Javi
commited on
Commit
·
ca67487
1
Parent(s):
89605b6
reduced bs results in slight improvement of metrics
Browse files- README.md +13 -12
- all_results.json +12 -12
- eval_nbest_predictions.json +2 -2
- eval_null_odds.json +0 -0
- eval_predictions.json +0 -0
- eval_results.json +8 -8
- pytorch_model.bin +1 -1
- train_results.json +4 -4
- trainer_state.json +252 -30
- training_args.bin +2 -2
README.md
CHANGED
|
@@ -10,27 +10,27 @@ widget:
|
|
| 10 |
|
| 11 |
This is the [BSC-TeMU/roberta-base-bne](https://huggingface.co/BSC-TeMU/roberta-base-bne) model ([source](https://github.com/PlanTL-SANIDAD/lm-spanish)) trained on the [squad_es v2.0.0](https://huggingface.co/datasets/squad_es) dataset ([source](https://github.com/ccasimiro88/TranslateAlignRetrieve)).
|
| 12 |
|
| 13 |
-
Current achievement: em=58.
|
| 14 |
|
| 15 |
Results:
|
| 16 |
|
| 17 |
```
|
| 18 |
{
|
| 19 |
"epoch": 4.0,
|
| 20 |
-
"eval_HasAns_exact": 48.
|
| 21 |
-
"eval_HasAns_f1":
|
| 22 |
"eval_HasAns_total": 5928,
|
| 23 |
-
"eval_NoAns_exact": 69.
|
| 24 |
-
"eval_NoAns_f1": 69.
|
| 25 |
"eval_NoAns_total": 5930,
|
| 26 |
-
"eval_best_exact": 58.
|
| 27 |
"eval_best_exact_thresh": 0.0,
|
| 28 |
-
"eval_best_f1":
|
| 29 |
"eval_best_f1_thresh": 0.0,
|
| 30 |
-
"eval_exact": 58.
|
| 31 |
-
"eval_f1":
|
| 32 |
"eval_samples": 12211,
|
| 33 |
-
"eval_total": 11858
|
| 34 |
}
|
| 35 |
```
|
| 36 |
|
|
@@ -48,8 +48,9 @@ python -m torch.distributed.launch --nproc_per_node=3 ./run_qa.py \
|
|
| 48 |
--max_seq_length 384 \
|
| 49 |
--doc_stride 128 \
|
| 50 |
--output_dir ./models/roberta-base-bne-squad-2.0-es/ \
|
| 51 |
-
--per_device_eval_batch_size=
|
| 52 |
-
--per_device_train_batch_size=
|
| 53 |
--version_2_with_negative \
|
| 54 |
--ddp_find_unused_parameters=False \
|
|
|
|
| 55 |
```
|
|
|
|
| 10 |
|
| 11 |
This is the [BSC-TeMU/roberta-base-bne](https://huggingface.co/BSC-TeMU/roberta-base-bne) model ([source](https://github.com/PlanTL-SANIDAD/lm-spanish)) trained on the [squad_es v2.0.0](https://huggingface.co/datasets/squad_es) dataset ([source](https://github.com/ccasimiro88/TranslateAlignRetrieve)).
|
| 12 |
|
| 13 |
+
Current achievement: em=58.80, f1=67.40
|
| 14 |
|
| 15 |
Results:
|
| 16 |
|
| 17 |
```
|
| 18 |
{
|
| 19 |
"epoch": 4.0,
|
| 20 |
+
"eval_HasAns_exact": 48.51551956815115,
|
| 21 |
+
"eval_HasAns_f1": 65.70745010262016,
|
| 22 |
"eval_HasAns_total": 5928,
|
| 23 |
+
"eval_NoAns_exact": 69.0893760539629,
|
| 24 |
+
"eval_NoAns_f1": 69.0893760539629,
|
| 25 |
"eval_NoAns_total": 5930,
|
| 26 |
+
"eval_best_exact": 58.804182830156854,
|
| 27 |
"eval_best_exact_thresh": 0.0,
|
| 28 |
+
"eval_best_f1": 67.39869828034618,
|
| 29 |
"eval_best_f1_thresh": 0.0,
|
| 30 |
+
"eval_exact": 58.804182830156854,
|
| 31 |
+
"eval_f1": 67.39869828034568,
|
| 32 |
"eval_samples": 12211,
|
| 33 |
+
"eval_total": 11858
|
| 34 |
}
|
| 35 |
```
|
| 36 |
|
|
|
|
| 48 |
--max_seq_length 384 \
|
| 49 |
--doc_stride 128 \
|
| 50 |
--output_dir ./models/roberta-base-bne-squad-2.0-es/ \
|
| 51 |
+
--per_device_eval_batch_size=8 \
|
| 52 |
+
--per_device_train_batch_size=8 \
|
| 53 |
--version_2_with_negative \
|
| 54 |
--ddp_find_unused_parameters=False \
|
| 55 |
+
--overwrite_output_dir \
|
| 56 |
```
|
all_results.json
CHANGED
|
@@ -1,22 +1,22 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 4.0,
|
| 3 |
-
"eval_HasAns_exact": 48.
|
| 4 |
-
"eval_HasAns_f1":
|
| 5 |
"eval_HasAns_total": 5928,
|
| 6 |
-
"eval_NoAns_exact": 69.
|
| 7 |
-
"eval_NoAns_f1": 69.
|
| 8 |
"eval_NoAns_total": 5930,
|
| 9 |
-
"eval_best_exact": 58.
|
| 10 |
"eval_best_exact_thresh": 0.0,
|
| 11 |
-
"eval_best_f1":
|
| 12 |
"eval_best_f1_thresh": 0.0,
|
| 13 |
-
"eval_exact": 58.
|
| 14 |
-
"eval_f1":
|
| 15 |
"eval_samples": 12211,
|
| 16 |
"eval_total": 11858,
|
| 17 |
-
"train_loss": 0.
|
| 18 |
-
"train_runtime":
|
| 19 |
"train_samples": 132248,
|
| 20 |
-
"train_samples_per_second":
|
| 21 |
-
"train_steps_per_second":
|
| 22 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 4.0,
|
| 3 |
+
"eval_HasAns_exact": 48.51551956815115,
|
| 4 |
+
"eval_HasAns_f1": 65.70745010262016,
|
| 5 |
"eval_HasAns_total": 5928,
|
| 6 |
+
"eval_NoAns_exact": 69.0893760539629,
|
| 7 |
+
"eval_NoAns_f1": 69.0893760539629,
|
| 8 |
"eval_NoAns_total": 5930,
|
| 9 |
+
"eval_best_exact": 58.804182830156854,
|
| 10 |
"eval_best_exact_thresh": 0.0,
|
| 11 |
+
"eval_best_f1": 67.39869828034618,
|
| 12 |
"eval_best_f1_thresh": 0.0,
|
| 13 |
+
"eval_exact": 58.804182830156854,
|
| 14 |
+
"eval_f1": 67.39869828034568,
|
| 15 |
"eval_samples": 12211,
|
| 16 |
"eval_total": 11858,
|
| 17 |
+
"train_loss": 0.7753000720575617,
|
| 18 |
+
"train_runtime": 7643.3678,
|
| 19 |
"train_samples": 132248,
|
| 20 |
+
"train_samples_per_second": 69.209,
|
| 21 |
+
"train_steps_per_second": 2.884
|
| 22 |
}
|
eval_nbest_predictions.json
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a3e881209214b8a20c2c8de0d9d0f656175171f9edd8e65d58b176ed2985955b
|
| 3 |
+
size 56429258
|
eval_null_odds.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
eval_predictions.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
eval_results.json
CHANGED
|
@@ -1,17 +1,17 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 4.0,
|
| 3 |
-
"eval_HasAns_exact": 48.
|
| 4 |
-
"eval_HasAns_f1":
|
| 5 |
"eval_HasAns_total": 5928,
|
| 6 |
-
"eval_NoAns_exact": 69.
|
| 7 |
-
"eval_NoAns_f1": 69.
|
| 8 |
"eval_NoAns_total": 5930,
|
| 9 |
-
"eval_best_exact": 58.
|
| 10 |
"eval_best_exact_thresh": 0.0,
|
| 11 |
-
"eval_best_f1":
|
| 12 |
"eval_best_f1_thresh": 0.0,
|
| 13 |
-
"eval_exact": 58.
|
| 14 |
-
"eval_f1":
|
| 15 |
"eval_samples": 12211,
|
| 16 |
"eval_total": 11858
|
| 17 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 4.0,
|
| 3 |
+
"eval_HasAns_exact": 48.51551956815115,
|
| 4 |
+
"eval_HasAns_f1": 65.70745010262016,
|
| 5 |
"eval_HasAns_total": 5928,
|
| 6 |
+
"eval_NoAns_exact": 69.0893760539629,
|
| 7 |
+
"eval_NoAns_f1": 69.0893760539629,
|
| 8 |
"eval_NoAns_total": 5930,
|
| 9 |
+
"eval_best_exact": 58.804182830156854,
|
| 10 |
"eval_best_exact_thresh": 0.0,
|
| 11 |
+
"eval_best_f1": 67.39869828034618,
|
| 12 |
"eval_best_f1_thresh": 0.0,
|
| 13 |
+
"eval_exact": 58.804182830156854,
|
| 14 |
+
"eval_f1": 67.39869828034568,
|
| 15 |
"eval_samples": 12211,
|
| 16 |
"eval_total": 11858
|
| 17 |
}
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 496301809
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7b21b5926e3de6d0e3f479d2586f077c03d1f91070a9ac58e9c17acea60ba920
|
| 3 |
size 496301809
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 4.0,
|
| 3 |
-
"train_loss": 0.
|
| 4 |
-
"train_runtime":
|
| 5 |
"train_samples": 132248,
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second":
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 4.0,
|
| 3 |
+
"train_loss": 0.7753000720575617,
|
| 4 |
+
"train_runtime": 7643.3678,
|
| 5 |
"train_samples": 132248,
|
| 6 |
+
"train_samples_per_second": 69.209,
|
| 7 |
+
"train_steps_per_second": 2.884
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -2,66 +2,288 @@
|
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
"epoch": 4.0,
|
| 5 |
-
"global_step":
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
-
"epoch": 0.
|
| 12 |
-
"learning_rate": 2.
|
| 13 |
-
"loss":
|
| 14 |
"step": 500
|
| 15 |
},
|
| 16 |
{
|
| 17 |
-
"epoch":
|
| 18 |
-
"learning_rate": 2.
|
| 19 |
-
"loss": 1.
|
| 20 |
"step": 1000
|
| 21 |
},
|
| 22 |
{
|
| 23 |
-
"epoch":
|
| 24 |
-
"learning_rate":
|
| 25 |
-
"loss": 1.
|
| 26 |
"step": 1500
|
| 27 |
},
|
| 28 |
{
|
| 29 |
-
"epoch":
|
| 30 |
-
"learning_rate":
|
| 31 |
-
"loss":
|
| 32 |
"step": 2000
|
| 33 |
},
|
| 34 |
{
|
| 35 |
-
"epoch":
|
| 36 |
-
"learning_rate":
|
| 37 |
-
"loss":
|
| 38 |
"step": 2500
|
| 39 |
},
|
| 40 |
{
|
| 41 |
-
"epoch":
|
| 42 |
-
"learning_rate":
|
| 43 |
-
"loss":
|
| 44 |
"step": 3000
|
| 45 |
},
|
| 46 |
{
|
| 47 |
-
"epoch":
|
| 48 |
-
"learning_rate":
|
| 49 |
-
"loss":
|
| 50 |
"step": 3500
|
| 51 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 52 |
{
|
| 53 |
"epoch": 4.0,
|
| 54 |
-
"step":
|
| 55 |
-
"total_flos": 1.
|
| 56 |
-
"train_loss": 0.
|
| 57 |
-
"train_runtime":
|
| 58 |
-
"train_samples_per_second":
|
| 59 |
-
"train_steps_per_second":
|
| 60 |
}
|
| 61 |
],
|
| 62 |
-
"max_steps":
|
| 63 |
"num_train_epochs": 4,
|
| 64 |
-
"total_flos": 1.
|
| 65 |
"trial_name": null,
|
| 66 |
"trial_params": null
|
| 67 |
}
|
|
|
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
"epoch": 4.0,
|
| 5 |
+
"global_step": 22044,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
+
"epoch": 0.09,
|
| 12 |
+
"learning_rate": 2.9319542732716386e-05,
|
| 13 |
+
"loss": 2.1551,
|
| 14 |
"step": 500
|
| 15 |
},
|
| 16 |
{
|
| 17 |
+
"epoch": 0.18,
|
| 18 |
+
"learning_rate": 2.863908546543277e-05,
|
| 19 |
+
"loss": 1.7229,
|
| 20 |
"step": 1000
|
| 21 |
},
|
| 22 |
{
|
| 23 |
+
"epoch": 0.27,
|
| 24 |
+
"learning_rate": 2.7958628198149155e-05,
|
| 25 |
+
"loss": 1.5902,
|
| 26 |
"step": 1500
|
| 27 |
},
|
| 28 |
{
|
| 29 |
+
"epoch": 0.36,
|
| 30 |
+
"learning_rate": 2.7278170930865543e-05,
|
| 31 |
+
"loss": 1.5139,
|
| 32 |
"step": 2000
|
| 33 |
},
|
| 34 |
{
|
| 35 |
+
"epoch": 0.45,
|
| 36 |
+
"learning_rate": 2.6597713663581925e-05,
|
| 37 |
+
"loss": 1.4848,
|
| 38 |
"step": 2500
|
| 39 |
},
|
| 40 |
{
|
| 41 |
+
"epoch": 0.54,
|
| 42 |
+
"learning_rate": 2.5917256396298313e-05,
|
| 43 |
+
"loss": 1.4533,
|
| 44 |
"step": 3000
|
| 45 |
},
|
| 46 |
{
|
| 47 |
+
"epoch": 0.64,
|
| 48 |
+
"learning_rate": 2.5236799129014698e-05,
|
| 49 |
+
"loss": 1.3801,
|
| 50 |
"step": 3500
|
| 51 |
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.73,
|
| 54 |
+
"learning_rate": 2.4556341861731083e-05,
|
| 55 |
+
"loss": 1.3509,
|
| 56 |
+
"step": 4000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.82,
|
| 60 |
+
"learning_rate": 2.3875884594447467e-05,
|
| 61 |
+
"loss": 1.3392,
|
| 62 |
+
"step": 4500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.91,
|
| 66 |
+
"learning_rate": 2.3195427327163856e-05,
|
| 67 |
+
"loss": 1.3046,
|
| 68 |
+
"step": 5000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 1.0,
|
| 72 |
+
"learning_rate": 2.251497005988024e-05,
|
| 73 |
+
"loss": 1.275,
|
| 74 |
+
"step": 5500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 1.09,
|
| 78 |
+
"learning_rate": 2.1834512792596625e-05,
|
| 79 |
+
"loss": 0.9101,
|
| 80 |
+
"step": 6000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 1.18,
|
| 84 |
+
"learning_rate": 2.115405552531301e-05,
|
| 85 |
+
"loss": 0.937,
|
| 86 |
+
"step": 6500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 1.27,
|
| 90 |
+
"learning_rate": 2.0473598258029398e-05,
|
| 91 |
+
"loss": 0.9173,
|
| 92 |
+
"step": 7000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 1.36,
|
| 96 |
+
"learning_rate": 1.979314099074578e-05,
|
| 97 |
+
"loss": 0.9003,
|
| 98 |
+
"step": 7500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.45,
|
| 102 |
+
"learning_rate": 1.9112683723462168e-05,
|
| 103 |
+
"loss": 0.9009,
|
| 104 |
+
"step": 8000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 1.54,
|
| 108 |
+
"learning_rate": 1.8432226456178553e-05,
|
| 109 |
+
"loss": 0.9359,
|
| 110 |
+
"step": 8500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 1.63,
|
| 114 |
+
"learning_rate": 1.7751769188894937e-05,
|
| 115 |
+
"loss": 0.9038,
|
| 116 |
+
"step": 9000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 1.72,
|
| 120 |
+
"learning_rate": 1.7071311921611322e-05,
|
| 121 |
+
"loss": 0.9211,
|
| 122 |
+
"step": 9500
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 1.81,
|
| 126 |
+
"learning_rate": 1.639085465432771e-05,
|
| 127 |
+
"loss": 0.8918,
|
| 128 |
+
"step": 10000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 1.91,
|
| 132 |
+
"learning_rate": 1.5710397387044092e-05,
|
| 133 |
+
"loss": 0.9076,
|
| 134 |
+
"step": 10500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 2.0,
|
| 138 |
+
"learning_rate": 1.502994011976048e-05,
|
| 139 |
+
"loss": 0.8857,
|
| 140 |
+
"step": 11000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 2.09,
|
| 144 |
+
"learning_rate": 1.4349482852476866e-05,
|
| 145 |
+
"loss": 0.5123,
|
| 146 |
+
"step": 11500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 2.18,
|
| 150 |
+
"learning_rate": 1.366902558519325e-05,
|
| 151 |
+
"loss": 0.481,
|
| 152 |
+
"step": 12000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 2.27,
|
| 156 |
+
"learning_rate": 1.2988568317909634e-05,
|
| 157 |
+
"loss": 0.4587,
|
| 158 |
+
"step": 12500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.36,
|
| 162 |
+
"learning_rate": 1.2308111050626021e-05,
|
| 163 |
+
"loss": 0.4773,
|
| 164 |
+
"step": 13000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 2.45,
|
| 168 |
+
"learning_rate": 1.1627653783342406e-05,
|
| 169 |
+
"loss": 0.4962,
|
| 170 |
+
"step": 13500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 2.54,
|
| 174 |
+
"learning_rate": 1.094719651605879e-05,
|
| 175 |
+
"loss": 0.5124,
|
| 176 |
+
"step": 14000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 2.63,
|
| 180 |
+
"learning_rate": 1.0266739248775177e-05,
|
| 181 |
+
"loss": 0.4949,
|
| 182 |
+
"step": 14500
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 2.72,
|
| 186 |
+
"learning_rate": 9.586281981491562e-06,
|
| 187 |
+
"loss": 0.4931,
|
| 188 |
+
"step": 15000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 2.81,
|
| 192 |
+
"learning_rate": 8.905824714207947e-06,
|
| 193 |
+
"loss": 0.4849,
|
| 194 |
+
"step": 15500
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 2.9,
|
| 198 |
+
"learning_rate": 8.225367446924333e-06,
|
| 199 |
+
"loss": 0.4878,
|
| 200 |
+
"step": 16000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 2.99,
|
| 204 |
+
"learning_rate": 7.544910179640718e-06,
|
| 205 |
+
"loss": 0.4949,
|
| 206 |
+
"step": 16500
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 3.08,
|
| 210 |
+
"learning_rate": 6.864452912357104e-06,
|
| 211 |
+
"loss": 0.2131,
|
| 212 |
+
"step": 17000
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 3.18,
|
| 216 |
+
"learning_rate": 6.183995645073489e-06,
|
| 217 |
+
"loss": 0.2022,
|
| 218 |
+
"step": 17500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.27,
|
| 222 |
+
"learning_rate": 5.503538377789875e-06,
|
| 223 |
+
"loss": 0.2071,
|
| 224 |
+
"step": 18000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 3.36,
|
| 228 |
+
"learning_rate": 4.82308111050626e-06,
|
| 229 |
+
"loss": 0.2067,
|
| 230 |
+
"step": 18500
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 3.45,
|
| 234 |
+
"learning_rate": 4.142623843222645e-06,
|
| 235 |
+
"loss": 0.2029,
|
| 236 |
+
"step": 19000
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 3.54,
|
| 240 |
+
"learning_rate": 3.4621665759390314e-06,
|
| 241 |
+
"loss": 0.1866,
|
| 242 |
+
"step": 19500
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 3.63,
|
| 246 |
+
"learning_rate": 2.7817093086554166e-06,
|
| 247 |
+
"loss": 0.2056,
|
| 248 |
+
"step": 20000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 3.72,
|
| 252 |
+
"learning_rate": 2.101252041371802e-06,
|
| 253 |
+
"loss": 0.1925,
|
| 254 |
+
"step": 20500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 3.81,
|
| 258 |
+
"learning_rate": 1.4207947740881873e-06,
|
| 259 |
+
"loss": 0.1866,
|
| 260 |
+
"step": 21000
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 3.9,
|
| 264 |
+
"learning_rate": 7.403375068045727e-07,
|
| 265 |
+
"loss": 0.1952,
|
| 266 |
+
"step": 21500
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 3.99,
|
| 270 |
+
"learning_rate": 5.988023952095808e-08,
|
| 271 |
+
"loss": 0.1866,
|
| 272 |
+
"step": 22000
|
| 273 |
+
},
|
| 274 |
{
|
| 275 |
"epoch": 4.0,
|
| 276 |
+
"step": 22044,
|
| 277 |
+
"total_flos": 1.0366869961703424e+17,
|
| 278 |
+
"train_loss": 0.7753000720575617,
|
| 279 |
+
"train_runtime": 7643.3678,
|
| 280 |
+
"train_samples_per_second": 69.209,
|
| 281 |
+
"train_steps_per_second": 2.884
|
| 282 |
}
|
| 283 |
],
|
| 284 |
+
"max_steps": 22044,
|
| 285 |
"num_train_epochs": 4,
|
| 286 |
+
"total_flos": 1.0366869961703424e+17,
|
| 287 |
"trial_name": null,
|
| 288 |
"trial_params": null
|
| 289 |
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c3771c9a065d16d78648b4802f002cd2b7aee20b36dedb4d59a7096165d89525
|
| 3 |
+
size 2671
|