End of training
Browse files- all_results.json +14 -0
- eval_results.json +9 -0
- pytorch_model.bin +1 -1
- train_results.json +8 -0
- trainer_state.json +25 -0
all_results.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.06,
|
| 3 |
+
"eval_loss": 171.4817352294922,
|
| 4 |
+
"eval_runtime": 11.6015,
|
| 5 |
+
"eval_samples": 301,
|
| 6 |
+
"eval_samples_per_second": 25.945,
|
| 7 |
+
"eval_steps_per_second": 1.638,
|
| 8 |
+
"eval_wer": 1.2434017595307918,
|
| 9 |
+
"train_loss": 94.35431518554688,
|
| 10 |
+
"train_runtime": 13.6182,
|
| 11 |
+
"train_samples": 704,
|
| 12 |
+
"train_samples_per_second": 2.937,
|
| 13 |
+
"train_steps_per_second": 0.734
|
| 14 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.06,
|
| 3 |
+
"eval_loss": 171.4817352294922,
|
| 4 |
+
"eval_runtime": 11.6015,
|
| 5 |
+
"eval_samples": 301,
|
| 6 |
+
"eval_samples_per_second": 25.945,
|
| 7 |
+
"eval_steps_per_second": 1.638,
|
| 8 |
+
"eval_wer": 1.2434017595307918
|
| 9 |
+
}
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 143910
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3840384d7411963f82280d5b229da655effefa63b316d8c8eb398cda9ca70344
|
| 3 |
size 143910
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.06,
|
| 3 |
+
"train_loss": 94.35431518554688,
|
| 4 |
+
"train_runtime": 13.6182,
|
| 5 |
+
"train_samples": 704,
|
| 6 |
+
"train_samples_per_second": 2.937,
|
| 7 |
+
"train_steps_per_second": 0.734
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.056818181818181816,
|
| 5 |
+
"global_step": 10,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.06,
|
| 12 |
+
"step": 10,
|
| 13 |
+
"total_flos": 988984857600.0,
|
| 14 |
+
"train_loss": 94.35431518554688,
|
| 15 |
+
"train_runtime": 13.6182,
|
| 16 |
+
"train_samples_per_second": 2.937,
|
| 17 |
+
"train_steps_per_second": 0.734
|
| 18 |
+
}
|
| 19 |
+
],
|
| 20 |
+
"max_steps": 10,
|
| 21 |
+
"num_train_epochs": 1,
|
| 22 |
+
"total_flos": 988984857600.0,
|
| 23 |
+
"trial_name": null,
|
| 24 |
+
"trial_params": null
|
| 25 |
+
}
|