Commit
•
04988e0
1
Parent(s):
fa9ae38
End of training
Browse files- README.md +15 -2
- all_results.json +15 -0
- eval_results.json +9 -0
- runs/Aug23_07-57-49_c73520ec3129/events.out.tfevents.1724457272.c73520ec3129.33.1 +3 -0
- train_results.json +9 -0
- trainer_state.json +1487 -0
README.md
CHANGED
|
@@ -3,11 +3,24 @@ license: apache-2.0
|
|
| 3 |
base_model: openai/whisper-large-v3
|
| 4 |
tags:
|
| 5 |
- generated_from_trainer
|
|
|
|
|
|
|
| 6 |
metrics:
|
| 7 |
- wer
|
| 8 |
model-index:
|
| 9 |
- name: whisper-large-v3-ft-cv-cy-train-all
|
| 10 |
-
results:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11 |
---
|
| 12 |
|
| 13 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
@@ -15,7 +28,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 15 |
|
| 16 |
# whisper-large-v3-ft-cv-cy-train-all
|
| 17 |
|
| 18 |
-
This model is a fine-tuned version of [openai/whisper-large-v3](https://huggingface.co/openai/whisper-large-v3) on
|
| 19 |
It achieves the following results on the evaluation set:
|
| 20 |
- Loss: 0.3638
|
| 21 |
- Wer: 0.1817
|
|
|
|
| 3 |
base_model: openai/whisper-large-v3
|
| 4 |
tags:
|
| 5 |
- generated_from_trainer
|
| 6 |
+
datasets:
|
| 7 |
+
- DewiBrynJones/commonvoice_18_0_cy
|
| 8 |
metrics:
|
| 9 |
- wer
|
| 10 |
model-index:
|
| 11 |
- name: whisper-large-v3-ft-cv-cy-train-all
|
| 12 |
+
results:
|
| 13 |
+
- task:
|
| 14 |
+
name: Automatic Speech Recognition
|
| 15 |
+
type: automatic-speech-recognition
|
| 16 |
+
dataset:
|
| 17 |
+
name: DewiBrynJones/commonvoice_18_0_cy default
|
| 18 |
+
type: DewiBrynJones/commonvoice_18_0_cy
|
| 19 |
+
args: default
|
| 20 |
+
metrics:
|
| 21 |
+
- name: Wer
|
| 22 |
+
type: wer
|
| 23 |
+
value: 0.18173684838363355
|
| 24 |
---
|
| 25 |
|
| 26 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
|
| 28 |
|
| 29 |
# whisper-large-v3-ft-cv-cy-train-all
|
| 30 |
|
| 31 |
+
This model is a fine-tuned version of [openai/whisper-large-v3](https://huggingface.co/openai/whisper-large-v3) on the DewiBrynJones/commonvoice_18_0_cy default dataset.
|
| 32 |
It achieves the following results on the evaluation set:
|
| 33 |
- Loss: 0.3638
|
| 34 |
- Wer: 0.1817
|
all_results.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.727626459143968,
|
| 3 |
+
"eval_loss": 0.36376532912254333,
|
| 4 |
+
"eval_runtime": 2112.3804,
|
| 5 |
+
"eval_samples": 5381,
|
| 6 |
+
"eval_samples_per_second": 2.547,
|
| 7 |
+
"eval_steps_per_second": 0.16,
|
| 8 |
+
"eval_wer": 0.18173684838363355,
|
| 9 |
+
"total_flos": 5.4332453168676864e+20,
|
| 10 |
+
"train_loss": 0.06844911024216563,
|
| 11 |
+
"train_runtime": 58020.7767,
|
| 12 |
+
"train_samples": 16439,
|
| 13 |
+
"train_samples_per_second": 2.758,
|
| 14 |
+
"train_steps_per_second": 0.086
|
| 15 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.727626459143968,
|
| 3 |
+
"eval_loss": 0.36376532912254333,
|
| 4 |
+
"eval_runtime": 2112.3804,
|
| 5 |
+
"eval_samples": 5381,
|
| 6 |
+
"eval_samples_per_second": 2.547,
|
| 7 |
+
"eval_steps_per_second": 0.16,
|
| 8 |
+
"eval_wer": 0.18173684838363355
|
| 9 |
+
}
|
runs/Aug23_07-57-49_c73520ec3129/events.out.tfevents.1724457272.c73520ec3129.33.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:99b9f8eab00c0349bb0cfe1ce30796c68b125170991b8dc6785cc8855fb04ee1
|
| 3 |
+
size 406
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.727626459143968,
|
| 3 |
+
"total_flos": 5.4332453168676864e+20,
|
| 4 |
+
"train_loss": 0.06844911024216563,
|
| 5 |
+
"train_runtime": 58020.7767,
|
| 6 |
+
"train_samples": 16439,
|
| 7 |
+
"train_samples_per_second": 2.758,
|
| 8 |
+
"train_steps_per_second": 0.086
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1487 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 9.727626459143968,
|
| 5 |
+
"eval_steps": 1000,
|
| 6 |
+
"global_step": 5000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.048638132295719845,
|
| 13 |
+
"grad_norm": 7.16464900970459,
|
| 14 |
+
"learning_rate": 5.000000000000001e-07,
|
| 15 |
+
"loss": 1.0297,
|
| 16 |
+
"step": 25
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.09727626459143969,
|
| 20 |
+
"grad_norm": 4.811949729919434,
|
| 21 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 22 |
+
"loss": 0.786,
|
| 23 |
+
"step": 50
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.14591439688715954,
|
| 27 |
+
"grad_norm": 4.57177734375,
|
| 28 |
+
"learning_rate": 1.5e-06,
|
| 29 |
+
"loss": 0.5046,
|
| 30 |
+
"step": 75
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.19455252918287938,
|
| 34 |
+
"grad_norm": 4.517965793609619,
|
| 35 |
+
"learning_rate": 2.0000000000000003e-06,
|
| 36 |
+
"loss": 0.4671,
|
| 37 |
+
"step": 100
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.24319066147859922,
|
| 41 |
+
"grad_norm": 4.065127849578857,
|
| 42 |
+
"learning_rate": 2.5e-06,
|
| 43 |
+
"loss": 0.4381,
|
| 44 |
+
"step": 125
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.2918287937743191,
|
| 48 |
+
"grad_norm": 3.5869154930114746,
|
| 49 |
+
"learning_rate": 3e-06,
|
| 50 |
+
"loss": 0.4062,
|
| 51 |
+
"step": 150
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.3404669260700389,
|
| 55 |
+
"grad_norm": 4.262431621551514,
|
| 56 |
+
"learning_rate": 3.5e-06,
|
| 57 |
+
"loss": 0.3951,
|
| 58 |
+
"step": 175
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.38910505836575876,
|
| 62 |
+
"grad_norm": 4.2176713943481445,
|
| 63 |
+
"learning_rate": 4.000000000000001e-06,
|
| 64 |
+
"loss": 0.3862,
|
| 65 |
+
"step": 200
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.4377431906614786,
|
| 69 |
+
"grad_norm": 3.8358681201934814,
|
| 70 |
+
"learning_rate": 4.5e-06,
|
| 71 |
+
"loss": 0.364,
|
| 72 |
+
"step": 225
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.48638132295719844,
|
| 76 |
+
"grad_norm": 3.8311147689819336,
|
| 77 |
+
"learning_rate": 5e-06,
|
| 78 |
+
"loss": 0.3425,
|
| 79 |
+
"step": 250
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.5350194552529183,
|
| 83 |
+
"grad_norm": 3.7137532234191895,
|
| 84 |
+
"learning_rate": 5.500000000000001e-06,
|
| 85 |
+
"loss": 0.339,
|
| 86 |
+
"step": 275
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.5836575875486382,
|
| 90 |
+
"grad_norm": 4.2145280838012695,
|
| 91 |
+
"learning_rate": 6e-06,
|
| 92 |
+
"loss": 0.3393,
|
| 93 |
+
"step": 300
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.632295719844358,
|
| 97 |
+
"grad_norm": 3.540754556655884,
|
| 98 |
+
"learning_rate": 6.5000000000000004e-06,
|
| 99 |
+
"loss": 0.3154,
|
| 100 |
+
"step": 325
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.6809338521400778,
|
| 104 |
+
"grad_norm": 3.7712759971618652,
|
| 105 |
+
"learning_rate": 7e-06,
|
| 106 |
+
"loss": 0.319,
|
| 107 |
+
"step": 350
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.7295719844357976,
|
| 111 |
+
"grad_norm": 3.6147518157958984,
|
| 112 |
+
"learning_rate": 7.500000000000001e-06,
|
| 113 |
+
"loss": 0.3127,
|
| 114 |
+
"step": 375
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.7782101167315175,
|
| 118 |
+
"grad_norm": 3.6948869228363037,
|
| 119 |
+
"learning_rate": 8.000000000000001e-06,
|
| 120 |
+
"loss": 0.2965,
|
| 121 |
+
"step": 400
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.8268482490272373,
|
| 125 |
+
"grad_norm": 3.350135087966919,
|
| 126 |
+
"learning_rate": 8.5e-06,
|
| 127 |
+
"loss": 0.3001,
|
| 128 |
+
"step": 425
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.8754863813229572,
|
| 132 |
+
"grad_norm": 3.7692646980285645,
|
| 133 |
+
"learning_rate": 9e-06,
|
| 134 |
+
"loss": 0.281,
|
| 135 |
+
"step": 450
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.9241245136186771,
|
| 139 |
+
"grad_norm": 3.2747561931610107,
|
| 140 |
+
"learning_rate": 9.5e-06,
|
| 141 |
+
"loss": 0.2707,
|
| 142 |
+
"step": 475
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.9727626459143969,
|
| 146 |
+
"grad_norm": 3.260094165802002,
|
| 147 |
+
"learning_rate": 1e-05,
|
| 148 |
+
"loss": 0.2708,
|
| 149 |
+
"step": 500
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 1.0214007782101167,
|
| 153 |
+
"grad_norm": 2.917370080947876,
|
| 154 |
+
"learning_rate": 9.944444444444445e-06,
|
| 155 |
+
"loss": 0.2415,
|
| 156 |
+
"step": 525
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 1.0700389105058365,
|
| 160 |
+
"grad_norm": 2.564760208129883,
|
| 161 |
+
"learning_rate": 9.88888888888889e-06,
|
| 162 |
+
"loss": 0.1714,
|
| 163 |
+
"step": 550
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 1.1186770428015564,
|
| 167 |
+
"grad_norm": 2.844514846801758,
|
| 168 |
+
"learning_rate": 9.833333333333333e-06,
|
| 169 |
+
"loss": 0.1842,
|
| 170 |
+
"step": 575
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 1.1673151750972763,
|
| 174 |
+
"grad_norm": 3.1123480796813965,
|
| 175 |
+
"learning_rate": 9.777777777777779e-06,
|
| 176 |
+
"loss": 0.171,
|
| 177 |
+
"step": 600
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 1.2159533073929962,
|
| 181 |
+
"grad_norm": 3.253563642501831,
|
| 182 |
+
"learning_rate": 9.722222222222223e-06,
|
| 183 |
+
"loss": 0.1773,
|
| 184 |
+
"step": 625
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 1.264591439688716,
|
| 188 |
+
"grad_norm": 2.3145253658294678,
|
| 189 |
+
"learning_rate": 9.666666666666667e-06,
|
| 190 |
+
"loss": 0.1779,
|
| 191 |
+
"step": 650
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 1.3132295719844358,
|
| 195 |
+
"grad_norm": 2.493607521057129,
|
| 196 |
+
"learning_rate": 9.611111111111112e-06,
|
| 197 |
+
"loss": 0.1717,
|
| 198 |
+
"step": 675
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 1.3618677042801557,
|
| 202 |
+
"grad_norm": 2.081571578979492,
|
| 203 |
+
"learning_rate": 9.555555555555556e-06,
|
| 204 |
+
"loss": 0.1747,
|
| 205 |
+
"step": 700
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 1.4105058365758754,
|
| 209 |
+
"grad_norm": 2.9359965324401855,
|
| 210 |
+
"learning_rate": 9.5e-06,
|
| 211 |
+
"loss": 0.1591,
|
| 212 |
+
"step": 725
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.4591439688715953,
|
| 216 |
+
"grad_norm": 2.9549202919006348,
|
| 217 |
+
"learning_rate": 9.444444444444445e-06,
|
| 218 |
+
"loss": 0.1556,
|
| 219 |
+
"step": 750
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 1.5077821011673151,
|
| 223 |
+
"grad_norm": 3.0995261669158936,
|
| 224 |
+
"learning_rate": 9.38888888888889e-06,
|
| 225 |
+
"loss": 0.1553,
|
| 226 |
+
"step": 775
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 1.556420233463035,
|
| 230 |
+
"grad_norm": 2.3519606590270996,
|
| 231 |
+
"learning_rate": 9.333333333333334e-06,
|
| 232 |
+
"loss": 0.1502,
|
| 233 |
+
"step": 800
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 1.605058365758755,
|
| 237 |
+
"grad_norm": 2.3519246578216553,
|
| 238 |
+
"learning_rate": 9.277777777777778e-06,
|
| 239 |
+
"loss": 0.1398,
|
| 240 |
+
"step": 825
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 1.6536964980544746,
|
| 244 |
+
"grad_norm": 2.292020797729492,
|
| 245 |
+
"learning_rate": 9.222222222222224e-06,
|
| 246 |
+
"loss": 0.1518,
|
| 247 |
+
"step": 850
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 1.7023346303501945,
|
| 251 |
+
"grad_norm": 2.7264564037323,
|
| 252 |
+
"learning_rate": 9.166666666666666e-06,
|
| 253 |
+
"loss": 0.1393,
|
| 254 |
+
"step": 875
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.7509727626459144,
|
| 258 |
+
"grad_norm": 2.202409267425537,
|
| 259 |
+
"learning_rate": 9.111111111111112e-06,
|
| 260 |
+
"loss": 0.1432,
|
| 261 |
+
"step": 900
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 1.7996108949416343,
|
| 265 |
+
"grad_norm": 2.6967380046844482,
|
| 266 |
+
"learning_rate": 9.055555555555556e-06,
|
| 267 |
+
"loss": 0.1492,
|
| 268 |
+
"step": 925
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 1.8482490272373542,
|
| 272 |
+
"grad_norm": 2.8609344959259033,
|
| 273 |
+
"learning_rate": 9e-06,
|
| 274 |
+
"loss": 0.1351,
|
| 275 |
+
"step": 950
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 1.8968871595330739,
|
| 279 |
+
"grad_norm": 1.9660629034042358,
|
| 280 |
+
"learning_rate": 8.944444444444446e-06,
|
| 281 |
+
"loss": 0.1378,
|
| 282 |
+
"step": 975
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 1.9455252918287937,
|
| 286 |
+
"grad_norm": 2.9408161640167236,
|
| 287 |
+
"learning_rate": 8.888888888888888e-06,
|
| 288 |
+
"loss": 0.1429,
|
| 289 |
+
"step": 1000
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 1.9455252918287937,
|
| 293 |
+
"eval_loss": 0.2754097580909729,
|
| 294 |
+
"eval_runtime": 2154.7003,
|
| 295 |
+
"eval_samples_per_second": 2.497,
|
| 296 |
+
"eval_steps_per_second": 0.156,
|
| 297 |
+
"eval_wer": 0.22082388961787744,
|
| 298 |
+
"step": 1000
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 1.9941634241245136,
|
| 302 |
+
"grad_norm": 2.337907075881958,
|
| 303 |
+
"learning_rate": 8.833333333333334e-06,
|
| 304 |
+
"loss": 0.137,
|
| 305 |
+
"step": 1025
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 2.0428015564202333,
|
| 309 |
+
"grad_norm": 1.9045900106430054,
|
| 310 |
+
"learning_rate": 8.777777777777778e-06,
|
| 311 |
+
"loss": 0.0728,
|
| 312 |
+
"step": 1050
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 2.0914396887159534,
|
| 316 |
+
"grad_norm": 2.0690503120422363,
|
| 317 |
+
"learning_rate": 8.722222222222224e-06,
|
| 318 |
+
"loss": 0.0615,
|
| 319 |
+
"step": 1075
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 2.140077821011673,
|
| 323 |
+
"grad_norm": 2.0722174644470215,
|
| 324 |
+
"learning_rate": 8.666666666666668e-06,
|
| 325 |
+
"loss": 0.0582,
|
| 326 |
+
"step": 1100
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 2.188715953307393,
|
| 330 |
+
"grad_norm": 1.7977361679077148,
|
| 331 |
+
"learning_rate": 8.611111111111112e-06,
|
| 332 |
+
"loss": 0.0589,
|
| 333 |
+
"step": 1125
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 2.237354085603113,
|
| 337 |
+
"grad_norm": 1.5375025272369385,
|
| 338 |
+
"learning_rate": 8.555555555555556e-06,
|
| 339 |
+
"loss": 0.0657,
|
| 340 |
+
"step": 1150
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 2.2859922178988326,
|
| 344 |
+
"grad_norm": 1.9439505338668823,
|
| 345 |
+
"learning_rate": 8.5e-06,
|
| 346 |
+
"loss": 0.0648,
|
| 347 |
+
"step": 1175
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 2.3346303501945527,
|
| 351 |
+
"grad_norm": 1.7909560203552246,
|
| 352 |
+
"learning_rate": 8.444444444444446e-06,
|
| 353 |
+
"loss": 0.0594,
|
| 354 |
+
"step": 1200
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 2.3832684824902723,
|
| 358 |
+
"grad_norm": 1.7498406171798706,
|
| 359 |
+
"learning_rate": 8.38888888888889e-06,
|
| 360 |
+
"loss": 0.0627,
|
| 361 |
+
"step": 1225
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 2.4319066147859925,
|
| 365 |
+
"grad_norm": 1.557906985282898,
|
| 366 |
+
"learning_rate": 8.333333333333334e-06,
|
| 367 |
+
"loss": 0.0585,
|
| 368 |
+
"step": 1250
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 2.480544747081712,
|
| 372 |
+
"grad_norm": 1.3372323513031006,
|
| 373 |
+
"learning_rate": 8.277777777777778e-06,
|
| 374 |
+
"loss": 0.0614,
|
| 375 |
+
"step": 1275
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 2.529182879377432,
|
| 379 |
+
"grad_norm": 1.8402326107025146,
|
| 380 |
+
"learning_rate": 8.222222222222222e-06,
|
| 381 |
+
"loss": 0.0601,
|
| 382 |
+
"step": 1300
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 2.5778210116731515,
|
| 386 |
+
"grad_norm": 2.003889560699463,
|
| 387 |
+
"learning_rate": 8.166666666666668e-06,
|
| 388 |
+
"loss": 0.0575,
|
| 389 |
+
"step": 1325
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 2.6264591439688716,
|
| 393 |
+
"grad_norm": 1.5208065509796143,
|
| 394 |
+
"learning_rate": 8.111111111111112e-06,
|
| 395 |
+
"loss": 0.0563,
|
| 396 |
+
"step": 1350
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 2.6750972762645917,
|
| 400 |
+
"grad_norm": 2.1009633541107178,
|
| 401 |
+
"learning_rate": 8.055555555555557e-06,
|
| 402 |
+
"loss": 0.0633,
|
| 403 |
+
"step": 1375
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 2.7237354085603114,
|
| 407 |
+
"grad_norm": 1.367943286895752,
|
| 408 |
+
"learning_rate": 8.000000000000001e-06,
|
| 409 |
+
"loss": 0.0572,
|
| 410 |
+
"step": 1400
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 2.772373540856031,
|
| 414 |
+
"grad_norm": 1.914768934249878,
|
| 415 |
+
"learning_rate": 7.944444444444445e-06,
|
| 416 |
+
"loss": 0.0541,
|
| 417 |
+
"step": 1425
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 2.8210116731517507,
|
| 421 |
+
"grad_norm": 2.1714184284210205,
|
| 422 |
+
"learning_rate": 7.88888888888889e-06,
|
| 423 |
+
"loss": 0.059,
|
| 424 |
+
"step": 1450
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 2.869649805447471,
|
| 428 |
+
"grad_norm": 1.951029658317566,
|
| 429 |
+
"learning_rate": 7.833333333333333e-06,
|
| 430 |
+
"loss": 0.0571,
|
| 431 |
+
"step": 1475
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 2.9182879377431905,
|
| 435 |
+
"grad_norm": 1.6378835439682007,
|
| 436 |
+
"learning_rate": 7.77777777777778e-06,
|
| 437 |
+
"loss": 0.0563,
|
| 438 |
+
"step": 1500
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 2.9669260700389106,
|
| 442 |
+
"grad_norm": 1.8460659980773926,
|
| 443 |
+
"learning_rate": 7.722222222222223e-06,
|
| 444 |
+
"loss": 0.0556,
|
| 445 |
+
"step": 1525
|
| 446 |
+
},
|
| 447 |
+
{
|
| 448 |
+
"epoch": 3.0155642023346303,
|
| 449 |
+
"grad_norm": 1.0332473516464233,
|
| 450 |
+
"learning_rate": 7.666666666666667e-06,
|
| 451 |
+
"loss": 0.045,
|
| 452 |
+
"step": 1550
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 3.06420233463035,
|
| 456 |
+
"grad_norm": 1.5045939683914185,
|
| 457 |
+
"learning_rate": 7.611111111111111e-06,
|
| 458 |
+
"loss": 0.0227,
|
| 459 |
+
"step": 1575
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 3.11284046692607,
|
| 463 |
+
"grad_norm": 2.1660637855529785,
|
| 464 |
+
"learning_rate": 7.555555555555556e-06,
|
| 465 |
+
"loss": 0.0275,
|
| 466 |
+
"step": 1600
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 3.1614785992217898,
|
| 470 |
+
"grad_norm": 1.159175992012024,
|
| 471 |
+
"learning_rate": 7.500000000000001e-06,
|
| 472 |
+
"loss": 0.024,
|
| 473 |
+
"step": 1625
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 3.21011673151751,
|
| 477 |
+
"grad_norm": 0.9240212440490723,
|
| 478 |
+
"learning_rate": 7.444444444444445e-06,
|
| 479 |
+
"loss": 0.0271,
|
| 480 |
+
"step": 1650
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 3.2587548638132295,
|
| 484 |
+
"grad_norm": 0.9973633885383606,
|
| 485 |
+
"learning_rate": 7.38888888888889e-06,
|
| 486 |
+
"loss": 0.0266,
|
| 487 |
+
"step": 1675
|
| 488 |
+
},
|
| 489 |
+
{
|
| 490 |
+
"epoch": 3.307392996108949,
|
| 491 |
+
"grad_norm": 1.5561943054199219,
|
| 492 |
+
"learning_rate": 7.333333333333333e-06,
|
| 493 |
+
"loss": 0.0267,
|
| 494 |
+
"step": 1700
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 3.3560311284046693,
|
| 498 |
+
"grad_norm": 1.2981663942337036,
|
| 499 |
+
"learning_rate": 7.277777777777778e-06,
|
| 500 |
+
"loss": 0.0284,
|
| 501 |
+
"step": 1725
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 3.404669260700389,
|
| 505 |
+
"grad_norm": 1.1085954904556274,
|
| 506 |
+
"learning_rate": 7.222222222222223e-06,
|
| 507 |
+
"loss": 0.0239,
|
| 508 |
+
"step": 1750
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"epoch": 3.453307392996109,
|
| 512 |
+
"grad_norm": 1.219834566116333,
|
| 513 |
+
"learning_rate": 7.166666666666667e-06,
|
| 514 |
+
"loss": 0.0244,
|
| 515 |
+
"step": 1775
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 3.501945525291829,
|
| 519 |
+
"grad_norm": 1.1478172540664673,
|
| 520 |
+
"learning_rate": 7.111111111111112e-06,
|
| 521 |
+
"loss": 0.0273,
|
| 522 |
+
"step": 1800
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 3.5505836575875485,
|
| 526 |
+
"grad_norm": 1.5546625852584839,
|
| 527 |
+
"learning_rate": 7.055555555555557e-06,
|
| 528 |
+
"loss": 0.0242,
|
| 529 |
+
"step": 1825
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"epoch": 3.5992217898832686,
|
| 533 |
+
"grad_norm": 1.5303540229797363,
|
| 534 |
+
"learning_rate": 7e-06,
|
| 535 |
+
"loss": 0.0248,
|
| 536 |
+
"step": 1850
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 3.6478599221789882,
|
| 540 |
+
"grad_norm": 1.2220957279205322,
|
| 541 |
+
"learning_rate": 6.944444444444445e-06,
|
| 542 |
+
"loss": 0.0268,
|
| 543 |
+
"step": 1875
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 3.6964980544747084,
|
| 547 |
+
"grad_norm": 1.076720952987671,
|
| 548 |
+
"learning_rate": 6.88888888888889e-06,
|
| 549 |
+
"loss": 0.0246,
|
| 550 |
+
"step": 1900
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 3.745136186770428,
|
| 554 |
+
"grad_norm": 0.6398268342018127,
|
| 555 |
+
"learning_rate": 6.833333333333334e-06,
|
| 556 |
+
"loss": 0.0251,
|
| 557 |
+
"step": 1925
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 3.7937743190661477,
|
| 561 |
+
"grad_norm": 1.0957626104354858,
|
| 562 |
+
"learning_rate": 6.777777777777779e-06,
|
| 563 |
+
"loss": 0.024,
|
| 564 |
+
"step": 1950
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 3.842412451361868,
|
| 568 |
+
"grad_norm": 1.3786637783050537,
|
| 569 |
+
"learning_rate": 6.7222222222222235e-06,
|
| 570 |
+
"loss": 0.0221,
|
| 571 |
+
"step": 1975
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 3.8910505836575875,
|
| 575 |
+
"grad_norm": 1.0472965240478516,
|
| 576 |
+
"learning_rate": 6.666666666666667e-06,
|
| 577 |
+
"loss": 0.0232,
|
| 578 |
+
"step": 2000
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 3.8910505836575875,
|
| 582 |
+
"eval_loss": 0.2916410565376282,
|
| 583 |
+
"eval_runtime": 2132.1371,
|
| 584 |
+
"eval_samples_per_second": 2.524,
|
| 585 |
+
"eval_steps_per_second": 0.158,
|
| 586 |
+
"eval_wer": 0.19905364030378941,
|
| 587 |
+
"step": 2000
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 3.9396887159533076,
|
| 591 |
+
"grad_norm": 1.0177409648895264,
|
| 592 |
+
"learning_rate": 6.6111111111111115e-06,
|
| 593 |
+
"loss": 0.0238,
|
| 594 |
+
"step": 2025
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 3.9883268482490273,
|
| 598 |
+
"grad_norm": 1.0486767292022705,
|
| 599 |
+
"learning_rate": 6.555555555555556e-06,
|
| 600 |
+
"loss": 0.0245,
|
| 601 |
+
"step": 2050
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 4.036964980544747,
|
| 605 |
+
"grad_norm": 0.7433611750602722,
|
| 606 |
+
"learning_rate": 6.5000000000000004e-06,
|
| 607 |
+
"loss": 0.0138,
|
| 608 |
+
"step": 2075
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 4.085603112840467,
|
| 612 |
+
"grad_norm": 0.4771675765514374,
|
| 613 |
+
"learning_rate": 6.444444444444445e-06,
|
| 614 |
+
"loss": 0.0109,
|
| 615 |
+
"step": 2100
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 4.134241245136187,
|
| 619 |
+
"grad_norm": 0.6232401132583618,
|
| 620 |
+
"learning_rate": 6.3888888888888885e-06,
|
| 621 |
+
"loss": 0.0109,
|
| 622 |
+
"step": 2125
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"epoch": 4.182879377431907,
|
| 626 |
+
"grad_norm": 0.6728478670120239,
|
| 627 |
+
"learning_rate": 6.333333333333333e-06,
|
| 628 |
+
"loss": 0.0103,
|
| 629 |
+
"step": 2150
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 4.2315175097276265,
|
| 633 |
+
"grad_norm": 1.6818970441818237,
|
| 634 |
+
"learning_rate": 6.277777777777778e-06,
|
| 635 |
+
"loss": 0.0114,
|
| 636 |
+
"step": 2175
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 4.280155642023346,
|
| 640 |
+
"grad_norm": 0.36247119307518005,
|
| 641 |
+
"learning_rate": 6.222222222222223e-06,
|
| 642 |
+
"loss": 0.011,
|
| 643 |
+
"step": 2200
|
| 644 |
+
},
|
| 645 |
+
{
|
| 646 |
+
"epoch": 4.328793774319066,
|
| 647 |
+
"grad_norm": 1.3438676595687866,
|
| 648 |
+
"learning_rate": 6.166666666666667e-06,
|
| 649 |
+
"loss": 0.0126,
|
| 650 |
+
"step": 2225
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 4.377431906614786,
|
| 654 |
+
"grad_norm": 0.8739299774169922,
|
| 655 |
+
"learning_rate": 6.111111111111112e-06,
|
| 656 |
+
"loss": 0.0104,
|
| 657 |
+
"step": 2250
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 4.426070038910506,
|
| 661 |
+
"grad_norm": 0.6817359328269958,
|
| 662 |
+
"learning_rate": 6.055555555555555e-06,
|
| 663 |
+
"loss": 0.0118,
|
| 664 |
+
"step": 2275
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 4.474708171206226,
|
| 668 |
+
"grad_norm": 0.5845357775688171,
|
| 669 |
+
"learning_rate": 6e-06,
|
| 670 |
+
"loss": 0.0125,
|
| 671 |
+
"step": 2300
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"epoch": 4.523346303501945,
|
| 675 |
+
"grad_norm": 1.067460298538208,
|
| 676 |
+
"learning_rate": 5.944444444444445e-06,
|
| 677 |
+
"loss": 0.0099,
|
| 678 |
+
"step": 2325
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"epoch": 4.571984435797665,
|
| 682 |
+
"grad_norm": 0.7725332379341125,
|
| 683 |
+
"learning_rate": 5.88888888888889e-06,
|
| 684 |
+
"loss": 0.0123,
|
| 685 |
+
"step": 2350
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 4.620622568093385,
|
| 689 |
+
"grad_norm": 0.8168444633483887,
|
| 690 |
+
"learning_rate": 5.833333333333334e-06,
|
| 691 |
+
"loss": 0.01,
|
| 692 |
+
"step": 2375
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 4.669260700389105,
|
| 696 |
+
"grad_norm": 0.7538411617279053,
|
| 697 |
+
"learning_rate": 5.777777777777778e-06,
|
| 698 |
+
"loss": 0.0098,
|
| 699 |
+
"step": 2400
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 4.717898832684825,
|
| 703 |
+
"grad_norm": 0.3141120970249176,
|
| 704 |
+
"learning_rate": 5.722222222222222e-06,
|
| 705 |
+
"loss": 0.0107,
|
| 706 |
+
"step": 2425
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 4.766536964980545,
|
| 710 |
+
"grad_norm": 0.887184202671051,
|
| 711 |
+
"learning_rate": 5.666666666666667e-06,
|
| 712 |
+
"loss": 0.0095,
|
| 713 |
+
"step": 2450
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 4.815175097276264,
|
| 717 |
+
"grad_norm": 0.5846179127693176,
|
| 718 |
+
"learning_rate": 5.611111111111112e-06,
|
| 719 |
+
"loss": 0.0099,
|
| 720 |
+
"step": 2475
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 4.863813229571985,
|
| 724 |
+
"grad_norm": 0.3938518464565277,
|
| 725 |
+
"learning_rate": 5.555555555555557e-06,
|
| 726 |
+
"loss": 0.0091,
|
| 727 |
+
"step": 2500
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 4.912451361867705,
|
| 731 |
+
"grad_norm": 0.9777388572692871,
|
| 732 |
+
"learning_rate": 5.500000000000001e-06,
|
| 733 |
+
"loss": 0.0093,
|
| 734 |
+
"step": 2525
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 4.961089494163424,
|
| 738 |
+
"grad_norm": 0.7862864136695862,
|
| 739 |
+
"learning_rate": 5.444444444444445e-06,
|
| 740 |
+
"loss": 0.0111,
|
| 741 |
+
"step": 2550
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 5.009727626459144,
|
| 745 |
+
"grad_norm": 0.7504919171333313,
|
| 746 |
+
"learning_rate": 5.388888888888889e-06,
|
| 747 |
+
"loss": 0.0094,
|
| 748 |
+
"step": 2575
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 5.058365758754864,
|
| 752 |
+
"grad_norm": 0.2351156324148178,
|
| 753 |
+
"learning_rate": 5.333333333333334e-06,
|
| 754 |
+
"loss": 0.0049,
|
| 755 |
+
"step": 2600
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 5.107003891050583,
|
| 759 |
+
"grad_norm": 0.13562794029712677,
|
| 760 |
+
"learning_rate": 5.2777777777777785e-06,
|
| 761 |
+
"loss": 0.0043,
|
| 762 |
+
"step": 2625
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 5.155642023346304,
|
| 766 |
+
"grad_norm": 0.33899495005607605,
|
| 767 |
+
"learning_rate": 5.2222222222222226e-06,
|
| 768 |
+
"loss": 0.0048,
|
| 769 |
+
"step": 2650
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 5.2042801556420235,
|
| 773 |
+
"grad_norm": 0.9408302307128906,
|
| 774 |
+
"learning_rate": 5.1666666666666675e-06,
|
| 775 |
+
"loss": 0.0051,
|
| 776 |
+
"step": 2675
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 5.252918287937743,
|
| 780 |
+
"grad_norm": 0.35573306679725647,
|
| 781 |
+
"learning_rate": 5.1111111111111115e-06,
|
| 782 |
+
"loss": 0.0053,
|
| 783 |
+
"step": 2700
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 5.301556420233463,
|
| 787 |
+
"grad_norm": 0.7676318883895874,
|
| 788 |
+
"learning_rate": 5.0555555555555555e-06,
|
| 789 |
+
"loss": 0.0042,
|
| 790 |
+
"step": 2725
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 5.3501945525291825,
|
| 794 |
+
"grad_norm": 0.6776612401008606,
|
| 795 |
+
"learning_rate": 5e-06,
|
| 796 |
+
"loss": 0.004,
|
| 797 |
+
"step": 2750
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 5.398832684824903,
|
| 801 |
+
"grad_norm": 1.0338044166564941,
|
| 802 |
+
"learning_rate": 4.944444444444445e-06,
|
| 803 |
+
"loss": 0.0047,
|
| 804 |
+
"step": 2775
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 5.447470817120623,
|
| 808 |
+
"grad_norm": 0.2744717001914978,
|
| 809 |
+
"learning_rate": 4.888888888888889e-06,
|
| 810 |
+
"loss": 0.0068,
|
| 811 |
+
"step": 2800
|
| 812 |
+
},
|
| 813 |
+
{
|
| 814 |
+
"epoch": 5.496108949416342,
|
| 815 |
+
"grad_norm": 0.7075155973434448,
|
| 816 |
+
"learning_rate": 4.833333333333333e-06,
|
| 817 |
+
"loss": 0.0061,
|
| 818 |
+
"step": 2825
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 5.544747081712062,
|
| 822 |
+
"grad_norm": 0.1820390671491623,
|
| 823 |
+
"learning_rate": 4.777777777777778e-06,
|
| 824 |
+
"loss": 0.0052,
|
| 825 |
+
"step": 2850
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 5.593385214007782,
|
| 829 |
+
"grad_norm": 0.33624935150146484,
|
| 830 |
+
"learning_rate": 4.722222222222222e-06,
|
| 831 |
+
"loss": 0.0038,
|
| 832 |
+
"step": 2875
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 5.642023346303502,
|
| 836 |
+
"grad_norm": 0.1461501270532608,
|
| 837 |
+
"learning_rate": 4.666666666666667e-06,
|
| 838 |
+
"loss": 0.0046,
|
| 839 |
+
"step": 2900
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 5.690661478599222,
|
| 843 |
+
"grad_norm": 0.33367106318473816,
|
| 844 |
+
"learning_rate": 4.611111111111112e-06,
|
| 845 |
+
"loss": 0.0042,
|
| 846 |
+
"step": 2925
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 5.739299610894942,
|
| 850 |
+
"grad_norm": 0.32830071449279785,
|
| 851 |
+
"learning_rate": 4.555555555555556e-06,
|
| 852 |
+
"loss": 0.0043,
|
| 853 |
+
"step": 2950
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 5.787937743190661,
|
| 857 |
+
"grad_norm": 1.2332854270935059,
|
| 858 |
+
"learning_rate": 4.5e-06,
|
| 859 |
+
"loss": 0.0039,
|
| 860 |
+
"step": 2975
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 5.836575875486381,
|
| 864 |
+
"grad_norm": 0.6161640286445618,
|
| 865 |
+
"learning_rate": 4.444444444444444e-06,
|
| 866 |
+
"loss": 0.0046,
|
| 867 |
+
"step": 3000
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 5.836575875486381,
|
| 871 |
+
"eval_loss": 0.3218872845172882,
|
| 872 |
+
"eval_runtime": 2114.549,
|
| 873 |
+
"eval_samples_per_second": 2.545,
|
| 874 |
+
"eval_steps_per_second": 0.159,
|
| 875 |
+
"eval_wer": 0.18784047079406735,
|
| 876 |
+
"step": 3000
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 5.885214007782102,
|
| 880 |
+
"grad_norm": 2.608947992324829,
|
| 881 |
+
"learning_rate": 4.388888888888889e-06,
|
| 882 |
+
"loss": 0.0061,
|
| 883 |
+
"step": 3025
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 5.933852140077821,
|
| 887 |
+
"grad_norm": 0.25549063086509705,
|
| 888 |
+
"learning_rate": 4.333333333333334e-06,
|
| 889 |
+
"loss": 0.0044,
|
| 890 |
+
"step": 3050
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 5.982490272373541,
|
| 894 |
+
"grad_norm": 1.4720717668533325,
|
| 895 |
+
"learning_rate": 4.277777777777778e-06,
|
| 896 |
+
"loss": 0.0051,
|
| 897 |
+
"step": 3075
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 6.031128404669261,
|
| 901 |
+
"grad_norm": 0.7722542881965637,
|
| 902 |
+
"learning_rate": 4.222222222222223e-06,
|
| 903 |
+
"loss": 0.0037,
|
| 904 |
+
"step": 3100
|
| 905 |
+
},
|
| 906 |
+
{
|
| 907 |
+
"epoch": 6.07976653696498,
|
| 908 |
+
"grad_norm": 0.35967299342155457,
|
| 909 |
+
"learning_rate": 4.166666666666667e-06,
|
| 910 |
+
"loss": 0.0029,
|
| 911 |
+
"step": 3125
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 6.1284046692607,
|
| 915 |
+
"grad_norm": 0.7472084164619446,
|
| 916 |
+
"learning_rate": 4.111111111111111e-06,
|
| 917 |
+
"loss": 0.0029,
|
| 918 |
+
"step": 3150
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"epoch": 6.1770428015564205,
|
| 922 |
+
"grad_norm": 0.08338561654090881,
|
| 923 |
+
"learning_rate": 4.055555555555556e-06,
|
| 924 |
+
"loss": 0.0023,
|
| 925 |
+
"step": 3175
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 6.22568093385214,
|
| 929 |
+
"grad_norm": 0.49297797679901123,
|
| 930 |
+
"learning_rate": 4.000000000000001e-06,
|
| 931 |
+
"loss": 0.0022,
|
| 932 |
+
"step": 3200
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 6.27431906614786,
|
| 936 |
+
"grad_norm": 0.08119112253189087,
|
| 937 |
+
"learning_rate": 3.944444444444445e-06,
|
| 938 |
+
"loss": 0.0019,
|
| 939 |
+
"step": 3225
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 6.3229571984435795,
|
| 943 |
+
"grad_norm": 0.13106492161750793,
|
| 944 |
+
"learning_rate": 3.88888888888889e-06,
|
| 945 |
+
"loss": 0.0027,
|
| 946 |
+
"step": 3250
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 6.3715953307393,
|
| 950 |
+
"grad_norm": 0.18573686480522156,
|
| 951 |
+
"learning_rate": 3.833333333333334e-06,
|
| 952 |
+
"loss": 0.003,
|
| 953 |
+
"step": 3275
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 6.42023346303502,
|
| 957 |
+
"grad_norm": 0.09228133410215378,
|
| 958 |
+
"learning_rate": 3.777777777777778e-06,
|
| 959 |
+
"loss": 0.004,
|
| 960 |
+
"step": 3300
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 6.468871595330739,
|
| 964 |
+
"grad_norm": 0.8169698715209961,
|
| 965 |
+
"learning_rate": 3.7222222222222225e-06,
|
| 966 |
+
"loss": 0.0029,
|
| 967 |
+
"step": 3325
|
| 968 |
+
},
|
| 969 |
+
{
|
| 970 |
+
"epoch": 6.517509727626459,
|
| 971 |
+
"grad_norm": 0.23130229115486145,
|
| 972 |
+
"learning_rate": 3.6666666666666666e-06,
|
| 973 |
+
"loss": 0.0026,
|
| 974 |
+
"step": 3350
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 6.566147859922179,
|
| 978 |
+
"grad_norm": 0.10798731446266174,
|
| 979 |
+
"learning_rate": 3.6111111111111115e-06,
|
| 980 |
+
"loss": 0.0036,
|
| 981 |
+
"step": 3375
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 6.614785992217898,
|
| 985 |
+
"grad_norm": 0.09145894646644592,
|
| 986 |
+
"learning_rate": 3.555555555555556e-06,
|
| 987 |
+
"loss": 0.0024,
|
| 988 |
+
"step": 3400
|
| 989 |
+
},
|
| 990 |
+
{
|
| 991 |
+
"epoch": 6.663424124513619,
|
| 992 |
+
"grad_norm": 0.3099443018436432,
|
| 993 |
+
"learning_rate": 3.5e-06,
|
| 994 |
+
"loss": 0.0018,
|
| 995 |
+
"step": 3425
|
| 996 |
+
},
|
| 997 |
+
{
|
| 998 |
+
"epoch": 6.712062256809339,
|
| 999 |
+
"grad_norm": 0.13762035965919495,
|
| 1000 |
+
"learning_rate": 3.444444444444445e-06,
|
| 1001 |
+
"loss": 0.0021,
|
| 1002 |
+
"step": 3450
|
| 1003 |
+
},
|
| 1004 |
+
{
|
| 1005 |
+
"epoch": 6.760700389105058,
|
| 1006 |
+
"grad_norm": 0.11253529787063599,
|
| 1007 |
+
"learning_rate": 3.3888888888888893e-06,
|
| 1008 |
+
"loss": 0.0012,
|
| 1009 |
+
"step": 3475
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"epoch": 6.809338521400778,
|
| 1013 |
+
"grad_norm": 0.054800793528556824,
|
| 1014 |
+
"learning_rate": 3.3333333333333333e-06,
|
| 1015 |
+
"loss": 0.0019,
|
| 1016 |
+
"step": 3500
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 6.857976653696498,
|
| 1020 |
+
"grad_norm": 0.06803842633962631,
|
| 1021 |
+
"learning_rate": 3.277777777777778e-06,
|
| 1022 |
+
"loss": 0.0024,
|
| 1023 |
+
"step": 3525
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 6.906614785992218,
|
| 1027 |
+
"grad_norm": 0.08114957064390182,
|
| 1028 |
+
"learning_rate": 3.2222222222222227e-06,
|
| 1029 |
+
"loss": 0.0019,
|
| 1030 |
+
"step": 3550
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"epoch": 6.955252918287938,
|
| 1034 |
+
"grad_norm": 0.06155654788017273,
|
| 1035 |
+
"learning_rate": 3.1666666666666667e-06,
|
| 1036 |
+
"loss": 0.0017,
|
| 1037 |
+
"step": 3575
|
| 1038 |
+
},
|
| 1039 |
+
{
|
| 1040 |
+
"epoch": 7.003891050583658,
|
| 1041 |
+
"grad_norm": 0.09375619143247604,
|
| 1042 |
+
"learning_rate": 3.1111111111111116e-06,
|
| 1043 |
+
"loss": 0.0015,
|
| 1044 |
+
"step": 3600
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"epoch": 7.052529182879377,
|
| 1048 |
+
"grad_norm": 0.04578198865056038,
|
| 1049 |
+
"learning_rate": 3.055555555555556e-06,
|
| 1050 |
+
"loss": 0.0011,
|
| 1051 |
+
"step": 3625
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 7.101167315175097,
|
| 1055 |
+
"grad_norm": 0.036920215934515,
|
| 1056 |
+
"learning_rate": 3e-06,
|
| 1057 |
+
"loss": 0.0012,
|
| 1058 |
+
"step": 3650
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 7.1498054474708175,
|
| 1062 |
+
"grad_norm": 0.07471055537462234,
|
| 1063 |
+
"learning_rate": 2.944444444444445e-06,
|
| 1064 |
+
"loss": 0.0009,
|
| 1065 |
+
"step": 3675
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 7.198443579766537,
|
| 1069 |
+
"grad_norm": 0.05370425805449486,
|
| 1070 |
+
"learning_rate": 2.888888888888889e-06,
|
| 1071 |
+
"loss": 0.0009,
|
| 1072 |
+
"step": 3700
|
| 1073 |
+
},
|
| 1074 |
+
{
|
| 1075 |
+
"epoch": 7.247081712062257,
|
| 1076 |
+
"grad_norm": 0.04795600846409798,
|
| 1077 |
+
"learning_rate": 2.8333333333333335e-06,
|
| 1078 |
+
"loss": 0.001,
|
| 1079 |
+
"step": 3725
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 7.2957198443579765,
|
| 1083 |
+
"grad_norm": 0.030901705846190453,
|
| 1084 |
+
"learning_rate": 2.7777777777777783e-06,
|
| 1085 |
+
"loss": 0.0007,
|
| 1086 |
+
"step": 3750
|
| 1087 |
+
},
|
| 1088 |
+
{
|
| 1089 |
+
"epoch": 7.344357976653696,
|
| 1090 |
+
"grad_norm": 0.05389130115509033,
|
| 1091 |
+
"learning_rate": 2.7222222222222224e-06,
|
| 1092 |
+
"loss": 0.0012,
|
| 1093 |
+
"step": 3775
|
| 1094 |
+
},
|
| 1095 |
+
{
|
| 1096 |
+
"epoch": 7.392996108949417,
|
| 1097 |
+
"grad_norm": 0.05646834522485733,
|
| 1098 |
+
"learning_rate": 2.666666666666667e-06,
|
| 1099 |
+
"loss": 0.0009,
|
| 1100 |
+
"step": 3800
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 7.441634241245136,
|
| 1104 |
+
"grad_norm": 0.04085018113255501,
|
| 1105 |
+
"learning_rate": 2.6111111111111113e-06,
|
| 1106 |
+
"loss": 0.001,
|
| 1107 |
+
"step": 3825
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 7.490272373540856,
|
| 1111 |
+
"grad_norm": 0.024638062343001366,
|
| 1112 |
+
"learning_rate": 2.5555555555555557e-06,
|
| 1113 |
+
"loss": 0.0009,
|
| 1114 |
+
"step": 3850
|
| 1115 |
+
},
|
| 1116 |
+
{
|
| 1117 |
+
"epoch": 7.538910505836576,
|
| 1118 |
+
"grad_norm": 0.03571788966655731,
|
| 1119 |
+
"learning_rate": 2.5e-06,
|
| 1120 |
+
"loss": 0.0011,
|
| 1121 |
+
"step": 3875
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 7.587548638132295,
|
| 1125 |
+
"grad_norm": 0.05909154936671257,
|
| 1126 |
+
"learning_rate": 2.4444444444444447e-06,
|
| 1127 |
+
"loss": 0.0007,
|
| 1128 |
+
"step": 3900
|
| 1129 |
+
},
|
| 1130 |
+
{
|
| 1131 |
+
"epoch": 7.636186770428015,
|
| 1132 |
+
"grad_norm": 0.028353404253721237,
|
| 1133 |
+
"learning_rate": 2.388888888888889e-06,
|
| 1134 |
+
"loss": 0.001,
|
| 1135 |
+
"step": 3925
|
| 1136 |
+
},
|
| 1137 |
+
{
|
| 1138 |
+
"epoch": 7.684824902723736,
|
| 1139 |
+
"grad_norm": 0.16029401123523712,
|
| 1140 |
+
"learning_rate": 2.3333333333333336e-06,
|
| 1141 |
+
"loss": 0.0009,
|
| 1142 |
+
"step": 3950
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 7.733463035019455,
|
| 1146 |
+
"grad_norm": 0.04438905417919159,
|
| 1147 |
+
"learning_rate": 2.277777777777778e-06,
|
| 1148 |
+
"loss": 0.0008,
|
| 1149 |
+
"step": 3975
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 7.782101167315175,
|
| 1153 |
+
"grad_norm": 0.035433579236269,
|
| 1154 |
+
"learning_rate": 2.222222222222222e-06,
|
| 1155 |
+
"loss": 0.0009,
|
| 1156 |
+
"step": 4000
|
| 1157 |
+
},
|
| 1158 |
+
{
|
| 1159 |
+
"epoch": 7.782101167315175,
|
| 1160 |
+
"eval_loss": 0.3454054296016693,
|
| 1161 |
+
"eval_runtime": 2148.9325,
|
| 1162 |
+
"eval_samples_per_second": 2.504,
|
| 1163 |
+
"eval_steps_per_second": 0.157,
|
| 1164 |
+
"eval_wer": 0.1831881983379061,
|
| 1165 |
+
"step": 4000
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 7.830739299610895,
|
| 1169 |
+
"grad_norm": 0.028996312990784645,
|
| 1170 |
+
"learning_rate": 2.166666666666667e-06,
|
| 1171 |
+
"loss": 0.0008,
|
| 1172 |
+
"step": 4025
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 7.879377431906615,
|
| 1176 |
+
"grad_norm": 0.038502488285303116,
|
| 1177 |
+
"learning_rate": 2.1111111111111114e-06,
|
| 1178 |
+
"loss": 0.0014,
|
| 1179 |
+
"step": 4050
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 7.928015564202335,
|
| 1183 |
+
"grad_norm": 0.06861083209514618,
|
| 1184 |
+
"learning_rate": 2.0555555555555555e-06,
|
| 1185 |
+
"loss": 0.0009,
|
| 1186 |
+
"step": 4075
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 7.976653696498055,
|
| 1190 |
+
"grad_norm": 0.07425787299871445,
|
| 1191 |
+
"learning_rate": 2.0000000000000003e-06,
|
| 1192 |
+
"loss": 0.0011,
|
| 1193 |
+
"step": 4100
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 8.025291828793774,
|
| 1197 |
+
"grad_norm": 0.025021173059940338,
|
| 1198 |
+
"learning_rate": 1.944444444444445e-06,
|
| 1199 |
+
"loss": 0.0006,
|
| 1200 |
+
"step": 4125
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 8.073929961089494,
|
| 1204 |
+
"grad_norm": 0.031610701233148575,
|
| 1205 |
+
"learning_rate": 1.888888888888889e-06,
|
| 1206 |
+
"loss": 0.0006,
|
| 1207 |
+
"step": 4150
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 8.122568093385214,
|
| 1211 |
+
"grad_norm": 0.02210886962711811,
|
| 1212 |
+
"learning_rate": 1.8333333333333333e-06,
|
| 1213 |
+
"loss": 0.0008,
|
| 1214 |
+
"step": 4175
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 8.171206225680933,
|
| 1218 |
+
"grad_norm": 0.03183664008975029,
|
| 1219 |
+
"learning_rate": 1.777777777777778e-06,
|
| 1220 |
+
"loss": 0.0007,
|
| 1221 |
+
"step": 4200
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 8.219844357976653,
|
| 1225 |
+
"grad_norm": 0.024653365835547447,
|
| 1226 |
+
"learning_rate": 1.7222222222222224e-06,
|
| 1227 |
+
"loss": 0.0005,
|
| 1228 |
+
"step": 4225
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 8.268482490272374,
|
| 1232 |
+
"grad_norm": 0.02229795977473259,
|
| 1233 |
+
"learning_rate": 1.6666666666666667e-06,
|
| 1234 |
+
"loss": 0.0005,
|
| 1235 |
+
"step": 4250
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 8.317120622568094,
|
| 1239 |
+
"grad_norm": 0.025678085163235664,
|
| 1240 |
+
"learning_rate": 1.6111111111111113e-06,
|
| 1241 |
+
"loss": 0.0005,
|
| 1242 |
+
"step": 4275
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 8.365758754863814,
|
| 1246 |
+
"grad_norm": 0.025166384875774384,
|
| 1247 |
+
"learning_rate": 1.5555555555555558e-06,
|
| 1248 |
+
"loss": 0.0005,
|
| 1249 |
+
"step": 4300
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 8.414396887159533,
|
| 1253 |
+
"grad_norm": 0.017260603606700897,
|
| 1254 |
+
"learning_rate": 1.5e-06,
|
| 1255 |
+
"loss": 0.0005,
|
| 1256 |
+
"step": 4325
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 8.463035019455253,
|
| 1260 |
+
"grad_norm": 0.016240952536463737,
|
| 1261 |
+
"learning_rate": 1.4444444444444445e-06,
|
| 1262 |
+
"loss": 0.0005,
|
| 1263 |
+
"step": 4350
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 8.511673151750973,
|
| 1267 |
+
"grad_norm": 0.021515797823667526,
|
| 1268 |
+
"learning_rate": 1.3888888888888892e-06,
|
| 1269 |
+
"loss": 0.0006,
|
| 1270 |
+
"step": 4375
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 8.560311284046692,
|
| 1274 |
+
"grad_norm": 0.023731861263513565,
|
| 1275 |
+
"learning_rate": 1.3333333333333334e-06,
|
| 1276 |
+
"loss": 0.0006,
|
| 1277 |
+
"step": 4400
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 8.608949416342412,
|
| 1281 |
+
"grad_norm": 0.027586709707975388,
|
| 1282 |
+
"learning_rate": 1.2777777777777779e-06,
|
| 1283 |
+
"loss": 0.0006,
|
| 1284 |
+
"step": 4425
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 8.657587548638132,
|
| 1288 |
+
"grad_norm": 0.04710078611969948,
|
| 1289 |
+
"learning_rate": 1.2222222222222223e-06,
|
| 1290 |
+
"loss": 0.0005,
|
| 1291 |
+
"step": 4450
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 8.706225680933851,
|
| 1295 |
+
"grad_norm": 0.030895188450813293,
|
| 1296 |
+
"learning_rate": 1.1666666666666668e-06,
|
| 1297 |
+
"loss": 0.0005,
|
| 1298 |
+
"step": 4475
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 8.754863813229573,
|
| 1302 |
+
"grad_norm": 0.04257024824619293,
|
| 1303 |
+
"learning_rate": 1.111111111111111e-06,
|
| 1304 |
+
"loss": 0.0006,
|
| 1305 |
+
"step": 4500
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 8.803501945525293,
|
| 1309 |
+
"grad_norm": 0.022679895162582397,
|
| 1310 |
+
"learning_rate": 1.0555555555555557e-06,
|
| 1311 |
+
"loss": 0.0006,
|
| 1312 |
+
"step": 4525
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 8.852140077821012,
|
| 1316 |
+
"grad_norm": 0.028052836656570435,
|
| 1317 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 1318 |
+
"loss": 0.0006,
|
| 1319 |
+
"step": 4550
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 8.900778210116732,
|
| 1323 |
+
"grad_norm": 0.022373031824827194,
|
| 1324 |
+
"learning_rate": 9.444444444444445e-07,
|
| 1325 |
+
"loss": 0.0005,
|
| 1326 |
+
"step": 4575
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 8.949416342412452,
|
| 1330 |
+
"grad_norm": 0.02304161712527275,
|
| 1331 |
+
"learning_rate": 8.88888888888889e-07,
|
| 1332 |
+
"loss": 0.0005,
|
| 1333 |
+
"step": 4600
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 8.998054474708171,
|
| 1337 |
+
"grad_norm": 0.020944086834788322,
|
| 1338 |
+
"learning_rate": 8.333333333333333e-07,
|
| 1339 |
+
"loss": 0.0005,
|
| 1340 |
+
"step": 4625
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 9.04669260700389,
|
| 1344 |
+
"grad_norm": 0.018772481009364128,
|
| 1345 |
+
"learning_rate": 7.777777777777779e-07,
|
| 1346 |
+
"loss": 0.0005,
|
| 1347 |
+
"step": 4650
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 9.09533073929961,
|
| 1351 |
+
"grad_norm": 0.019685110077261925,
|
| 1352 |
+
"learning_rate": 7.222222222222222e-07,
|
| 1353 |
+
"loss": 0.0005,
|
| 1354 |
+
"step": 4675
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 9.14396887159533,
|
| 1358 |
+
"grad_norm": 0.017540233209729195,
|
| 1359 |
+
"learning_rate": 6.666666666666667e-07,
|
| 1360 |
+
"loss": 0.0004,
|
| 1361 |
+
"step": 4700
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 9.19260700389105,
|
| 1365 |
+
"grad_norm": 0.022134965285658836,
|
| 1366 |
+
"learning_rate": 6.111111111111112e-07,
|
| 1367 |
+
"loss": 0.0004,
|
| 1368 |
+
"step": 4725
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 9.24124513618677,
|
| 1372 |
+
"grad_norm": 0.01725279912352562,
|
| 1373 |
+
"learning_rate": 5.555555555555555e-07,
|
| 1374 |
+
"loss": 0.0005,
|
| 1375 |
+
"step": 4750
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 9.289883268482491,
|
| 1379 |
+
"grad_norm": 0.01975845918059349,
|
| 1380 |
+
"learning_rate": 5.000000000000001e-07,
|
| 1381 |
+
"loss": 0.0004,
|
| 1382 |
+
"step": 4775
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 9.33852140077821,
|
| 1386 |
+
"grad_norm": 0.021399999037384987,
|
| 1387 |
+
"learning_rate": 4.444444444444445e-07,
|
| 1388 |
+
"loss": 0.0005,
|
| 1389 |
+
"step": 4800
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 9.38715953307393,
|
| 1393 |
+
"grad_norm": 0.020681940019130707,
|
| 1394 |
+
"learning_rate": 3.8888888888888895e-07,
|
| 1395 |
+
"loss": 0.0004,
|
| 1396 |
+
"step": 4825
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 9.43579766536965,
|
| 1400 |
+
"grad_norm": 0.020026126876473427,
|
| 1401 |
+
"learning_rate": 3.3333333333333335e-07,
|
| 1402 |
+
"loss": 0.0005,
|
| 1403 |
+
"step": 4850
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 9.48443579766537,
|
| 1407 |
+
"grad_norm": 0.030826667323708534,
|
| 1408 |
+
"learning_rate": 2.7777777777777776e-07,
|
| 1409 |
+
"loss": 0.0004,
|
| 1410 |
+
"step": 4875
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 9.53307392996109,
|
| 1414 |
+
"grad_norm": 0.017975399270653725,
|
| 1415 |
+
"learning_rate": 2.2222222222222224e-07,
|
| 1416 |
+
"loss": 0.0004,
|
| 1417 |
+
"step": 4900
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 9.581712062256809,
|
| 1421 |
+
"grad_norm": 0.014251296408474445,
|
| 1422 |
+
"learning_rate": 1.6666666666666668e-07,
|
| 1423 |
+
"loss": 0.0005,
|
| 1424 |
+
"step": 4925
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 9.630350194552529,
|
| 1428 |
+
"grad_norm": 0.018128497526049614,
|
| 1429 |
+
"learning_rate": 1.1111111111111112e-07,
|
| 1430 |
+
"loss": 0.0005,
|
| 1431 |
+
"step": 4950
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 9.678988326848248,
|
| 1435 |
+
"grad_norm": 0.018901441246271133,
|
| 1436 |
+
"learning_rate": 5.555555555555556e-08,
|
| 1437 |
+
"loss": 0.0005,
|
| 1438 |
+
"step": 4975
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 9.727626459143968,
|
| 1442 |
+
"grad_norm": 0.01760680228471756,
|
| 1443 |
+
"learning_rate": 0.0,
|
| 1444 |
+
"loss": 0.0004,
|
| 1445 |
+
"step": 5000
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 9.727626459143968,
|
| 1449 |
+
"eval_loss": 0.36376532912254333,
|
| 1450 |
+
"eval_runtime": 2157.1617,
|
| 1451 |
+
"eval_samples_per_second": 2.494,
|
| 1452 |
+
"eval_steps_per_second": 0.156,
|
| 1453 |
+
"eval_wer": 0.18173684838363355,
|
| 1454 |
+
"step": 5000
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 9.727626459143968,
|
| 1458 |
+
"step": 5000,
|
| 1459 |
+
"total_flos": 5.4332453168676864e+20,
|
| 1460 |
+
"train_loss": 0.06844911024216563,
|
| 1461 |
+
"train_runtime": 58020.7767,
|
| 1462 |
+
"train_samples_per_second": 2.758,
|
| 1463 |
+
"train_steps_per_second": 0.086
|
| 1464 |
+
}
|
| 1465 |
+
],
|
| 1466 |
+
"logging_steps": 25,
|
| 1467 |
+
"max_steps": 5000,
|
| 1468 |
+
"num_input_tokens_seen": 0,
|
| 1469 |
+
"num_train_epochs": 10,
|
| 1470 |
+
"save_steps": 1000,
|
| 1471 |
+
"stateful_callbacks": {
|
| 1472 |
+
"TrainerControl": {
|
| 1473 |
+
"args": {
|
| 1474 |
+
"should_epoch_stop": false,
|
| 1475 |
+
"should_evaluate": false,
|
| 1476 |
+
"should_log": false,
|
| 1477 |
+
"should_save": true,
|
| 1478 |
+
"should_training_stop": true
|
| 1479 |
+
},
|
| 1480 |
+
"attributes": {}
|
| 1481 |
+
}
|
| 1482 |
+
},
|
| 1483 |
+
"total_flos": 5.4332453168676864e+20,
|
| 1484 |
+
"train_batch_size": 16,
|
| 1485 |
+
"trial_name": null,
|
| 1486 |
+
"trial_params": null
|
| 1487 |
+
}
|