630
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 1.2475633528265107,
|
| 462 |
+
"grad_norm": 5.344657897949219,
|
| 463 |
+
"learning_rate": 2.920727745289149e-05,
|
| 464 |
+
"loss": 3.024,
|
| 465 |
+
"step": 640
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 1.267056530214425,
|
| 469 |
+
"grad_norm": 6.377889633178711,
|
| 470 |
+
"learning_rate": 2.888239116309292e-05,
|
| 471 |
+
"loss": 3.1357,
|
| 472 |
+
"step": 650
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 1.286549707602339,
|
| 476 |
+
"grad_norm": 5.2032694816589355,
|
| 477 |
+
"learning_rate": 2.855750487329435e-05,
|
| 478 |
+
"loss": 3.0854,
|
| 479 |
+
"step": 660
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 1.3060428849902534,
|
| 483 |
+
"grad_norm": 5.197242736816406,
|
| 484 |
+
"learning_rate": 2.823261858349578e-05,
|
| 485 |
+
"loss": 3.1208,
|
| 486 |
+
"step": 670
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 1.3255360623781676,
|
| 490 |
+
"grad_norm": 5.073383808135986,
|
| 491 |
+
"learning_rate": 2.7907732293697207e-05,
|
| 492 |
+
"loss": 2.9982,
|
| 493 |
+
"step": 680
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 1.345029239766082,
|
| 497 |
+
"grad_norm": 5.227969646453857,
|
| 498 |
+
"learning_rate": 2.7582846003898637e-05,
|
| 499 |
+
"loss": 3.0834,
|
| 500 |
+
"step": 690
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 1.364522417153996,
|
| 504 |
+
"grad_norm": 5.831137657165527,
|
| 505 |
+
"learning_rate": 2.7257959714100068e-05,
|
| 506 |
+
"loss": 3.0019,
|
| 507 |
+
"step": 700
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 1.3840155945419104,
|
| 511 |
+
"grad_norm": 5.306942939758301,
|
| 512 |
+
"learning_rate": 2.6933073424301498e-05,
|
| 513 |
+
"loss": 2.9489,
|
| 514 |
+
"step": 710
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 1.4035087719298245,
|
| 518 |
+
"grad_norm": 5.346749305725098,
|
| 519 |
+
"learning_rate": 2.6608187134502928e-05,
|
| 520 |
+
"loss": 3.0281,
|
| 521 |
+
"step": 720
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 1.4230019493177388,
|
| 525 |
+
"grad_norm": 4.3056817054748535,
|
| 526 |
+
"learning_rate": 2.6283300844704355e-05,
|
| 527 |
+
"loss": 2.8552,
|
| 528 |
+
"step": 730
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 1.442495126705653,
|
| 532 |
+
"grad_norm": 8.335524559020996,
|
| 533 |
+
"learning_rate": 2.5958414554905785e-05,
|
| 534 |
+
"loss": 2.9833,
|
| 535 |
+
"step": 740
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 1.4619883040935673,
|
| 539 |
+
"grad_norm": 4.510199069976807,
|
| 540 |
+
"learning_rate": 2.5633528265107216e-05,
|
| 541 |
+
"loss": 3.01,
|
| 542 |
+
"step": 750
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 1.4814814814814814,
|
| 546 |
+
"grad_norm": 4.329857349395752,
|
| 547 |
+
"learning_rate": 2.5308641975308646e-05,
|
| 548 |
+
"loss": 3.0841,
|
| 549 |
+
"step": 760
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 1.5009746588693957,
|
| 553 |
+
"grad_norm": 4.561777114868164,
|
| 554 |
+
"learning_rate": 2.4983755685510073e-05,
|
| 555 |
+
"loss": 2.9037,
|
| 556 |
+
"step": 770
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 1.52046783625731,
|
| 560 |
+
"grad_norm": 4.895258903503418,
|
| 561 |
+
"learning_rate": 2.46588693957115e-05,
|
| 562 |
+
"loss": 2.9441,
|
| 563 |
+
"step": 780
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 1.5399610136452242,
|
| 567 |
+
"grad_norm": 4.836617946624756,
|
| 568 |
+
"learning_rate": 2.433398310591293e-05,
|
| 569 |
+
"loss": 2.9684,
|
| 570 |
+
"step": 790
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 1.5594541910331383,
|
| 574 |
+
"grad_norm": 5.523194789886475,
|
| 575 |
+
"learning_rate": 2.400909681611436e-05,
|
| 576 |
+
"loss": 3.0216,
|
| 577 |
+
"step": 800
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 1.5789473684210527,
|
| 581 |
+
"grad_norm": 6.470391750335693,
|
| 582 |
+
"learning_rate": 2.368421052631579e-05,
|
| 583 |
+
"loss": 2.9688,
|
| 584 |
+
"step": 810
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 1.598440545808967,
|
| 588 |
+
"grad_norm": 5.374258995056152,
|
| 589 |
+
"learning_rate": 2.3359324236517217e-05,
|
| 590 |
+
"loss": 2.934,
|
| 591 |
+
"step": 820
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 1.6179337231968811,
|
| 595 |
+
"grad_norm": 4.702390670776367,
|
| 596 |
+
"learning_rate": 2.3034437946718648e-05,
|
| 597 |
+
"loss": 2.8229,
|
| 598 |
+
"step": 830
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 1.6374269005847952,
|
| 602 |
+
"grad_norm": 4.931224822998047,
|
| 603 |
+
"learning_rate": 2.2709551656920078e-05,
|
| 604 |
+
"loss": 2.9748,
|
| 605 |
+
"step": 840
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 1.6569200779727096,
|
| 609 |
+
"grad_norm": 5.478194713592529,
|
| 610 |
+
"learning_rate": 2.2384665367121508e-05,
|
| 611 |
+
"loss": 2.8791,
|
| 612 |
+
"step": 850
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 1.676413255360624,
|
| 616 |
+
"grad_norm": 4.936557769775391,
|
| 617 |
+
"learning_rate": 2.205977907732294e-05,
|
| 618 |
+
"loss": 2.9022,
|
| 619 |
+
"step": 860
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 1.695906432748538,
|
| 623 |
+
"grad_norm": 5.581218242645264,
|
| 624 |
+
"learning_rate": 2.1734892787524365e-05,
|
| 625 |
+
"loss": 3.0293,
|
| 626 |
+
"step": 870
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 1.7153996101364521,
|
| 630 |
+
"grad_norm": 6.841440677642822,
|
| 631 |
+
"learning_rate": 2.1410006497725796e-05,
|
| 632 |
+
"loss": 2.9343,
|
| 633 |
+
"step": 880
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 1.7348927875243665,
|
| 637 |
+
"grad_norm": 5.218410968780518,
|
| 638 |
+
"learning_rate": 2.1085120207927226e-05,
|
| 639 |
+
"loss": 3.0231,
|
| 640 |
+
"step": 890
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 1.7543859649122808,
|
| 644 |
+
"grad_norm": 5.5271077156066895,
|
| 645 |
+
"learning_rate": 2.0760233918128656e-05,
|
| 646 |
+
"loss": 2.8563,
|
| 647 |
+
"step": 900
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 1.773879142300195,
|
| 651 |
+
"grad_norm": 5.889744758605957,
|
| 652 |
+
"learning_rate": 2.0435347628330083e-05,
|
| 653 |
+
"loss": 3.0004,
|
| 654 |
+
"step": 910
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 1.793372319688109,
|
| 658 |
+
"grad_norm": 5.243134498596191,
|
| 659 |
+
"learning_rate": 2.0110461338531513e-05,
|
| 660 |
+
"loss": 2.7516,
|
| 661 |
+
"step": 920
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 1.8128654970760234,
|
| 665 |
+
"grad_norm": 5.98087739944458,
|
| 666 |
+
"learning_rate": 1.9785575048732944e-05,
|
| 667 |
+
"loss": 3.0529,
|
| 668 |
+
"step": 930
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 1.8323586744639377,
|
| 672 |
+
"grad_norm": 4.677685737609863,
|
| 673 |
+
"learning_rate": 1.9460688758934374e-05,
|
| 674 |
+
"loss": 2.8891,
|
| 675 |
+
"step": 940
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 1.8518518518518519,
|
| 679 |
+
"grad_norm": 6.311639308929443,
|
| 680 |
+
"learning_rate": 1.91358024691358e-05,
|
| 681 |
+
"loss": 2.9617,
|
| 682 |
+
"step": 950
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 1.871345029239766,
|
| 686 |
+
"grad_norm": 4.64074182510376,
|
| 687 |
+
"learning_rate": 1.881091617933723e-05,
|
| 688 |
+
"loss": 2.9111,
|
| 689 |
+
"step": 960
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 1.8908382066276803,
|
| 693 |
+
"grad_norm": 4.606050491333008,
|
| 694 |
+
"learning_rate": 1.848602988953866e-05,
|
| 695 |
+
"loss": 2.8621,
|
| 696 |
+
"step": 970
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 1.9103313840155947,
|
| 700 |
+
"grad_norm": 4.810323238372803,
|
| 701 |
+
"learning_rate": 1.816114359974009e-05,
|
| 702 |
+
"loss": 3.017,
|
| 703 |
+
"step": 980
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 1.9298245614035088,
|
| 707 |
+
"grad_norm": 5.766260147094727,
|
| 708 |
+
"learning_rate": 1.7836257309941522e-05,
|
| 709 |
+
"loss": 2.8715,
|
| 710 |
+
"step": 990
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 1.949317738791423,
|
| 714 |
+
"grad_norm": 5.515121936798096,
|
| 715 |
+
"learning_rate": 1.751137102014295e-05,
|
| 716 |
+
"loss": 2.9338,
|
| 717 |
+
"step": 1000
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 1.9688109161793372,
|
| 721 |
+
"grad_norm": 5.856773376464844,
|
| 722 |
+
"learning_rate": 1.718648473034438e-05,
|
| 723 |
+
"loss": 2.9503,
|
| 724 |
+
"step": 1010
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 1.9883040935672516,
|
| 728 |
+
"grad_norm": 6.5240373611450195,
|
| 729 |
+
"learning_rate": 1.686159844054581e-05,
|
| 730 |
+
"loss": 2.9444,
|
| 731 |
+
"step": 1020
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 2.0,
|
| 735 |
+
"eval_loss": 3.203155994415283,
|
| 736 |
+
"eval_runtime": 21.2886,
|
| 737 |
+
"eval_samples_per_second": 21.42,
|
| 738 |
+
"eval_steps_per_second": 2.677,
|
| 739 |
+
"step": 1026
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"epoch": 2.007797270955166,
|
| 743 |
+
"grad_norm": 4.038070201873779,
|
| 744 |
+
"learning_rate": 1.653671215074724e-05,
|
| 745 |
+
"loss": 2.8103,
|
| 746 |
+
"step": 1030
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 2.02729044834308,
|
| 750 |
+
"grad_norm": 5.510829925537109,
|
| 751 |
+
"learning_rate": 1.6211825860948666e-05,
|
| 752 |
+
"loss": 2.6078,
|
| 753 |
+
"step": 1040
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 2.046783625730994,
|
| 757 |
+
"grad_norm": 6.110142707824707,
|
| 758 |
+
"learning_rate": 1.5886939571150097e-05,
|
| 759 |
+
"loss": 2.6043,
|
| 760 |
+
"step": 1050
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 2.0662768031189085,
|
| 764 |
+
"grad_norm": 7.630992412567139,
|
| 765 |
+
"learning_rate": 1.5562053281351527e-05,
|
| 766 |
+
"loss": 2.6449,
|
| 767 |
+
"step": 1060
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 2.0857699805068224,
|
| 771 |
+
"grad_norm": 6.36288595199585,
|
| 772 |
+
"learning_rate": 1.5237166991552956e-05,
|
| 773 |
+
"loss": 2.6561,
|
| 774 |
+
"step": 1070
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 2.1052631578947367,
|
| 778 |
+
"grad_norm": 6.1103363037109375,
|
| 779 |
+
"learning_rate": 1.4912280701754386e-05,
|
| 780 |
+
"loss": 2.615,
|
| 781 |
+
"step": 1080
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 2.124756335282651,
|
| 785 |
+
"grad_norm": 7.310168266296387,
|
| 786 |
+
"learning_rate": 1.4587394411955816e-05,
|
| 787 |
+
"loss": 2.6722,
|
| 788 |
+
"step": 1090
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 2.1442495126705654,
|
| 792 |
+
"grad_norm": 5.365858554840088,
|
| 793 |
+
"learning_rate": 1.4262508122157245e-05,
|
| 794 |
+
"loss": 2.624,
|
| 795 |
+
"step": 1100
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 2.1637426900584797,
|
| 799 |
+
"grad_norm": 5.3909735679626465,
|
| 800 |
+
"learning_rate": 1.3937621832358675e-05,
|
| 801 |
+
"loss": 2.595,
|
| 802 |
+
"step": 1110
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 2.1832358674463936,
|
| 806 |
+
"grad_norm": 6.050861835479736,
|
| 807 |
+
"learning_rate": 1.3612735542560104e-05,
|
| 808 |
+
"loss": 2.5981,
|
| 809 |
+
"step": 1120
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 2.202729044834308,
|
| 813 |
+
"grad_norm": 6.927775859832764,
|
| 814 |
+
"learning_rate": 1.3287849252761534e-05,
|
| 815 |
+
"loss": 2.6114,
|
| 816 |
+
"step": 1130
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 2.2222222222222223,
|
| 820 |
+
"grad_norm": 6.272125244140625,
|
| 821 |
+
"learning_rate": 1.2962962962962962e-05,
|
| 822 |
+
"loss": 2.654,
|
| 823 |
+
"step": 1140
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 2.241715399610136,
|
| 827 |
+
"grad_norm": 5.306847095489502,
|
| 828 |
+
"learning_rate": 1.2638076673164393e-05,
|
| 829 |
+
"loss": 2.7056,
|
| 830 |
+
"step": 1150
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 2.2612085769980506,
|
| 834 |
+
"grad_norm": 5.6090850830078125,
|
| 835 |
+
"learning_rate": 1.2313190383365821e-05,
|
| 836 |
+
"loss": 2.6322,
|
| 837 |
+
"step": 1160
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 2.280701754385965,
|
| 841 |
+
"grad_norm": 4.8249382972717285,
|
| 842 |
+
"learning_rate": 1.1988304093567252e-05,
|
| 843 |
+
"loss": 2.6714,
|
| 844 |
+
"step": 1170
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 2.3001949317738792,
|
| 848 |
+
"grad_norm": 6.645869731903076,
|
| 849 |
+
"learning_rate": 1.166341780376868e-05,
|
| 850 |
+
"loss": 2.7006,
|
| 851 |
+
"step": 1180
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 2.3196881091617936,
|
| 855 |
+
"grad_norm": 5.666390895843506,
|
| 856 |
+
"learning_rate": 1.133853151397011e-05,
|
| 857 |
+
"loss": 2.6726,
|
| 858 |
+
"step": 1190
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 2.3391812865497075,
|
| 862 |
+
"grad_norm": 5.168858528137207,
|
| 863 |
+
"learning_rate": 1.1013645224171539e-05,
|
| 864 |
+
"loss": 2.4944,
|
| 865 |
+
"step": 1200
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 2.358674463937622,
|
| 869 |
+
"grad_norm": 5.98241662979126,
|
| 870 |
+
"learning_rate": 1.068875893437297e-05,
|
| 871 |
+
"loss": 2.6001,
|
| 872 |
+
"step": 1210
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 2.378167641325536,
|
| 876 |
+
"grad_norm": 5.360602855682373,
|
| 877 |
+
"learning_rate": 1.03638726445744e-05,
|
| 878 |
+
"loss": 2.8167,
|
| 879 |
+
"step": 1220
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 2.39766081871345,
|
| 883 |
+
"grad_norm": 6.712647914886475,
|
| 884 |
+
"learning_rate": 1.0038986354775828e-05,
|
| 885 |
+
"loss": 2.6197,
|
| 886 |
+
"step": 1230
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 2.4171539961013644,
|
| 890 |
+
"grad_norm": 5.139374256134033,
|
| 891 |
+
"learning_rate": 9.714100064977258e-06,
|
| 892 |
+
"loss": 2.6276,
|
| 893 |
+
"step": 1240
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 2.4366471734892787,
|
| 897 |
+
"grad_norm": 5.239111423492432,
|
| 898 |
+
"learning_rate": 9.389213775178687e-06,
|
| 899 |
+
"loss": 2.5737,
|
| 900 |
+
"step": 1250
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 2.456140350877193,
|
| 904 |
+
"grad_norm": 5.82625675201416,
|
| 905 |
+
"learning_rate": 9.064327485380117e-06,
|
| 906 |
+
"loss": 2.7264,
|
| 907 |
+
"step": 1260
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 2.4756335282651074,
|
| 911 |
+
"grad_norm": 7.747376441955566,
|
| 912 |
+
"learning_rate": 8.739441195581546e-06,
|
| 913 |
+
"loss": 2.5372,
|
| 914 |
+
"step": 1270
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 2.4951267056530213,
|
| 918 |
+
"grad_norm": 6.986497402191162,
|
| 919 |
+
"learning_rate": 8.414554905782976e-06,
|
| 920 |
+
"loss": 2.6583,
|
| 921 |
+
"step": 1280
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 2.5146198830409356,
|
| 925 |
+
"grad_norm": 5.268970012664795,
|
| 926 |
+
"learning_rate": 8.089668615984405e-06,
|
| 927 |
+
"loss": 2.5864,
|
| 928 |
+
"step": 1290
|
| 929 |
+
},
|
| 930 |
+
{
|
| 931 |
+
"epoch": 2.53411306042885,
|
| 932 |
+
"grad_norm": 6.267032623291016,
|
| 933 |
+
"learning_rate": 7.764782326185835e-06,
|
| 934 |
+
"loss": 2.5389,
|
| 935 |
+
"step": 1300
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 2.553606237816764,
|
| 939 |
+
"grad_norm": 5.6785173416137695,
|
| 940 |
+
"learning_rate": 7.4398960363872645e-06,
|
| 941 |
+
"loss": 2.6147,
|
| 942 |
+
"step": 1310
|
| 943 |
+
},
|
| 944 |
+
{
|
| 945 |
+
"epoch": 2.573099415204678,
|
| 946 |
+
"grad_norm": 7.006417751312256,
|
| 947 |
+
"learning_rate": 7.115009746588694e-06,
|
| 948 |
+
"loss": 2.6028,
|
| 949 |
+
"step": 1320
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 2.5925925925925926,
|
| 953 |
+
"grad_norm": 5.888213634490967,
|
| 954 |
+
"learning_rate": 6.790123456790123e-06,
|
| 955 |
+
"loss": 2.7019,
|
| 956 |
+
"step": 1330
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 2.612085769980507,
|
| 960 |
+
"grad_norm": 6.327919960021973,
|
| 961 |
+
"learning_rate": 6.465237166991553e-06,
|
| 962 |
+
"loss": 2.6776,
|
| 963 |
+
"step": 1340
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 2.6315789473684212,
|
| 967 |
+
"grad_norm": 5.6416096687316895,
|
| 968 |
+
"learning_rate": 6.140350877192982e-06,
|
| 969 |
+
"loss": 2.6533,
|
| 970 |
+
"step": 1350
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 2.651072124756335,
|
| 974 |
+
"grad_norm": 6.070097923278809,
|
| 975 |
+
"learning_rate": 5.815464587394412e-06,
|
| 976 |
+
"loss": 2.674,
|
| 977 |
+
"step": 1360
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 2.6705653021442495,
|
| 981 |
+
"grad_norm": 5.943949222564697,
|
| 982 |
+
"learning_rate": 5.490578297595841e-06,
|
| 983 |
+
"loss": 2.5782,
|
| 984 |
+
"step": 1370
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 2.690058479532164,
|
| 988 |
+
"grad_norm": 5.129433631896973,
|
| 989 |
+
"learning_rate": 5.165692007797271e-06,
|
| 990 |
+
"loss": 2.7243,
|
| 991 |
+
"step": 1380
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 2.7095516569200777,
|
| 995 |
+
"grad_norm": 6.323672771453857,
|
| 996 |
+
"learning_rate": 4.840805717998701e-06,
|
| 997 |
+
"loss": 2.7187,
|
| 998 |
+
"step": 1390
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 2.729044834307992,
|
| 1002 |
+
"grad_norm": 4.658057689666748,
|
| 1003 |
+
"learning_rate": 4.51591942820013e-06,
|
| 1004 |
+
"loss": 2.6158,
|
| 1005 |
+
"step": 1400
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 2.7485380116959064,
|
| 1009 |
+
"grad_norm": 5.387426376342773,
|
| 1010 |
+
"learning_rate": 4.19103313840156e-06,
|
| 1011 |
+
"loss": 2.5866,
|
| 1012 |
+
"step": 1410
|
| 1013 |
+
},
|
| 1014 |
+
{
|
| 1015 |
+
"epoch": 2.7680311890838207,
|
| 1016 |
+
"grad_norm": 6.695394992828369,
|
| 1017 |
+
"learning_rate": 3.866146848602989e-06,
|
| 1018 |
+
"loss": 2.5415,
|
| 1019 |
+
"step": 1420
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 2.787524366471735,
|
| 1023 |
+
"grad_norm": 5.209081172943115,
|
| 1024 |
+
"learning_rate": 3.5412605588044185e-06,
|
| 1025 |
+
"loss": 2.4233,
|
| 1026 |
+
"step": 1430
|
| 1027 |
+
},
|
| 1028 |
+
{
|
| 1029 |
+
"epoch": 2.807017543859649,
|
| 1030 |
+
"grad_norm": 4.750288009643555,
|
| 1031 |
+
"learning_rate": 3.216374269005848e-06,
|
| 1032 |
+
"loss": 2.5685,
|
| 1033 |
+
"step": 1440
|
| 1034 |
+
},
|
| 1035 |
+
{
|
| 1036 |
+
"epoch": 2.8265107212475633,
|
| 1037 |
+
"grad_norm": 5.388522148132324,
|
| 1038 |
+
"learning_rate": 2.8914879792072773e-06,
|
| 1039 |
+
"loss": 2.6053,
|
| 1040 |
+
"step": 1450
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 2.8460038986354776,
|
| 1044 |
+
"grad_norm": 6.151064395904541,
|
| 1045 |
+
"learning_rate": 2.5666016894087067e-06,
|
| 1046 |
+
"loss": 2.6842,
|
| 1047 |
+
"step": 1460
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 2.8654970760233915,
|
| 1051 |
+
"grad_norm": 6.249871253967285,
|
| 1052 |
+
"learning_rate": 2.2417153996101366e-06,
|
| 1053 |
+
"loss": 2.505,
|
| 1054 |
+
"step": 1470
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"epoch": 2.884990253411306,
|
| 1058 |
+
"grad_norm": 7.69942045211792,
|
| 1059 |
+
"learning_rate": 1.916829109811566e-06,
|
| 1060 |
+
"loss": 2.6107,
|
| 1061 |
+
"step": 1480
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"epoch": 2.9044834307992202,
|
| 1065 |
+
"grad_norm": 5.715646266937256,
|
| 1066 |
+
"learning_rate": 1.5919428200129955e-06,
|
| 1067 |
+
"loss": 2.6015,
|
| 1068 |
+
"step": 1490
|
| 1069 |
+
},
|
| 1070 |
+
{
|
| 1071 |
+
"epoch": 2.9239766081871346,
|
| 1072 |
+
"grad_norm": 5.447291851043701,
|
| 1073 |
+
"learning_rate": 1.2670565302144249e-06,
|
| 1074 |
+
"loss": 2.7594,
|
| 1075 |
+
"step": 1500
|
| 1076 |
+
}
|
| 1077 |
+
],
|
| 1078 |
+
"logging_steps": 10,
|
| 1079 |
+
"max_steps": 1539,
|
| 1080 |
+
"num_input_tokens_seen": 0,
|
| 1081 |
+
"num_train_epochs": 3,
|
| 1082 |
+
"save_steps": 500,
|
| 1083 |
+
"stateful_callbacks": {
|
| 1084 |
+
"TrainerControl": {
|
| 1085 |
+
"args": {
|
| 1086 |
+
"should_epoch_stop": false,
|
| 1087 |
+
"should_evaluate": false,
|
| 1088 |
+
"should_log": false,
|
| 1089 |
+
"should_save": true,
|
| 1090 |
+
"should_training_stop": false
|
| 1091 |
+
},
|
| 1092 |
+
"attributes": {}
|
| 1093 |
+
}
|
| 1094 |
+
},
|
| 1095 |
+
"total_flos": 3133936631808000.0,
|
| 1096 |
+
"train_batch_size": 8,
|
| 1097 |
+
"trial_name": null,
|
| 1098 |
+
"trial_params": null
|
| 1099 |
+
}
|
checkpoint-1500/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6b0e13a27dc9eb93f3cc3f0f69eeb907a00fef0cf3de646ea9710a3354ece317
|
| 3 |
+
size 5304
|
checkpoint-1539/config.json
ADDED
|
@@ -0,0 +1,41 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "ai-forever/rugpt3small_based_on_gpt2",
|
| 3 |
+
"activation_function": "gelu_new",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.1,
|
| 8 |
+
"bos_token_id": 1,
|
| 9 |
+
"embd_pdrop": 0.1,
|
| 10 |
+
"eos_token_id": 2,
|
| 11 |
+
"gradient_checkpointing": false,
|
| 12 |
+
"id2label": {
|
| 13 |
+
"0": "LABEL_0"
|
| 14 |
+
},
|
| 15 |
+
"initializer_range": 0.02,
|
| 16 |
+
"label2id": {
|
| 17 |
+
"LABEL_0": 0
|
| 18 |
+
},
|
| 19 |
+
"layer_norm_epsilon": 1e-05,
|
| 20 |
+
"model_type": "gpt2",
|
| 21 |
+
"n_ctx": 2048,
|
| 22 |
+
"n_embd": 768,
|
| 23 |
+
"n_head": 12,
|
| 24 |
+
"n_inner": null,
|
| 25 |
+
"n_layer": 12,
|
| 26 |
+
"n_positions": 2048,
|
| 27 |
+
"pad_token_id": 0,
|
| 28 |
+
"reorder_and_upcast_attn": false,
|
| 29 |
+
"resid_pdrop": 0.1,
|
| 30 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 31 |
+
"scale_attn_weights": true,
|
| 32 |
+
"summary_activation": null,
|
| 33 |
+
"summary_first_dropout": 0.1,
|
| 34 |
+
"summary_proj_to_labels": true,
|
| 35 |
+
"summary_type": "cls_index",
|
| 36 |
+
"summary_use_proj": true,
|
| 37 |
+
"torch_dtype": "float32",
|
| 38 |
+
"transformers_version": "4.48.3",
|
| 39 |
+
"use_cache": true,
|
| 40 |
+
"vocab_size": 50264
|
| 41 |
+
}
|
checkpoint-1539/generation_config.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"pad_token_id": 0,
|
| 6 |
+
"transformers_version": "4.48.3"
|
| 7 |
+
}
|
checkpoint-1539/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2f7a48593bb8608da895e5a9952affea45da965be8b7b500ac3f4c09d7ffcdb1
|
| 3 |
+
size 500941440
|
checkpoint-1539/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3d6968e87fd52909b2799987342af0fc39e79e6f00080dacbb9d6105914a2bbf
|
| 3 |
+
size 1001976762
|
checkpoint-1539/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d3d56993867fda4e2d99c57ad37ecc52c95af68683f7ad453cc2b1d442da5952
|
| 3 |
+
size 14244
|
checkpoint-1539/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b726bb56f04207a87f82d3f843f536e33e1283a51ba025e626232662b69d66c4
|
| 3 |
+
size 1064
|
checkpoint-1539/trainer_state.json
ADDED
|
@@ -0,0 +1,1120 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1539,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.01949317738791423,
|
| 13 |
+
"grad_norm": 9.292141914367676,
|
| 14 |
+
"learning_rate": 4.967511371020143e-05,
|
| 15 |
+
"loss": 6.2548,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.03898635477582846,
|
| 20 |
+
"grad_norm": 7.766495227813721,
|
| 21 |
+
"learning_rate": 4.935022742040286e-05,
|
| 22 |
+
"loss": 5.0175,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.05847953216374269,
|
| 27 |
+
"grad_norm": 6.4375128746032715,
|
| 28 |
+
"learning_rate": 4.902534113060429e-05,
|
| 29 |
+
"loss": 4.59,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.07797270955165692,
|
| 34 |
+
"grad_norm": 7.395759105682373,
|
| 35 |
+
"learning_rate": 4.8700454840805724e-05,
|
| 36 |
+
"loss": 4.2891,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.09746588693957114,
|
| 41 |
+
"grad_norm": 5.542599678039551,
|
| 42 |
+
"learning_rate": 4.837556855100715e-05,
|
| 43 |
+
"loss": 4.1904,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.11695906432748537,
|
| 48 |
+
"grad_norm": 8.749999046325684,
|
| 49 |
+
"learning_rate": 4.805068226120858e-05,
|
| 50 |
+
"loss": 4.1512,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.1364522417153996,
|
| 55 |
+
"grad_norm": 6.6702446937561035,
|
| 56 |
+
"learning_rate": 4.772579597141001e-05,
|
| 57 |
+
"loss": 3.9695,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.15594541910331383,
|
| 62 |
+
"grad_norm": 13.8934326171875,
|
| 63 |
+
"learning_rate": 4.740090968161144e-05,
|
| 64 |
+
"loss": 3.9657,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.17543859649122806,
|
| 69 |
+
"grad_norm": 15.161931037902832,
|
| 70 |
+
"learning_rate": 4.707602339181287e-05,
|
| 71 |
+
"loss": 3.9312,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.1949317738791423,
|
| 76 |
+
"grad_norm": 8.096311569213867,
|
| 77 |
+
"learning_rate": 4.67511371020143e-05,
|
| 78 |
+
"loss": 3.9407,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.21442495126705652,
|
| 83 |
+
"grad_norm": 6.077565670013428,
|
| 84 |
+
"learning_rate": 4.642625081221573e-05,
|
| 85 |
+
"loss": 3.7438,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.23391812865497075,
|
| 90 |
+
"grad_norm": 5.909548282623291,
|
| 91 |
+
"learning_rate": 4.610136452241716e-05,
|
| 92 |
+
"loss": 3.8495,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.253411306042885,
|
| 97 |
+
"grad_norm": 6.919253349304199,
|
| 98 |
+
"learning_rate": 4.577647823261859e-05,
|
| 99 |
+
"loss": 3.8178,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.2729044834307992,
|
| 104 |
+
"grad_norm": 5.3416218757629395,
|
| 105 |
+
"learning_rate": 4.545159194282001e-05,
|
| 106 |
+
"loss": 3.6388,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.29239766081871343,
|
| 111 |
+
"grad_norm": 6.484707832336426,
|
| 112 |
+
"learning_rate": 4.512670565302144e-05,
|
| 113 |
+
"loss": 3.8361,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.31189083820662766,
|
| 118 |
+
"grad_norm": 6.068100929260254,
|
| 119 |
+
"learning_rate": 4.480181936322287e-05,
|
| 120 |
+
"loss": 3.7164,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.3313840155945419,
|
| 125 |
+
"grad_norm": 6.298447608947754,
|
| 126 |
+
"learning_rate": 4.4476933073424304e-05,
|
| 127 |
+
"loss": 3.6589,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.3508771929824561,
|
| 132 |
+
"grad_norm": 4.752441883087158,
|
| 133 |
+
"learning_rate": 4.4152046783625734e-05,
|
| 134 |
+
"loss": 3.7328,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.37037037037037035,
|
| 139 |
+
"grad_norm": 5.050961971282959,
|
| 140 |
+
"learning_rate": 4.3827160493827164e-05,
|
| 141 |
+
"loss": 3.7162,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.3898635477582846,
|
| 146 |
+
"grad_norm": 4.73190975189209,
|
| 147 |
+
"learning_rate": 4.3502274204028594e-05,
|
| 148 |
+
"loss": 3.6748,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.4093567251461988,
|
| 153 |
+
"grad_norm": 10.872115135192871,
|
| 154 |
+
"learning_rate": 4.3177387914230025e-05,
|
| 155 |
+
"loss": 3.7618,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.42884990253411304,
|
| 160 |
+
"grad_norm": 6.119602680206299,
|
| 161 |
+
"learning_rate": 4.285250162443145e-05,
|
| 162 |
+
"loss": 3.747,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.44834307992202727,
|
| 167 |
+
"grad_norm": 5.332015514373779,
|
| 168 |
+
"learning_rate": 4.252761533463288e-05,
|
| 169 |
+
"loss": 3.6214,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.4678362573099415,
|
| 174 |
+
"grad_norm": 5.08021354675293,
|
| 175 |
+
"learning_rate": 4.220272904483431e-05,
|
| 176 |
+
"loss": 3.5122,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.4873294346978557,
|
| 181 |
+
"grad_norm": 4.609364986419678,
|
| 182 |
+
"learning_rate": 4.187784275503574e-05,
|
| 183 |
+
"loss": 3.5281,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.50682261208577,
|
| 188 |
+
"grad_norm": 24.73488998413086,
|
| 189 |
+
"learning_rate": 4.155295646523717e-05,
|
| 190 |
+
"loss": 3.7142,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.5263157894736842,
|
| 195 |
+
"grad_norm": 5.869173526763916,
|
| 196 |
+
"learning_rate": 4.12280701754386e-05,
|
| 197 |
+
"loss": 3.577,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.5458089668615984,
|
| 202 |
+
"grad_norm": 5.6849470138549805,
|
| 203 |
+
"learning_rate": 4.090318388564003e-05,
|
| 204 |
+
"loss": 3.6146,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.5653021442495126,
|
| 209 |
+
"grad_norm": 5.919517517089844,
|
| 210 |
+
"learning_rate": 4.057829759584146e-05,
|
| 211 |
+
"loss": 3.5417,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.5847953216374269,
|
| 216 |
+
"grad_norm": 5.177787780761719,
|
| 217 |
+
"learning_rate": 4.025341130604289e-05,
|
| 218 |
+
"loss": 3.5453,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.6042884990253411,
|
| 223 |
+
"grad_norm": 5.569265842437744,
|
| 224 |
+
"learning_rate": 3.9928525016244314e-05,
|
| 225 |
+
"loss": 3.5203,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.6237816764132553,
|
| 230 |
+
"grad_norm": 6.63994026184082,
|
| 231 |
+
"learning_rate": 3.9603638726445744e-05,
|
| 232 |
+
"loss": 3.5504,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.6432748538011696,
|
| 237 |
+
"grad_norm": 5.785546779632568,
|
| 238 |
+
"learning_rate": 3.9278752436647174e-05,
|
| 239 |
+
"loss": 3.5587,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.6627680311890838,
|
| 244 |
+
"grad_norm": 5.489815711975098,
|
| 245 |
+
"learning_rate": 3.8953866146848605e-05,
|
| 246 |
+
"loss": 3.5612,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.682261208576998,
|
| 251 |
+
"grad_norm": 4.650988578796387,
|
| 252 |
+
"learning_rate": 3.8628979857050035e-05,
|
| 253 |
+
"loss": 3.5062,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.7017543859649122,
|
| 258 |
+
"grad_norm": 5.7823872566223145,
|
| 259 |
+
"learning_rate": 3.8304093567251465e-05,
|
| 260 |
+
"loss": 3.5819,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.7212475633528265,
|
| 265 |
+
"grad_norm": 7.364371299743652,
|
| 266 |
+
"learning_rate": 3.7979207277452896e-05,
|
| 267 |
+
"loss": 3.5845,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.7407407407407407,
|
| 272 |
+
"grad_norm": 6.099256992340088,
|
| 273 |
+
"learning_rate": 3.7654320987654326e-05,
|
| 274 |
+
"loss": 3.5787,
|
| 275 |
+
"step": 380
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.7602339181286549,
|
| 279 |
+
"grad_norm": 5.115094184875488,
|
| 280 |
+
"learning_rate": 3.7329434697855756e-05,
|
| 281 |
+
"loss": 3.5901,
|
| 282 |
+
"step": 390
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.7797270955165692,
|
| 286 |
+
"grad_norm": 5.4442830085754395,
|
| 287 |
+
"learning_rate": 3.700454840805718e-05,
|
| 288 |
+
"loss": 3.4565,
|
| 289 |
+
"step": 400
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.7992202729044834,
|
| 293 |
+
"grad_norm": 6.6567816734313965,
|
| 294 |
+
"learning_rate": 3.667966211825861e-05,
|
| 295 |
+
"loss": 3.5481,
|
| 296 |
+
"step": 410
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.8187134502923976,
|
| 300 |
+
"grad_norm": 5.831213474273682,
|
| 301 |
+
"learning_rate": 3.635477582846004e-05,
|
| 302 |
+
"loss": 3.4274,
|
| 303 |
+
"step": 420
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.8382066276803118,
|
| 307 |
+
"grad_norm": 5.754021167755127,
|
| 308 |
+
"learning_rate": 3.602988953866147e-05,
|
| 309 |
+
"loss": 3.423,
|
| 310 |
+
"step": 430
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.8576998050682261,
|
| 314 |
+
"grad_norm": 5.590385437011719,
|
| 315 |
+
"learning_rate": 3.57050032488629e-05,
|
| 316 |
+
"loss": 3.6208,
|
| 317 |
+
"step": 440
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.8771929824561403,
|
| 321 |
+
"grad_norm": 4.159304141998291,
|
| 322 |
+
"learning_rate": 3.538011695906433e-05,
|
| 323 |
+
"loss": 3.4107,
|
| 324 |
+
"step": 450
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.8966861598440545,
|
| 328 |
+
"grad_norm": 5.699474334716797,
|
| 329 |
+
"learning_rate": 3.505523066926576e-05,
|
| 330 |
+
"loss": 3.3759,
|
| 331 |
+
"step": 460
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.9161793372319688,
|
| 335 |
+
"grad_norm": 5.0284929275512695,
|
| 336 |
+
"learning_rate": 3.473034437946719e-05,
|
| 337 |
+
"loss": 3.4719,
|
| 338 |
+
"step": 470
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.935672514619883,
|
| 342 |
+
"grad_norm": 4.884040832519531,
|
| 343 |
+
"learning_rate": 3.440545808966862e-05,
|
| 344 |
+
"loss": 3.5534,
|
| 345 |
+
"step": 480
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.9551656920077972,
|
| 349 |
+
"grad_norm": 5.428916931152344,
|
| 350 |
+
"learning_rate": 3.4080571799870045e-05,
|
| 351 |
+
"loss": 3.3919,
|
| 352 |
+
"step": 490
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.9746588693957114,
|
| 356 |
+
"grad_norm": 5.549111843109131,
|
| 357 |
+
"learning_rate": 3.3755685510071476e-05,
|
| 358 |
+
"loss": 3.2799,
|
| 359 |
+
"step": 500
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.9941520467836257,
|
| 363 |
+
"grad_norm": 5.919485092163086,
|
| 364 |
+
"learning_rate": 3.3430799220272906e-05,
|
| 365 |
+
"loss": 3.6097,
|
| 366 |
+
"step": 510
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 1.0,
|
| 370 |
+
"eval_loss": 3.2824156284332275,
|
| 371 |
+
"eval_runtime": 21.5834,
|
| 372 |
+
"eval_samples_per_second": 21.127,
|
| 373 |
+
"eval_steps_per_second": 2.641,
|
| 374 |
+
"step": 513
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 1.01364522417154,
|
| 378 |
+
"grad_norm": 4.9640631675720215,
|
| 379 |
+
"learning_rate": 3.3105912930474336e-05,
|
| 380 |
+
"loss": 3.2341,
|
| 381 |
+
"step": 520
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 1.0331384015594542,
|
| 385 |
+
"grad_norm": 5.887562274932861,
|
| 386 |
+
"learning_rate": 3.2781026640675766e-05,
|
| 387 |
+
"loss": 3.0186,
|
| 388 |
+
"step": 530
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 1.0526315789473684,
|
| 392 |
+
"grad_norm": 7.111718654632568,
|
| 393 |
+
"learning_rate": 3.24561403508772e-05,
|
| 394 |
+
"loss": 3.0381,
|
| 395 |
+
"step": 540
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 1.0721247563352827,
|
| 399 |
+
"grad_norm": 6.662885665893555,
|
| 400 |
+
"learning_rate": 3.213125406107863e-05,
|
| 401 |
+
"loss": 3.0068,
|
| 402 |
+
"step": 550
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 1.0916179337231968,
|
| 406 |
+
"grad_norm": 5.402489185333252,
|
| 407 |
+
"learning_rate": 3.180636777128006e-05,
|
| 408 |
+
"loss": 3.0492,
|
| 409 |
+
"step": 560
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 1.1111111111111112,
|
| 413 |
+
"grad_norm": 6.615454196929932,
|
| 414 |
+
"learning_rate": 3.148148148148148e-05,
|
| 415 |
+
"loss": 2.9823,
|
| 416 |
+
"step": 570
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 1.1306042884990253,
|
| 420 |
+
"grad_norm": 6.344214916229248,
|
| 421 |
+
"learning_rate": 3.115659519168291e-05,
|
| 422 |
+
"loss": 2.863,
|
| 423 |
+
"step": 580
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 1.1500974658869396,
|
| 427 |
+
"grad_norm": 6.1259589195251465,
|
| 428 |
+
"learning_rate": 3.083170890188434e-05,
|
| 429 |
+
"loss": 2.9253,
|
| 430 |
+
"step": 590
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 1.1695906432748537,
|
| 434 |
+
"grad_norm": 5.922159671783447,
|
| 435 |
+
"learning_rate": 3.050682261208577e-05,
|
| 436 |
+
"loss": 3.1142,
|
| 437 |
+
"step": 600
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 1.189083820662768,
|
| 441 |
+
"grad_norm": 4.965208530426025,
|
| 442 |
+
"learning_rate": 3.0181936322287202e-05,
|
| 443 |
+
"loss": 3.0236,
|
| 444 |
+
"step": 610
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 1.2085769980506822,
|
| 448 |
+
"grad_norm": 4.726465225219727,
|
| 449 |
+
"learning_rate": 2.9857050032488632e-05,
|
| 450 |
+
"loss": 3.114,
|
| 451 |
+
"step": 620
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 1.2280701754385965,
|
| 455 |
+
"grad_norm": 5.802026748657227,
|
| 456 |
+
"learning_rate": 2.9532163742690062e-05,
|
| 457 |
+
"loss": 2.9944,
|
| 458 |
+
"step": 630
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 1.2475633528265107,
|
| 462 |
+
"grad_norm": 5.344657897949219,
|
| 463 |
+
"learning_rate": 2.920727745289149e-05,
|
| 464 |
+
"loss": 3.024,
|
| 465 |
+
"step": 640
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 1.267056530214425,
|
| 469 |
+
"grad_norm": 6.377889633178711,
|
| 470 |
+
"learning_rate": 2.888239116309292e-05,
|
| 471 |
+
"loss": 3.1357,
|
| 472 |
+
"step": 650
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 1.286549707602339,
|
| 476 |
+
"grad_norm": 5.2032694816589355,
|
| 477 |
+
"learning_rate": 2.855750487329435e-05,
|
| 478 |
+
"loss": 3.0854,
|
| 479 |
+
"step": 660
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 1.3060428849902534,
|
| 483 |
+
"grad_norm": 5.197242736816406,
|
| 484 |
+
"learning_rate": 2.823261858349578e-05,
|
| 485 |
+
"loss": 3.1208,
|
| 486 |
+
"step": 670
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 1.3255360623781676,
|
| 490 |
+
"grad_norm": 5.073383808135986,
|
| 491 |
+
"learning_rate": 2.7907732293697207e-05,
|
| 492 |
+
"loss": 2.9982,
|
| 493 |
+
"step": 680
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 1.345029239766082,
|
| 497 |
+
"grad_norm": 5.227969646453857,
|
| 498 |
+
"learning_rate": 2.7582846003898637e-05,
|
| 499 |
+
"loss": 3.0834,
|
| 500 |
+
"step": 690
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 1.364522417153996,
|
| 504 |
+
"grad_norm": 5.831137657165527,
|
| 505 |
+
"learning_rate": 2.7257959714100068e-05,
|
| 506 |
+
"loss": 3.0019,
|
| 507 |
+
"step": 700
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 1.3840155945419104,
|
| 511 |
+
"grad_norm": 5.306942939758301,
|
| 512 |
+
"learning_rate": 2.6933073424301498e-05,
|
| 513 |
+
"loss": 2.9489,
|
| 514 |
+
"step": 710
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 1.4035087719298245,
|
| 518 |
+
"grad_norm": 5.346749305725098,
|
| 519 |
+
"learning_rate": 2.6608187134502928e-05,
|
| 520 |
+
"loss": 3.0281,
|
| 521 |
+
"step": 720
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 1.4230019493177388,
|
| 525 |
+
"grad_norm": 4.3056817054748535,
|
| 526 |
+
"learning_rate": 2.6283300844704355e-05,
|
| 527 |
+
"loss": 2.8552,
|
| 528 |
+
"step": 730
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 1.442495126705653,
|
| 532 |
+
"grad_norm": 8.335524559020996,
|
| 533 |
+
"learning_rate": 2.5958414554905785e-05,
|
| 534 |
+
"loss": 2.9833,
|
| 535 |
+
"step": 740
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 1.4619883040935673,
|
| 539 |
+
"grad_norm": 4.510199069976807,
|
| 540 |
+
"learning_rate": 2.5633528265107216e-05,
|
| 541 |
+
"loss": 3.01,
|
| 542 |
+
"step": 750
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 1.4814814814814814,
|
| 546 |
+
"grad_norm": 4.329857349395752,
|
| 547 |
+
"learning_rate": 2.5308641975308646e-05,
|
| 548 |
+
"loss": 3.0841,
|
| 549 |
+
"step": 760
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 1.5009746588693957,
|
| 553 |
+
"grad_norm": 4.561777114868164,
|
| 554 |
+
"learning_rate": 2.4983755685510073e-05,
|
| 555 |
+
"loss": 2.9037,
|
| 556 |
+
"step": 770
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 1.52046783625731,
|
| 560 |
+
"grad_norm": 4.895258903503418,
|
| 561 |
+
"learning_rate": 2.46588693957115e-05,
|
| 562 |
+
"loss": 2.9441,
|
| 563 |
+
"step": 780
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 1.5399610136452242,
|
| 567 |
+
"grad_norm": 4.836617946624756,
|
| 568 |
+
"learning_rate": 2.433398310591293e-05,
|
| 569 |
+
"loss": 2.9684,
|
| 570 |
+
"step": 790
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 1.5594541910331383,
|
| 574 |
+
"grad_norm": 5.523194789886475,
|
| 575 |
+
"learning_rate": 2.400909681611436e-05,
|
| 576 |
+
"loss": 3.0216,
|
| 577 |
+
"step": 800
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 1.5789473684210527,
|
| 581 |
+
"grad_norm": 6.470391750335693,
|
| 582 |
+
"learning_rate": 2.368421052631579e-05,
|
| 583 |
+
"loss": 2.9688,
|
| 584 |
+
"step": 810
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 1.598440545808967,
|
| 588 |
+
"grad_norm": 5.374258995056152,
|
| 589 |
+
"learning_rate": 2.3359324236517217e-05,
|
| 590 |
+
"loss": 2.934,
|
| 591 |
+
"step": 820
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 1.6179337231968811,
|
| 595 |
+
"grad_norm": 4.702390670776367,
|
| 596 |
+
"learning_rate": 2.3034437946718648e-05,
|
| 597 |
+
"loss": 2.8229,
|
| 598 |
+
"step": 830
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 1.6374269005847952,
|
| 602 |
+
"grad_norm": 4.931224822998047,
|
| 603 |
+
"learning_rate": 2.2709551656920078e-05,
|
| 604 |
+
"loss": 2.9748,
|
| 605 |
+
"step": 840
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 1.6569200779727096,
|
| 609 |
+
"grad_norm": 5.478194713592529,
|
| 610 |
+
"learning_rate": 2.2384665367121508e-05,
|
| 611 |
+
"loss": 2.8791,
|
| 612 |
+
"step": 850
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 1.676413255360624,
|
| 616 |
+
"grad_norm": 4.936557769775391,
|
| 617 |
+
"learning_rate": 2.205977907732294e-05,
|
| 618 |
+
"loss": 2.9022,
|
| 619 |
+
"step": 860
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 1.695906432748538,
|
| 623 |
+
"grad_norm": 5.581218242645264,
|
| 624 |
+
"learning_rate": 2.1734892787524365e-05,
|
| 625 |
+
"loss": 3.0293,
|
| 626 |
+
"step": 870
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 1.7153996101364521,
|
| 630 |
+
"grad_norm": 6.841440677642822,
|
| 631 |
+
"learning_rate": 2.1410006497725796e-05,
|
| 632 |
+
"loss": 2.9343,
|
| 633 |
+
"step": 880
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 1.7348927875243665,
|
| 637 |
+
"grad_norm": 5.218410968780518,
|
| 638 |
+
"learning_rate": 2.1085120207927226e-05,
|
| 639 |
+
"loss": 3.0231,
|
| 640 |
+
"step": 890
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 1.7543859649122808,
|
| 644 |
+
"grad_norm": 5.5271077156066895,
|
| 645 |
+
"learning_rate": 2.0760233918128656e-05,
|
| 646 |
+
"loss": 2.8563,
|
| 647 |
+
"step": 900
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 1.773879142300195,
|
| 651 |
+
"grad_norm": 5.889744758605957,
|
| 652 |
+
"learning_rate": 2.0435347628330083e-05,
|
| 653 |
+
"loss": 3.0004,
|
| 654 |
+
"step": 910
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 1.793372319688109,
|
| 658 |
+
"grad_norm": 5.243134498596191,
|
| 659 |
+
"learning_rate": 2.0110461338531513e-05,
|
| 660 |
+
"loss": 2.7516,
|
| 661 |
+
"step": 920
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 1.8128654970760234,
|
| 665 |
+
"grad_norm": 5.98087739944458,
|
| 666 |
+
"learning_rate": 1.9785575048732944e-05,
|
| 667 |
+
"loss": 3.0529,
|
| 668 |
+
"step": 930
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 1.8323586744639377,
|
| 672 |
+
"grad_norm": 4.677685737609863,
|
| 673 |
+
"learning_rate": 1.9460688758934374e-05,
|
| 674 |
+
"loss": 2.8891,
|
| 675 |
+
"step": 940
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 1.8518518518518519,
|
| 679 |
+
"grad_norm": 6.311639308929443,
|
| 680 |
+
"learning_rate": 1.91358024691358e-05,
|
| 681 |
+
"loss": 2.9617,
|
| 682 |
+
"step": 950
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 1.871345029239766,
|
| 686 |
+
"grad_norm": 4.64074182510376,
|
| 687 |
+
"learning_rate": 1.881091617933723e-05,
|
| 688 |
+
"loss": 2.9111,
|
| 689 |
+
"step": 960
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 1.8908382066276803,
|
| 693 |
+
"grad_norm": 4.606050491333008,
|
| 694 |
+
"learning_rate": 1.848602988953866e-05,
|
| 695 |
+
"loss": 2.8621,
|
| 696 |
+
"step": 970
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 1.9103313840155947,
|
| 700 |
+
"grad_norm": 4.810323238372803,
|
| 701 |
+
"learning_rate": 1.816114359974009e-05,
|
| 702 |
+
"loss": 3.017,
|
| 703 |
+
"step": 980
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 1.9298245614035088,
|
| 707 |
+
"grad_norm": 5.766260147094727,
|
| 708 |
+
"learning_rate": 1.7836257309941522e-05,
|
| 709 |
+
"loss": 2.8715,
|
| 710 |
+
"step": 990
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 1.949317738791423,
|
| 714 |
+
"grad_norm": 5.515121936798096,
|
| 715 |
+
"learning_rate": 1.751137102014295e-05,
|
| 716 |
+
"loss": 2.9338,
|
| 717 |
+
"step": 1000
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 1.9688109161793372,
|
| 721 |
+
"grad_norm": 5.856773376464844,
|
| 722 |
+
"learning_rate": 1.718648473034438e-05,
|
| 723 |
+
"loss": 2.9503,
|
| 724 |
+
"step": 1010
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 1.9883040935672516,
|
| 728 |
+
"grad_norm": 6.5240373611450195,
|
| 729 |
+
"learning_rate": 1.686159844054581e-05,
|
| 730 |
+
"loss": 2.9444,
|
| 731 |
+
"step": 1020
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 2.0,
|
| 735 |
+
"eval_loss": 3.203155994415283,
|
| 736 |
+
"eval_runtime": 21.2886,
|
| 737 |
+
"eval_samples_per_second": 21.42,
|
| 738 |
+
"eval_steps_per_second": 2.677,
|
| 739 |
+
"step": 1026
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"epoch": 2.007797270955166,
|
| 743 |
+
"grad_norm": 4.038070201873779,
|
| 744 |
+
"learning_rate": 1.653671215074724e-05,
|
| 745 |
+
"loss": 2.8103,
|
| 746 |
+
"step": 1030
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 2.02729044834308,
|
| 750 |
+
"grad_norm": 5.510829925537109,
|
| 751 |
+
"learning_rate": 1.6211825860948666e-05,
|
| 752 |
+
"loss": 2.6078,
|
| 753 |
+
"step": 1040
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 2.046783625730994,
|
| 757 |
+
"grad_norm": 6.110142707824707,
|
| 758 |
+
"learning_rate": 1.5886939571150097e-05,
|
| 759 |
+
"loss": 2.6043,
|
| 760 |
+
"step": 1050
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 2.0662768031189085,
|
| 764 |
+
"grad_norm": 7.630992412567139,
|
| 765 |
+
"learning_rate": 1.5562053281351527e-05,
|
| 766 |
+
"loss": 2.6449,
|
| 767 |
+
"step": 1060
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 2.0857699805068224,
|
| 771 |
+
"grad_norm": 6.36288595199585,
|
| 772 |
+
"learning_rate": 1.5237166991552956e-05,
|
| 773 |
+
"loss": 2.6561,
|
| 774 |
+
"step": 1070
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 2.1052631578947367,
|
| 778 |
+
"grad_norm": 6.1103363037109375,
|
| 779 |
+
"learning_rate": 1.4912280701754386e-05,
|
| 780 |
+
"loss": 2.615,
|
| 781 |
+
"step": 1080
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 2.124756335282651,
|
| 785 |
+
"grad_norm": 7.310168266296387,
|
| 786 |
+
"learning_rate": 1.4587394411955816e-05,
|
| 787 |
+
"loss": 2.6722,
|
| 788 |
+
"step": 1090
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 2.1442495126705654,
|
| 792 |
+
"grad_norm": 5.365858554840088,
|
| 793 |
+
"learning_rate": 1.4262508122157245e-05,
|
| 794 |
+
"loss": 2.624,
|
| 795 |
+
"step": 1100
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 2.1637426900584797,
|
| 799 |
+
"grad_norm": 5.3909735679626465,
|
| 800 |
+
"learning_rate": 1.3937621832358675e-05,
|
| 801 |
+
"loss": 2.595,
|
| 802 |
+
"step": 1110
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 2.1832358674463936,
|
| 806 |
+
"grad_norm": 6.050861835479736,
|
| 807 |
+
"learning_rate": 1.3612735542560104e-05,
|
| 808 |
+
"loss": 2.5981,
|
| 809 |
+
"step": 1120
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 2.202729044834308,
|
| 813 |
+
"grad_norm": 6.927775859832764,
|
| 814 |
+
"learning_rate": 1.3287849252761534e-05,
|
| 815 |
+
"loss": 2.6114,
|
| 816 |
+
"step": 1130
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 2.2222222222222223,
|
| 820 |
+
"grad_norm": 6.272125244140625,
|
| 821 |
+
"learning_rate": 1.2962962962962962e-05,
|
| 822 |
+
"loss": 2.654,
|
| 823 |
+
"step": 1140
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 2.241715399610136,
|
| 827 |
+
"grad_norm": 5.306847095489502,
|
| 828 |
+
"learning_rate": 1.2638076673164393e-05,
|
| 829 |
+
"loss": 2.7056,
|
| 830 |
+
"step": 1150
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 2.2612085769980506,
|
| 834 |
+
"grad_norm": 5.6090850830078125,
|
| 835 |
+
"learning_rate": 1.2313190383365821e-05,
|
| 836 |
+
"loss": 2.6322,
|
| 837 |
+
"step": 1160
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 2.280701754385965,
|
| 841 |
+
"grad_norm": 4.8249382972717285,
|
| 842 |
+
"learning_rate": 1.1988304093567252e-05,
|
| 843 |
+
"loss": 2.6714,
|
| 844 |
+
"step": 1170
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 2.3001949317738792,
|
| 848 |
+
"grad_norm": 6.645869731903076,
|
| 849 |
+
"learning_rate": 1.166341780376868e-05,
|
| 850 |
+
"loss": 2.7006,
|
| 851 |
+
"step": 1180
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 2.3196881091617936,
|
| 855 |
+
"grad_norm": 5.666390895843506,
|
| 856 |
+
"learning_rate": 1.133853151397011e-05,
|
| 857 |
+
"loss": 2.6726,
|
| 858 |
+
"step": 1190
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 2.3391812865497075,
|
| 862 |
+
"grad_norm": 5.168858528137207,
|
| 863 |
+
"learning_rate": 1.1013645224171539e-05,
|
| 864 |
+
"loss": 2.4944,
|
| 865 |
+
"step": 1200
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 2.358674463937622,
|
| 869 |
+
"grad_norm": 5.98241662979126,
|
| 870 |
+
"learning_rate": 1.068875893437297e-05,
|
| 871 |
+
"loss": 2.6001,
|
| 872 |
+
"step": 1210
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 2.378167641325536,
|
| 876 |
+
"grad_norm": 5.360602855682373,
|
| 877 |
+
"learning_rate": 1.03638726445744e-05,
|
| 878 |
+
"loss": 2.8167,
|
| 879 |
+
"step": 1220
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 2.39766081871345,
|
| 883 |
+
"grad_norm": 6.712647914886475,
|
| 884 |
+
"learning_rate": 1.0038986354775828e-05,
|
| 885 |
+
"loss": 2.6197,
|
| 886 |
+
"step": 1230
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 2.4171539961013644,
|
| 890 |
+
"grad_norm": 5.139374256134033,
|
| 891 |
+
"learning_rate": 9.714100064977258e-06,
|
| 892 |
+
"loss": 2.6276,
|
| 893 |
+
"step": 1240
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 2.4366471734892787,
|
| 897 |
+
"grad_norm": 5.239111423492432,
|
| 898 |
+
"learning_rate": 9.389213775178687e-06,
|
| 899 |
+
"loss": 2.5737,
|
| 900 |
+
"step": 1250
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 2.456140350877193,
|
| 904 |
+
"grad_norm": 5.82625675201416,
|
| 905 |
+
"learning_rate": 9.064327485380117e-06,
|
| 906 |
+
"loss": 2.7264,
|
| 907 |
+
"step": 1260
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 2.4756335282651074,
|
| 911 |
+
"grad_norm": 7.747376441955566,
|
| 912 |
+
"learning_rate": 8.739441195581546e-06,
|
| 913 |
+
"loss": 2.5372,
|
| 914 |
+
"step": 1270
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 2.4951267056530213,
|
| 918 |
+
"grad_norm": 6.986497402191162,
|
| 919 |
+
"learning_rate": 8.414554905782976e-06,
|
| 920 |
+
"loss": 2.6583,
|
| 921 |
+
"step": 1280
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 2.5146198830409356,
|
| 925 |
+
"grad_norm": 5.268970012664795,
|
| 926 |
+
"learning_rate": 8.089668615984405e-06,
|
| 927 |
+
"loss": 2.5864,
|
| 928 |
+
"step": 1290
|
| 929 |
+
},
|
| 930 |
+
{
|
| 931 |
+
"epoch": 2.53411306042885,
|
| 932 |
+
"grad_norm": 6.267032623291016,
|
| 933 |
+
"learning_rate": 7.764782326185835e-06,
|
| 934 |
+
"loss": 2.5389,
|
| 935 |
+
"step": 1300
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 2.553606237816764,
|
| 939 |
+
"grad_norm": 5.6785173416137695,
|
| 940 |
+
"learning_rate": 7.4398960363872645e-06,
|
| 941 |
+
"loss": 2.6147,
|
| 942 |
+
"step": 1310
|
| 943 |
+
},
|
| 944 |
+
{
|
| 945 |
+
"epoch": 2.573099415204678,
|
| 946 |
+
"grad_norm": 7.006417751312256,
|
| 947 |
+
"learning_rate": 7.115009746588694e-06,
|
| 948 |
+
"loss": 2.6028,
|
| 949 |
+
"step": 1320
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 2.5925925925925926,
|
| 953 |
+
"grad_norm": 5.888213634490967,
|
| 954 |
+
"learning_rate": 6.790123456790123e-06,
|
| 955 |
+
"loss": 2.7019,
|
| 956 |
+
"step": 1330
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 2.612085769980507,
|
| 960 |
+
"grad_norm": 6.327919960021973,
|
| 961 |
+
"learning_rate": 6.465237166991553e-06,
|
| 962 |
+
"loss": 2.6776,
|
| 963 |
+
"step": 1340
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 2.6315789473684212,
|
| 967 |
+
"grad_norm": 5.6416096687316895,
|
| 968 |
+
"learning_rate": 6.140350877192982e-06,
|
| 969 |
+
"loss": 2.6533,
|
| 970 |
+
"step": 1350
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 2.651072124756335,
|
| 974 |
+
"grad_norm": 6.070097923278809,
|
| 975 |
+
"learning_rate": 5.815464587394412e-06,
|
| 976 |
+
"loss": 2.674,
|
| 977 |
+
"step": 1360
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 2.6705653021442495,
|
| 981 |
+
"grad_norm": 5.943949222564697,
|
| 982 |
+
"learning_rate": 5.490578297595841e-06,
|
| 983 |
+
"loss": 2.5782,
|
| 984 |
+
"step": 1370
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 2.690058479532164,
|
| 988 |
+
"grad_norm": 5.129433631896973,
|
| 989 |
+
"learning_rate": 5.165692007797271e-06,
|
| 990 |
+
"loss": 2.7243,
|
| 991 |
+
"step": 1380
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 2.7095516569200777,
|
| 995 |
+
"grad_norm": 6.323672771453857,
|
| 996 |
+
"learning_rate": 4.840805717998701e-06,
|
| 997 |
+
"loss": 2.7187,
|
| 998 |
+
"step": 1390
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 2.729044834307992,
|
| 1002 |
+
"grad_norm": 4.658057689666748,
|
| 1003 |
+
"learning_rate": 4.51591942820013e-06,
|
| 1004 |
+
"loss": 2.6158,
|
| 1005 |
+
"step": 1400
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 2.7485380116959064,
|
| 1009 |
+
"grad_norm": 5.387426376342773,
|
| 1010 |
+
"learning_rate": 4.19103313840156e-06,
|
| 1011 |
+
"loss": 2.5866,
|
| 1012 |
+
"step": 1410
|
| 1013 |
+
},
|
| 1014 |
+
{
|
| 1015 |
+
"epoch": 2.7680311890838207,
|
| 1016 |
+
"grad_norm": 6.695394992828369,
|
| 1017 |
+
"learning_rate": 3.866146848602989e-06,
|
| 1018 |
+
"loss": 2.5415,
|
| 1019 |
+
"step": 1420
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 2.787524366471735,
|
| 1023 |
+
"grad_norm": 5.209081172943115,
|
| 1024 |
+
"learning_rate": 3.5412605588044185e-06,
|
| 1025 |
+
"loss": 2.4233,
|
| 1026 |
+
"step": 1430
|
| 1027 |
+
},
|
| 1028 |
+
{
|
| 1029 |
+
"epoch": 2.807017543859649,
|
| 1030 |
+
"grad_norm": 4.750288009643555,
|
| 1031 |
+
"learning_rate": 3.216374269005848e-06,
|
| 1032 |
+
"loss": 2.5685,
|
| 1033 |
+
"step": 1440
|
| 1034 |
+
},
|
| 1035 |
+
{
|
| 1036 |
+
"epoch": 2.8265107212475633,
|
| 1037 |
+
"grad_norm": 5.388522148132324,
|
| 1038 |
+
"learning_rate": 2.8914879792072773e-06,
|
| 1039 |
+
"loss": 2.6053,
|
| 1040 |
+
"step": 1450
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 2.8460038986354776,
|
| 1044 |
+
"grad_norm": 6.151064395904541,
|
| 1045 |
+
"learning_rate": 2.5666016894087067e-06,
|
| 1046 |
+
"loss": 2.6842,
|
| 1047 |
+
"step": 1460
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 2.8654970760233915,
|
| 1051 |
+
"grad_norm": 6.249871253967285,
|
| 1052 |
+
"learning_rate": 2.2417153996101366e-06,
|
| 1053 |
+
"loss": 2.505,
|
| 1054 |
+
"step": 1470
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"epoch": 2.884990253411306,
|
| 1058 |
+
"grad_norm": 7.69942045211792,
|
| 1059 |
+
"learning_rate": 1.916829109811566e-06,
|
| 1060 |
+
"loss": 2.6107,
|
| 1061 |
+
"step": 1480
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"epoch": 2.9044834307992202,
|
| 1065 |
+
"grad_norm": 5.715646266937256,
|
| 1066 |
+
"learning_rate": 1.5919428200129955e-06,
|
| 1067 |
+
"loss": 2.6015,
|
| 1068 |
+
"step": 1490
|
| 1069 |
+
},
|
| 1070 |
+
{
|
| 1071 |
+
"epoch": 2.9239766081871346,
|
| 1072 |
+
"grad_norm": 5.447291851043701,
|
| 1073 |
+
"learning_rate": 1.2670565302144249e-06,
|
| 1074 |
+
"loss": 2.7594,
|
| 1075 |
+
"step": 1500
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"epoch": 2.943469785575049,
|
| 1079 |
+
"grad_norm": 5.8685526847839355,
|
| 1080 |
+
"learning_rate": 9.421702404158544e-07,
|
| 1081 |
+
"loss": 2.5403,
|
| 1082 |
+
"step": 1510
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 2.962962962962963,
|
| 1086 |
+
"grad_norm": 5.752800941467285,
|
| 1087 |
+
"learning_rate": 6.17283950617284e-07,
|
| 1088 |
+
"loss": 2.5594,
|
| 1089 |
+
"step": 1520
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 2.982456140350877,
|
| 1093 |
+
"grad_norm": 6.230057716369629,
|
| 1094 |
+
"learning_rate": 2.9239766081871344e-07,
|
| 1095 |
+
"loss": 2.6494,
|
| 1096 |
+
"step": 1530
|
| 1097 |
+
}
|
| 1098 |
+
],
|
| 1099 |
+
"logging_steps": 10,
|
| 1100 |
+
"max_steps": 1539,
|
| 1101 |
+
"num_input_tokens_seen": 0,
|
| 1102 |
+
"num_train_epochs": 3,
|
| 1103 |
+
"save_steps": 500,
|
| 1104 |
+
"stateful_callbacks": {
|
| 1105 |
+
"TrainerControl": {
|
| 1106 |
+
"args": {
|
| 1107 |
+
"should_epoch_stop": false,
|
| 1108 |
+
"should_evaluate": false,
|
| 1109 |
+
"should_log": false,
|
| 1110 |
+
"should_save": true,
|
| 1111 |
+
"should_training_stop": true
|
| 1112 |
+
},
|
| 1113 |
+
"attributes": {}
|
| 1114 |
+
}
|
| 1115 |
+
},
|
| 1116 |
+
"total_flos": 3214675869696000.0,
|
| 1117 |
+
"train_batch_size": 8,
|
| 1118 |
+
"trial_name": null,
|
| 1119 |
+
"trial_params": null
|
| 1120 |
+
}
|
checkpoint-1539/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6b0e13a27dc9eb93f3cc3f0f69eeb907a00fef0cf3de646ea9710a3354ece317
|
| 3 |
+
size 5304
|
checkpoint-500/config.json
ADDED
|
@@ -0,0 +1,41 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "ai-forever/rugpt3small_based_on_gpt2",
|
| 3 |
+
"activation_function": "gelu_new",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.1,
|
| 8 |
+
"bos_token_id": 1,
|
| 9 |
+
"embd_pdrop": 0.1,
|
| 10 |
+
"eos_token_id": 2,
|
| 11 |
+
"gradient_checkpointing": false,
|
| 12 |
+
"id2label": {
|
| 13 |
+
"0": "LABEL_0"
|
| 14 |
+
},
|
| 15 |
+
"initializer_range": 0.02,
|
| 16 |
+
"label2id": {
|
| 17 |
+
"LABEL_0": 0
|
| 18 |
+
},
|
| 19 |
+
"layer_norm_epsilon": 1e-05,
|
| 20 |
+
"model_type": "gpt2",
|
| 21 |
+
"n_ctx": 2048,
|
| 22 |
+
"n_embd": 768,
|
| 23 |
+
"n_head": 12,
|
| 24 |
+
"n_inner": null,
|
| 25 |
+
"n_layer": 12,
|
| 26 |
+
"n_positions": 2048,
|
| 27 |
+
"pad_token_id": 0,
|
| 28 |
+
"reorder_and_upcast_attn": false,
|
| 29 |
+
"resid_pdrop": 0.1,
|
| 30 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 31 |
+
"scale_attn_weights": true,
|
| 32 |
+
"summary_activation": null,
|
| 33 |
+
"summary_first_dropout": 0.1,
|
| 34 |
+
"summary_proj_to_labels": true,
|
| 35 |
+
"summary_type": "cls_index",
|
| 36 |
+
"summary_use_proj": true,
|
| 37 |
+
"torch_dtype": "float32",
|
| 38 |
+
"transformers_version": "4.48.3",
|
| 39 |
+
"use_cache": true,
|
| 40 |
+
"vocab_size": 50264
|
| 41 |
+
}
|
checkpoint-500/generation_config.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 1,
|
| 4 |
+
"eos_token_id": 2,
|
| 5 |
+
"pad_token_id": 0,
|
| 6 |
+
"transformers_version": "4.48.3"
|
| 7 |
+
}
|
checkpoint-500/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8831e9f8525373b3623fc321b372ecb5f82f1ba3dc23b6bdb210906ad38bf84f
|
| 3 |
+
size 500941440
|
checkpoint-500/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:198bcee269ed09c4ccf4ec22dcd51e945182df2ac65d09c8cbcfe5721f7791ba
|
| 3 |
+
size 1001976762
|
checkpoint-500/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:57c9b192a78a9279027dc0a850ab2b7ecba5404e3a5cde418abf6c26d8396529
|
| 3 |
+
size 14244
|
checkpoint-500/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:057c886cf0c33f96a03a676f56706efbc201608efe481f0a3be264f08bdb78a2
|
| 3 |
+
size 1064
|
checkpoint-500/trainer_state.json
ADDED
|
@@ -0,0 +1,383 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9746588693957114,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 500,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.01949317738791423,
|
| 13 |
+
"grad_norm": 9.292141914367676,
|
| 14 |
+
"learning_rate": 4.967511371020143e-05,
|
| 15 |
+
"loss": 6.2548,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.03898635477582846,
|
| 20 |
+
"grad_norm": 7.766495227813721,
|
| 21 |
+
"learning_rate": 4.935022742040286e-05,
|
| 22 |
+
"loss": 5.0175,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.05847953216374269,
|
| 27 |
+
"grad_norm": 6.4375128746032715,
|
| 28 |
+
"learning_rate": 4.902534113060429e-05,
|
| 29 |
+
"loss": 4.59,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.07797270955165692,
|
| 34 |
+
"grad_norm": 7.395759105682373,
|
| 35 |
+
"learning_rate": 4.8700454840805724e-05,
|
| 36 |
+
"loss": 4.2891,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.09746588693957114,
|
| 41 |
+
"grad_norm": 5.542599678039551,
|
| 42 |
+
"learning_rate": 4.837556855100715e-05,
|
| 43 |
+
"loss": 4.1904,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.11695906432748537,
|
| 48 |
+
"grad_norm": 8.749999046325684,
|
| 49 |
+
"learning_rate": 4.805068226120858e-05,
|
| 50 |
+
"loss": 4.1512,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.1364522417153996,
|
| 55 |
+
"grad_norm": 6.6702446937561035,
|
| 56 |
+
"learning_rate": 4.772579597141001e-05,
|
| 57 |
+
"loss": 3.9695,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.15594541910331383,
|
| 62 |
+
"grad_norm": 13.8934326171875,
|
| 63 |
+
"learning_rate": 4.740090968161144e-05,
|
| 64 |
+
"loss": 3.9657,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.17543859649122806,
|
| 69 |
+
"grad_norm": 15.161931037902832,
|
| 70 |
+
"learning_rate": 4.707602339181287e-05,
|
| 71 |
+
"loss": 3.9312,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.1949317738791423,
|
| 76 |
+
"grad_norm": 8.096311569213867,
|
| 77 |
+
"learning_rate": 4.67511371020143e-05,
|
| 78 |
+
"loss": 3.9407,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.21442495126705652,
|
| 83 |
+
"grad_norm": 6.077565670013428,
|
| 84 |
+
"learning_rate": 4.642625081221573e-05,
|
| 85 |
+
"loss": 3.7438,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.23391812865497075,
|
| 90 |
+
"grad_norm": 5.909548282623291,
|
| 91 |
+
"learning_rate": 4.610136452241716e-05,
|
| 92 |
+
"loss": 3.8495,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.253411306042885,
|
| 97 |
+
"grad_norm": 6.919253349304199,
|
| 98 |
+
"learning_rate": 4.577647823261859e-05,
|
| 99 |
+
"loss": 3.8178,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.2729044834307992,
|
| 104 |
+
"grad_norm": 5.3416218757629395,
|
| 105 |
+
"learning_rate": 4.545159194282001e-05,
|
| 106 |
+
"loss": 3.6388,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.29239766081871343,
|
| 111 |
+
"grad_norm": 6.484707832336426,
|
| 112 |
+
"learning_rate": 4.512670565302144e-05,
|
| 113 |
+
"loss": 3.8361,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.31189083820662766,
|
| 118 |
+
"grad_norm": 6.068100929260254,
|
| 119 |
+
"learning_rate": 4.480181936322287e-05,
|
| 120 |
+
"loss": 3.7164,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.3313840155945419,
|
| 125 |
+
"grad_norm": 6.298447608947754,
|
| 126 |
+
"learning_rate": 4.4476933073424304e-05,
|
| 127 |
+
"loss": 3.6589,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.3508771929824561,
|
| 132 |
+
"grad_norm": 4.752441883087158,
|
| 133 |
+
"learning_rate": 4.4152046783625734e-05,
|
| 134 |
+
"loss": 3.7328,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.37037037037037035,
|
| 139 |
+
"grad_norm": 5.050961971282959,
|
| 140 |
+
"learning_rate": 4.3827160493827164e-05,
|
| 141 |
+
"loss": 3.7162,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.3898635477582846,
|
| 146 |
+
"grad_norm": 4.73190975189209,
|
| 147 |
+
"learning_rate": 4.3502274204028594e-05,
|
| 148 |
+
"loss": 3.6748,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.4093567251461988,
|
| 153 |
+
"grad_norm": 10.872115135192871,
|
| 154 |
+
"learning_rate": 4.3177387914230025e-05,
|
| 155 |
+
"loss": 3.7618,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.42884990253411304,
|
| 160 |
+
"grad_norm": 6.119602680206299,
|
| 161 |
+
"learning_rate": 4.285250162443145e-05,
|
| 162 |
+
"loss": 3.747,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.44834307992202727,
|
| 167 |
+
"grad_norm": 5.332015514373779,
|
| 168 |
+
"learning_rate": 4.252761533463288e-05,
|
| 169 |
+
"loss": 3.6214,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.4678362573099415,
|
| 174 |
+
"grad_norm": 5.08021354675293,
|
| 175 |
+
"learning_rate": 4.220272904483431e-05,
|
| 176 |
+
"loss": 3.5122,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.4873294346978557,
|
| 181 |
+
"grad_norm": 4.609364986419678,
|
| 182 |
+
"learning_rate": 4.187784275503574e-05,
|
| 183 |
+
"loss": 3.5281,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.50682261208577,
|
| 188 |
+
"grad_norm": 24.73488998413086,
|
| 189 |
+
"learning_rate": 4.155295646523717e-05,
|
| 190 |
+
"loss": 3.7142,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.5263157894736842,
|
| 195 |
+
"grad_norm": 5.869173526763916,
|
| 196 |
+
"learning_rate": 4.12280701754386e-05,
|
| 197 |
+
"loss": 3.577,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.5458089668615984,
|
| 202 |
+
"grad_norm": 5.6849470138549805,
|
| 203 |
+
"learning_rate": 4.090318388564003e-05,
|
| 204 |
+
"loss": 3.6146,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.5653021442495126,
|
| 209 |
+
"grad_norm": 5.919517517089844,
|
| 210 |
+
"learning_rate": 4.057829759584146e-05,
|
| 211 |
+
"loss": 3.5417,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.5847953216374269,
|
| 216 |
+
"grad_norm": 5.177787780761719,
|
| 217 |
+
"learning_rate": 4.025341130604289e-05,
|
| 218 |
+
"loss": 3.5453,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.6042884990253411,
|
| 223 |
+
"grad_norm": 5.569265842437744,
|
| 224 |
+
"learning_rate": 3.9928525016244314e-05,
|
| 225 |
+
"loss": 3.5203,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.6237816764132553,
|
| 230 |
+
"grad_norm": 6.63994026184082,
|
| 231 |
+
"learning_rate": 3.9603638726445744e-05,
|
| 232 |
+
"loss": 3.5504,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.6432748538011696,
|
| 237 |
+
"grad_norm": 5.785546779632568,
|
| 238 |
+
"learning_rate": 3.9278752436647174e-05,
|
| 239 |
+
"loss": 3.5587,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.6627680311890838,
|
| 244 |
+
"grad_norm": 5.489815711975098,
|
| 245 |
+
"learning_rate": 3.8953866146848605e-05,
|
| 246 |
+
"loss": 3.5612,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.682261208576998,
|
| 251 |
+
"grad_norm": 4.650988578796387,
|
| 252 |
+
"learning_rate": 3.8628979857050035e-05,
|
| 253 |
+
"loss": 3.5062,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.7017543859649122,
|
| 258 |
+
"grad_norm": 5.7823872566223145,
|
| 259 |
+
"learning_rate": 3.8304093567251465e-05,
|
| 260 |
+
"loss": 3.5819,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.7212475633528265,
|
| 265 |
+
"grad_norm": 7.364371299743652,
|
| 266 |
+
"learning_rate": 3.7979207277452896e-05,
|
| 267 |
+
"loss": 3.5845,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.7407407407407407,
|
| 272 |
+
"grad_norm": 6.099256992340088,
|
| 273 |
+
"learning_rate": 3.7654320987654326e-05,
|
| 274 |
+
"loss": 3.5787,
|
| 275 |
+
"step": 380
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.7602339181286549,
|
| 279 |
+
"grad_norm": 5.115094184875488,
|
| 280 |
+
"learning_rate": 3.7329434697855756e-05,
|
| 281 |
+
"loss": 3.5901,
|
| 282 |
+
"step": 390
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.7797270955165692,
|
| 286 |
+
"grad_norm": 5.4442830085754395,
|
| 287 |
+
"learning_rate": 3.700454840805718e-05,
|
| 288 |
+
"loss": 3.4565,
|
| 289 |
+
"step": 400
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.7992202729044834,
|
| 293 |
+
"grad_norm": 6.6567816734313965,
|
| 294 |
+
"learning_rate": 3.667966211825861e-05,
|
| 295 |
+
"loss": 3.5481,
|
| 296 |
+
"step": 410
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.8187134502923976,
|
| 300 |
+
"grad_norm": 5.831213474273682,
|
| 301 |
+
"learning_rate": 3.635477582846004e-05,
|
| 302 |
+
"loss": 3.4274,
|
| 303 |
+
"step": 420
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.8382066276803118,
|
| 307 |
+
"grad_norm": 5.754021167755127,
|
| 308 |
+
"learning_rate": 3.602988953866147e-05,
|
| 309 |
+
"loss": 3.423,
|
| 310 |
+
"step": 430
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.8576998050682261,
|
| 314 |
+
"grad_norm": 5.590385437011719,
|
| 315 |
+
"learning_rate": 3.57050032488629e-05,
|
| 316 |
+
"loss": 3.6208,
|
| 317 |
+
"step": 440
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.8771929824561403,
|
| 321 |
+
"grad_norm": 4.159304141998291,
|
| 322 |
+
"learning_rate": 3.538011695906433e-05,
|
| 323 |
+
"loss": 3.4107,
|
| 324 |
+
"step": 450
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.8966861598440545,
|
| 328 |
+
"grad_norm": 5.699474334716797,
|
| 329 |
+
"learning_rate": 3.505523066926576e-05,
|
| 330 |
+
"loss": 3.3759,
|
| 331 |
+
"step": 460
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.9161793372319688,
|
| 335 |
+
"grad_norm": 5.0284929275512695,
|
| 336 |
+
"learning_rate": 3.473034437946719e-05,
|
| 337 |
+
"loss": 3.4719,
|
| 338 |
+
"step": 470
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.935672514619883,
|
| 342 |
+
"grad_norm": 4.884040832519531,
|
| 343 |
+
"learning_rate": 3.440545808966862e-05,
|
| 344 |
+
"loss": 3.5534,
|
| 345 |
+
"step": 480
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.9551656920077972,
|
| 349 |
+
"grad_norm": 5.428916931152344,
|
| 350 |
+
"learning_rate": 3.4080571799870045e-05,
|
| 351 |
+
"loss": 3.3919,
|
| 352 |
+
"step": 490
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.9746588693957114,
|
| 356 |
+
"grad_norm": 5.549111843109131,
|
| 357 |
+
"learning_rate": 3.3755685510071476e-05,
|
| 358 |
+
"loss": 3.2799,
|
| 359 |
+
"step": 500
|
| 360 |
+
}
|
| 361 |
+
],
|
| 362 |
+
"logging_steps": 10,
|
| 363 |
+
"max_steps": 1539,
|
| 364 |
+
"num_input_tokens_seen": 0,
|
| 365 |
+
"num_train_epochs": 3,
|
| 366 |
+
"save_steps": 500,
|
| 367 |
+
"stateful_callbacks": {
|
| 368 |
+
"TrainerControl": {
|
| 369 |
+
"args": {
|
| 370 |
+
"should_epoch_stop": false,
|
| 371 |
+
"should_evaluate": false,
|
| 372 |
+
"should_log": false,
|
| 373 |
+
"should_save": true,
|
| 374 |
+
"should_training_stop": false
|
| 375 |
+
},
|
| 376 |
+
"attributes": {}
|
| 377 |
+
}
|
| 378 |
+
},
|
| 379 |
+
"total_flos": 1045168128000000.0,
|
| 380 |
+
"train_batch_size": 8,
|
| 381 |
+
"trial_name": null,
|
| 382 |
+
"trial_params": null
|
| 383 |
+
}
|
checkpoint-500/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6b0e13a27dc9eb93f3cc3f0f69eeb907a00fef0cf3de646ea9710a3354ece317
|
| 3 |
+
size 5304
|
runs/Feb17_10-24-44_8cfcbcdf5939/events.out.tfevents.1739787888.8cfcbcdf5939.1018.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:27b1419e5806a1ecb99be75f28106f64e4e9cda164eb59a7000a8ccf46bbbd40
|
| 3 |
+
size 5325
|
runs/Feb17_10-27-50_8cfcbcdf5939/events.out.tfevents.1739788072.8cfcbcdf5939.1018.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b44e8dc60711ae0fd040a5cea28e074fbcbd1019be2e92065f6977bf5d56ca9e
|
| 3 |
+
size 38726
|