End of training
Browse files- .gitattributes +1 -0
- README.md +9 -6
- all_results.json +11 -5
- eval_results.json +6 -6
- train_results.json +8 -0
- trainer_state.json +2724 -0
- wandb/run-20241007_131849-0rbzerob/files/output.log +0 -0
- wandb/run-20241007_131849-0rbzerob/run-0rbzerob.wandb +0 -0
.gitattributes
CHANGED
|
@@ -35,3 +35,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
wandb/run-20241005_141414-821qpm7o/files/output.log filter=lfs diff=lfs merge=lfs -text
|
| 37 |
wandb/run-20241005_141414-821qpm7o/run-821qpm7o.wandb filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
wandb/run-20241005_141414-821qpm7o/files/output.log filter=lfs diff=lfs merge=lfs -text
|
| 37 |
wandb/run-20241005_141414-821qpm7o/run-821qpm7o.wandb filter=lfs diff=lfs merge=lfs -text
|
| 38 |
+
wandb/run-20241007_131849-0rbzerob/run-0rbzerob.wandb filter=lfs diff=lfs merge=lfs -text
|
README.md
CHANGED
|
@@ -1,22 +1,25 @@
|
|
| 1 |
---
|
| 2 |
library_name: transformers
|
|
|
|
|
|
|
| 3 |
license: apache-2.0
|
| 4 |
base_model: openai/whisper-large-v3
|
| 5 |
tags:
|
|
|
|
| 6 |
- generated_from_trainer
|
| 7 |
datasets:
|
| 8 |
-
- common_voice_17_0
|
| 9 |
metrics:
|
| 10 |
- wer
|
| 11 |
model-index:
|
| 12 |
-
- name:
|
| 13 |
results:
|
| 14 |
- task:
|
| 15 |
name: Automatic Speech Recognition
|
| 16 |
type: automatic-speech-recognition
|
| 17 |
dataset:
|
| 18 |
-
name: common_voice_17_0
|
| 19 |
-
type: common_voice_17_0
|
| 20 |
config: eu
|
| 21 |
split: test
|
| 22 |
args: eu
|
|
@@ -29,9 +32,9 @@ model-index:
|
|
| 29 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 30 |
should probably proofread and complete it, then remove this comment. -->
|
| 31 |
|
| 32 |
-
#
|
| 33 |
|
| 34 |
-
This model is a fine-tuned version of [openai/whisper-large-v3](https://huggingface.co/openai/whisper-large-v3) on the common_voice_17_0 dataset.
|
| 35 |
It achieves the following results on the evaluation set:
|
| 36 |
- Loss: 0.1259
|
| 37 |
- Wer: 7.2154
|
|
|
|
| 1 |
---
|
| 2 |
library_name: transformers
|
| 3 |
+
language:
|
| 4 |
+
- eu
|
| 5 |
license: apache-2.0
|
| 6 |
base_model: openai/whisper-large-v3
|
| 7 |
tags:
|
| 8 |
+
- whisper-event
|
| 9 |
- generated_from_trainer
|
| 10 |
datasets:
|
| 11 |
+
- mozilla-foundation/common_voice_17_0
|
| 12 |
metrics:
|
| 13 |
- wer
|
| 14 |
model-index:
|
| 15 |
+
- name: Whisper Large Basque
|
| 16 |
results:
|
| 17 |
- task:
|
| 18 |
name: Automatic Speech Recognition
|
| 19 |
type: automatic-speech-recognition
|
| 20 |
dataset:
|
| 21 |
+
name: mozilla-foundation/common_voice_17_0 eu
|
| 22 |
+
type: mozilla-foundation/common_voice_17_0
|
| 23 |
config: eu
|
| 24 |
split: test
|
| 25 |
args: eu
|
|
|
|
| 32 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 33 |
should probably proofread and complete it, then remove this comment. -->
|
| 34 |
|
| 35 |
+
# Whisper Large Basque
|
| 36 |
|
| 37 |
+
This model is a fine-tuned version of [openai/whisper-large-v3](https://huggingface.co/openai/whisper-large-v3) on the mozilla-foundation/common_voice_17_0 eu dataset.
|
| 38 |
It achieves the following results on the evaluation set:
|
| 39 |
- Loss: 0.1259
|
| 40 |
- Wer: 7.2154
|
all_results.json
CHANGED
|
@@ -1,8 +1,14 @@
|
|
| 1 |
{
|
| 2 |
-
"
|
|
|
|
| 3 |
"eval_model_preparation_time": 0.0102,
|
| 4 |
-
"eval_runtime":
|
| 5 |
-
"eval_samples_per_second": 3.
|
| 6 |
-
"eval_steps_per_second": 0.
|
| 7 |
-
"eval_wer":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 5.048,
|
| 3 |
+
"eval_loss": 0.12588092684745789,
|
| 4 |
"eval_model_preparation_time": 0.0102,
|
| 5 |
+
"eval_runtime": 4097.4891,
|
| 6 |
+
"eval_samples_per_second": 3.326,
|
| 7 |
+
"eval_steps_per_second": 0.416,
|
| 8 |
+
"eval_wer": 7.215361500971087,
|
| 9 |
+
"total_flos": 4.891718061785088e+20,
|
| 10 |
+
"train_loss": 0.0,
|
| 11 |
+
"train_runtime": 289.8068,
|
| 12 |
+
"train_samples_per_second": 552.092,
|
| 13 |
+
"train_steps_per_second": 34.506
|
| 14 |
}
|
eval_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"
|
| 3 |
-
"
|
| 4 |
-
"eval_runtime":
|
| 5 |
-
"eval_samples_per_second": 3.
|
| 6 |
-
"eval_steps_per_second": 0.
|
| 7 |
-
"eval_wer":
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 5.048,
|
| 3 |
+
"eval_loss": 0.12588092684745789,
|
| 4 |
+
"eval_runtime": 4097.4891,
|
| 5 |
+
"eval_samples_per_second": 3.326,
|
| 6 |
+
"eval_steps_per_second": 0.416,
|
| 7 |
+
"eval_wer": 7.215361500971087
|
| 8 |
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.048,
|
| 3 |
+
"total_flos": 4.891718061785088e+20,
|
| 4 |
+
"train_loss": 0.0,
|
| 5 |
+
"train_runtime": 289.8068,
|
| 6 |
+
"train_samples_per_second": 552.092,
|
| 7 |
+
"train_steps_per_second": 34.506
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2724 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 7.215361500971087,
|
| 3 |
+
"best_model_checkpoint": "./checkpoint-9000",
|
| 4 |
+
"epoch": 5.048,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 9000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0025,
|
| 13 |
+
"grad_norm": 6.131621360778809,
|
| 14 |
+
"learning_rate": 2.1875e-07,
|
| 15 |
+
"loss": 0.9345,
|
| 16 |
+
"step": 25
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.005,
|
| 20 |
+
"grad_norm": 6.021520137786865,
|
| 21 |
+
"learning_rate": 4.375e-07,
|
| 22 |
+
"loss": 0.8231,
|
| 23 |
+
"step": 50
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.0075,
|
| 27 |
+
"grad_norm": 5.526496410369873,
|
| 28 |
+
"learning_rate": 6.5625e-07,
|
| 29 |
+
"loss": 0.5623,
|
| 30 |
+
"step": 75
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.01,
|
| 34 |
+
"grad_norm": 4.9277825355529785,
|
| 35 |
+
"learning_rate": 8.75e-07,
|
| 36 |
+
"loss": 0.4173,
|
| 37 |
+
"step": 100
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.0125,
|
| 41 |
+
"grad_norm": 4.292990684509277,
|
| 42 |
+
"learning_rate": 1.09375e-06,
|
| 43 |
+
"loss": 0.385,
|
| 44 |
+
"step": 125
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.015,
|
| 48 |
+
"grad_norm": 5.749295234680176,
|
| 49 |
+
"learning_rate": 1.3125e-06,
|
| 50 |
+
"loss": 0.3931,
|
| 51 |
+
"step": 150
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.0175,
|
| 55 |
+
"grad_norm": 3.8306965827941895,
|
| 56 |
+
"learning_rate": 1.5312499999999997e-06,
|
| 57 |
+
"loss": 0.3516,
|
| 58 |
+
"step": 175
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.02,
|
| 62 |
+
"grad_norm": 4.687748908996582,
|
| 63 |
+
"learning_rate": 1.75e-06,
|
| 64 |
+
"loss": 0.3235,
|
| 65 |
+
"step": 200
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.0225,
|
| 69 |
+
"grad_norm": 4.232759952545166,
|
| 70 |
+
"learning_rate": 1.96875e-06,
|
| 71 |
+
"loss": 0.3314,
|
| 72 |
+
"step": 225
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.025,
|
| 76 |
+
"grad_norm": 4.185751914978027,
|
| 77 |
+
"learning_rate": 2.1875e-06,
|
| 78 |
+
"loss": 0.309,
|
| 79 |
+
"step": 250
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.0275,
|
| 83 |
+
"grad_norm": 4.818612098693848,
|
| 84 |
+
"learning_rate": 2.40625e-06,
|
| 85 |
+
"loss": 0.2991,
|
| 86 |
+
"step": 275
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.03,
|
| 90 |
+
"grad_norm": 4.171736717224121,
|
| 91 |
+
"learning_rate": 2.625e-06,
|
| 92 |
+
"loss": 0.2832,
|
| 93 |
+
"step": 300
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.0325,
|
| 97 |
+
"grad_norm": 5.217376708984375,
|
| 98 |
+
"learning_rate": 2.8437499999999997e-06,
|
| 99 |
+
"loss": 0.2873,
|
| 100 |
+
"step": 325
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.035,
|
| 104 |
+
"grad_norm": 4.671106815338135,
|
| 105 |
+
"learning_rate": 3.0624999999999995e-06,
|
| 106 |
+
"loss": 0.2957,
|
| 107 |
+
"step": 350
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.0375,
|
| 111 |
+
"grad_norm": 3.9175262451171875,
|
| 112 |
+
"learning_rate": 3.2812499999999997e-06,
|
| 113 |
+
"loss": 0.2634,
|
| 114 |
+
"step": 375
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.04,
|
| 118 |
+
"grad_norm": 4.647582054138184,
|
| 119 |
+
"learning_rate": 3.5e-06,
|
| 120 |
+
"loss": 0.2541,
|
| 121 |
+
"step": 400
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.0425,
|
| 125 |
+
"grad_norm": 3.25675368309021,
|
| 126 |
+
"learning_rate": 3.7187499999999998e-06,
|
| 127 |
+
"loss": 0.2244,
|
| 128 |
+
"step": 425
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.045,
|
| 132 |
+
"grad_norm": 4.597206115722656,
|
| 133 |
+
"learning_rate": 3.9375e-06,
|
| 134 |
+
"loss": 0.2492,
|
| 135 |
+
"step": 450
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.0475,
|
| 139 |
+
"grad_norm": 4.602332592010498,
|
| 140 |
+
"learning_rate": 4.156249999999999e-06,
|
| 141 |
+
"loss": 0.246,
|
| 142 |
+
"step": 475
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.05,
|
| 146 |
+
"grad_norm": 3.6419622898101807,
|
| 147 |
+
"learning_rate": 4.375e-06,
|
| 148 |
+
"loss": 0.2208,
|
| 149 |
+
"step": 500
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.05,
|
| 153 |
+
"eval_loss": 0.2592349350452423,
|
| 154 |
+
"eval_runtime": 4116.5906,
|
| 155 |
+
"eval_samples_per_second": 3.311,
|
| 156 |
+
"eval_steps_per_second": 0.414,
|
| 157 |
+
"eval_wer": 20.691487412510533,
|
| 158 |
+
"step": 500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.0525,
|
| 162 |
+
"grad_norm": 3.6599488258361816,
|
| 163 |
+
"learning_rate": 4.363486842105263e-06,
|
| 164 |
+
"loss": 0.2539,
|
| 165 |
+
"step": 525
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.055,
|
| 169 |
+
"grad_norm": 3.6934616565704346,
|
| 170 |
+
"learning_rate": 4.351973684210526e-06,
|
| 171 |
+
"loss": 0.2313,
|
| 172 |
+
"step": 550
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 0.0575,
|
| 176 |
+
"grad_norm": 3.7546138763427734,
|
| 177 |
+
"learning_rate": 4.340460526315789e-06,
|
| 178 |
+
"loss": 0.2272,
|
| 179 |
+
"step": 575
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 0.06,
|
| 183 |
+
"grad_norm": 3.096877098083496,
|
| 184 |
+
"learning_rate": 4.3289473684210525e-06,
|
| 185 |
+
"loss": 0.2373,
|
| 186 |
+
"step": 600
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 0.0625,
|
| 190 |
+
"grad_norm": 3.572812795639038,
|
| 191 |
+
"learning_rate": 4.3174342105263155e-06,
|
| 192 |
+
"loss": 0.2285,
|
| 193 |
+
"step": 625
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 0.065,
|
| 197 |
+
"grad_norm": 3.3494396209716797,
|
| 198 |
+
"learning_rate": 4.3059210526315785e-06,
|
| 199 |
+
"loss": 0.2293,
|
| 200 |
+
"step": 650
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.0675,
|
| 204 |
+
"grad_norm": 3.5156869888305664,
|
| 205 |
+
"learning_rate": 4.2944078947368415e-06,
|
| 206 |
+
"loss": 0.2063,
|
| 207 |
+
"step": 675
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.07,
|
| 211 |
+
"grad_norm": 3.698807716369629,
|
| 212 |
+
"learning_rate": 4.282894736842105e-06,
|
| 213 |
+
"loss": 0.2113,
|
| 214 |
+
"step": 700
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"epoch": 0.0725,
|
| 218 |
+
"grad_norm": 3.716585636138916,
|
| 219 |
+
"learning_rate": 4.271381578947368e-06,
|
| 220 |
+
"loss": 0.2055,
|
| 221 |
+
"step": 725
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.075,
|
| 225 |
+
"grad_norm": 4.204227924346924,
|
| 226 |
+
"learning_rate": 4.2598684210526314e-06,
|
| 227 |
+
"loss": 0.2114,
|
| 228 |
+
"step": 750
|
| 229 |
+
},
|
| 230 |
+
{
|
| 231 |
+
"epoch": 0.0775,
|
| 232 |
+
"grad_norm": 3.479562282562256,
|
| 233 |
+
"learning_rate": 4.2483552631578944e-06,
|
| 234 |
+
"loss": 0.2224,
|
| 235 |
+
"step": 775
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 0.08,
|
| 239 |
+
"grad_norm": 4.5203094482421875,
|
| 240 |
+
"learning_rate": 4.2368421052631575e-06,
|
| 241 |
+
"loss": 0.2523,
|
| 242 |
+
"step": 800
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.0825,
|
| 246 |
+
"grad_norm": 3.6081738471984863,
|
| 247 |
+
"learning_rate": 4.2253289473684205e-06,
|
| 248 |
+
"loss": 0.2383,
|
| 249 |
+
"step": 825
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.085,
|
| 253 |
+
"grad_norm": 3.2602758407592773,
|
| 254 |
+
"learning_rate": 4.2138157894736835e-06,
|
| 255 |
+
"loss": 0.1808,
|
| 256 |
+
"step": 850
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 0.0875,
|
| 260 |
+
"grad_norm": 3.6786868572235107,
|
| 261 |
+
"learning_rate": 4.202302631578947e-06,
|
| 262 |
+
"loss": 0.1747,
|
| 263 |
+
"step": 875
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 0.09,
|
| 267 |
+
"grad_norm": 3.1120803356170654,
|
| 268 |
+
"learning_rate": 4.19078947368421e-06,
|
| 269 |
+
"loss": 0.1662,
|
| 270 |
+
"step": 900
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 0.0925,
|
| 274 |
+
"grad_norm": 3.1962203979492188,
|
| 275 |
+
"learning_rate": 4.179276315789473e-06,
|
| 276 |
+
"loss": 0.1771,
|
| 277 |
+
"step": 925
|
| 278 |
+
},
|
| 279 |
+
{
|
| 280 |
+
"epoch": 0.095,
|
| 281 |
+
"grad_norm": 3.172363758087158,
|
| 282 |
+
"learning_rate": 4.167763157894736e-06,
|
| 283 |
+
"loss": 0.1751,
|
| 284 |
+
"step": 950
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.0975,
|
| 288 |
+
"grad_norm": 2.4304590225219727,
|
| 289 |
+
"learning_rate": 4.156249999999999e-06,
|
| 290 |
+
"loss": 0.1701,
|
| 291 |
+
"step": 975
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.1,
|
| 295 |
+
"grad_norm": 3.193345308303833,
|
| 296 |
+
"learning_rate": 4.144736842105262e-06,
|
| 297 |
+
"loss": 0.1489,
|
| 298 |
+
"step": 1000
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 0.1,
|
| 302 |
+
"eval_loss": 0.1971057653427124,
|
| 303 |
+
"eval_runtime": 4130.6867,
|
| 304 |
+
"eval_samples_per_second": 3.3,
|
| 305 |
+
"eval_steps_per_second": 0.413,
|
| 306 |
+
"eval_wer": 14.68265601524424,
|
| 307 |
+
"step": 1000
|
| 308 |
+
},
|
| 309 |
+
{
|
| 310 |
+
"epoch": 0.1025,
|
| 311 |
+
"grad_norm": 3.322065591812134,
|
| 312 |
+
"learning_rate": 4.133223684210526e-06,
|
| 313 |
+
"loss": 0.1701,
|
| 314 |
+
"step": 1025
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.105,
|
| 318 |
+
"grad_norm": 3.5462722778320312,
|
| 319 |
+
"learning_rate": 4.121710526315789e-06,
|
| 320 |
+
"loss": 0.1875,
|
| 321 |
+
"step": 1050
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 0.1075,
|
| 325 |
+
"grad_norm": 3.39326810836792,
|
| 326 |
+
"learning_rate": 4.110197368421052e-06,
|
| 327 |
+
"loss": 0.1506,
|
| 328 |
+
"step": 1075
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 0.11,
|
| 332 |
+
"grad_norm": 2.9165821075439453,
|
| 333 |
+
"learning_rate": 4.098684210526315e-06,
|
| 334 |
+
"loss": 0.1525,
|
| 335 |
+
"step": 1100
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 0.1125,
|
| 339 |
+
"grad_norm": 3.262007236480713,
|
| 340 |
+
"learning_rate": 4.087171052631578e-06,
|
| 341 |
+
"loss": 0.157,
|
| 342 |
+
"step": 1125
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 0.115,
|
| 346 |
+
"grad_norm": 2.4523119926452637,
|
| 347 |
+
"learning_rate": 4.075657894736842e-06,
|
| 348 |
+
"loss": 0.1416,
|
| 349 |
+
"step": 1150
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 0.1175,
|
| 353 |
+
"grad_norm": 2.7651798725128174,
|
| 354 |
+
"learning_rate": 4.064144736842105e-06,
|
| 355 |
+
"loss": 0.1527,
|
| 356 |
+
"step": 1175
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.12,
|
| 360 |
+
"grad_norm": 3.609523296356201,
|
| 361 |
+
"learning_rate": 4.052631578947368e-06,
|
| 362 |
+
"loss": 0.1822,
|
| 363 |
+
"step": 1200
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 0.1225,
|
| 367 |
+
"grad_norm": 3.8101985454559326,
|
| 368 |
+
"learning_rate": 4.041118421052631e-06,
|
| 369 |
+
"loss": 0.1703,
|
| 370 |
+
"step": 1225
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 0.125,
|
| 374 |
+
"grad_norm": 3.8921287059783936,
|
| 375 |
+
"learning_rate": 4.029605263157894e-06,
|
| 376 |
+
"loss": 0.1924,
|
| 377 |
+
"step": 1250
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 0.1275,
|
| 381 |
+
"grad_norm": 4.463279724121094,
|
| 382 |
+
"learning_rate": 4.018092105263157e-06,
|
| 383 |
+
"loss": 0.1818,
|
| 384 |
+
"step": 1275
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 0.13,
|
| 388 |
+
"grad_norm": 3.6556308269500732,
|
| 389 |
+
"learning_rate": 4.00657894736842e-06,
|
| 390 |
+
"loss": 0.1726,
|
| 391 |
+
"step": 1300
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 0.1325,
|
| 395 |
+
"grad_norm": 2.98067569732666,
|
| 396 |
+
"learning_rate": 3.995065789473683e-06,
|
| 397 |
+
"loss": 0.174,
|
| 398 |
+
"step": 1325
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.135,
|
| 402 |
+
"grad_norm": 2.8287429809570312,
|
| 403 |
+
"learning_rate": 3.983552631578947e-06,
|
| 404 |
+
"loss": 0.1631,
|
| 405 |
+
"step": 1350
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 0.1375,
|
| 409 |
+
"grad_norm": 2.6438794136047363,
|
| 410 |
+
"learning_rate": 3.97203947368421e-06,
|
| 411 |
+
"loss": 0.1475,
|
| 412 |
+
"step": 1375
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 0.14,
|
| 416 |
+
"grad_norm": 3.513123035430908,
|
| 417 |
+
"learning_rate": 3.960526315789473e-06,
|
| 418 |
+
"loss": 0.1457,
|
| 419 |
+
"step": 1400
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 0.1425,
|
| 423 |
+
"grad_norm": 2.4688916206359863,
|
| 424 |
+
"learning_rate": 3.949013157894737e-06,
|
| 425 |
+
"loss": 0.1375,
|
| 426 |
+
"step": 1425
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 0.145,
|
| 430 |
+
"grad_norm": 4.005943775177002,
|
| 431 |
+
"learning_rate": 3.9375e-06,
|
| 432 |
+
"loss": 0.1623,
|
| 433 |
+
"step": 1450
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 0.1475,
|
| 437 |
+
"grad_norm": 2.91786789894104,
|
| 438 |
+
"learning_rate": 3.925986842105263e-06,
|
| 439 |
+
"loss": 0.1701,
|
| 440 |
+
"step": 1475
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.15,
|
| 444 |
+
"grad_norm": 3.5332415103912354,
|
| 445 |
+
"learning_rate": 3.914473684210526e-06,
|
| 446 |
+
"loss": 0.1973,
|
| 447 |
+
"step": 1500
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 0.15,
|
| 451 |
+
"eval_loss": 0.17469166219234467,
|
| 452 |
+
"eval_runtime": 4132.0041,
|
| 453 |
+
"eval_samples_per_second": 3.299,
|
| 454 |
+
"eval_steps_per_second": 0.412,
|
| 455 |
+
"eval_wer": 12.377697973542453,
|
| 456 |
+
"step": 1500
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 0.1525,
|
| 460 |
+
"grad_norm": 4.05070686340332,
|
| 461 |
+
"learning_rate": 3.902960526315789e-06,
|
| 462 |
+
"loss": 0.1796,
|
| 463 |
+
"step": 1525
|
| 464 |
+
},
|
| 465 |
+
{
|
| 466 |
+
"epoch": 0.155,
|
| 467 |
+
"grad_norm": 2.989821195602417,
|
| 468 |
+
"learning_rate": 3.891447368421052e-06,
|
| 469 |
+
"loss": 0.1561,
|
| 470 |
+
"step": 1550
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.1575,
|
| 474 |
+
"grad_norm": 2.9603219032287598,
|
| 475 |
+
"learning_rate": 3.879934210526315e-06,
|
| 476 |
+
"loss": 0.1609,
|
| 477 |
+
"step": 1575
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 0.16,
|
| 481 |
+
"grad_norm": 3.2663583755493164,
|
| 482 |
+
"learning_rate": 3.868421052631579e-06,
|
| 483 |
+
"loss": 0.1833,
|
| 484 |
+
"step": 1600
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"epoch": 0.1625,
|
| 488 |
+
"grad_norm": 3.459775686264038,
|
| 489 |
+
"learning_rate": 3.856907894736842e-06,
|
| 490 |
+
"loss": 0.1727,
|
| 491 |
+
"step": 1625
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 0.165,
|
| 495 |
+
"grad_norm": 3.427720069885254,
|
| 496 |
+
"learning_rate": 3.845394736842105e-06,
|
| 497 |
+
"loss": 0.181,
|
| 498 |
+
"step": 1650
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 0.1675,
|
| 502 |
+
"grad_norm": 4.471118450164795,
|
| 503 |
+
"learning_rate": 3.833881578947368e-06,
|
| 504 |
+
"loss": 0.1536,
|
| 505 |
+
"step": 1675
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 0.17,
|
| 509 |
+
"grad_norm": 3.1428306102752686,
|
| 510 |
+
"learning_rate": 3.822368421052632e-06,
|
| 511 |
+
"loss": 0.1372,
|
| 512 |
+
"step": 1700
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 1.0021,
|
| 516 |
+
"grad_norm": 2.8270132541656494,
|
| 517 |
+
"learning_rate": 3.8108552631578944e-06,
|
| 518 |
+
"loss": 0.1454,
|
| 519 |
+
"step": 1725
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 1.0046,
|
| 523 |
+
"grad_norm": 3.0873589515686035,
|
| 524 |
+
"learning_rate": 3.799342105263158e-06,
|
| 525 |
+
"loss": 0.1303,
|
| 526 |
+
"step": 1750
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 1.0071,
|
| 530 |
+
"grad_norm": 3.187711000442505,
|
| 531 |
+
"learning_rate": 3.787828947368421e-06,
|
| 532 |
+
"loss": 0.1383,
|
| 533 |
+
"step": 1775
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 1.0096,
|
| 537 |
+
"grad_norm": 3.1710643768310547,
|
| 538 |
+
"learning_rate": 3.776315789473684e-06,
|
| 539 |
+
"loss": 0.1626,
|
| 540 |
+
"step": 1800
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 1.0121,
|
| 544 |
+
"grad_norm": 3.4516818523406982,
|
| 545 |
+
"learning_rate": 3.7648026315789473e-06,
|
| 546 |
+
"loss": 0.1405,
|
| 547 |
+
"step": 1825
|
| 548 |
+
},
|
| 549 |
+
{
|
| 550 |
+
"epoch": 1.0146,
|
| 551 |
+
"grad_norm": 2.930408000946045,
|
| 552 |
+
"learning_rate": 3.7532894736842103e-06,
|
| 553 |
+
"loss": 0.143,
|
| 554 |
+
"step": 1850
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 1.0171,
|
| 558 |
+
"grad_norm": 3.066941261291504,
|
| 559 |
+
"learning_rate": 3.7417763157894733e-06,
|
| 560 |
+
"loss": 0.1437,
|
| 561 |
+
"step": 1875
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 1.0196,
|
| 565 |
+
"grad_norm": 3.389916181564331,
|
| 566 |
+
"learning_rate": 3.7302631578947363e-06,
|
| 567 |
+
"loss": 0.1289,
|
| 568 |
+
"step": 1900
|
| 569 |
+
},
|
| 570 |
+
{
|
| 571 |
+
"epoch": 1.0221,
|
| 572 |
+
"grad_norm": 3.048574209213257,
|
| 573 |
+
"learning_rate": 3.7187499999999998e-06,
|
| 574 |
+
"loss": 0.1415,
|
| 575 |
+
"step": 1925
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 1.0246,
|
| 579 |
+
"grad_norm": 2.5267295837402344,
|
| 580 |
+
"learning_rate": 3.7072368421052628e-06,
|
| 581 |
+
"loss": 0.1386,
|
| 582 |
+
"step": 1950
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 1.0271,
|
| 586 |
+
"grad_norm": 3.151757001876831,
|
| 587 |
+
"learning_rate": 3.6957236842105258e-06,
|
| 588 |
+
"loss": 0.1436,
|
| 589 |
+
"step": 1975
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 1.0296,
|
| 593 |
+
"grad_norm": 3.629039764404297,
|
| 594 |
+
"learning_rate": 3.684210526315789e-06,
|
| 595 |
+
"loss": 0.1353,
|
| 596 |
+
"step": 2000
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 1.0296,
|
| 600 |
+
"eval_loss": 0.1527385264635086,
|
| 601 |
+
"eval_runtime": 4116.9756,
|
| 602 |
+
"eval_samples_per_second": 3.311,
|
| 603 |
+
"eval_steps_per_second": 0.414,
|
| 604 |
+
"eval_wer": 10.719520685990693,
|
| 605 |
+
"step": 2000
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 1.0321,
|
| 609 |
+
"grad_norm": 1.8788173198699951,
|
| 610 |
+
"learning_rate": 3.6726973684210522e-06,
|
| 611 |
+
"loss": 0.1322,
|
| 612 |
+
"step": 2025
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 1.0346,
|
| 616 |
+
"grad_norm": 2.587233066558838,
|
| 617 |
+
"learning_rate": 3.6611842105263157e-06,
|
| 618 |
+
"loss": 0.1176,
|
| 619 |
+
"step": 2050
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 1.0371,
|
| 623 |
+
"grad_norm": 4.001532077789307,
|
| 624 |
+
"learning_rate": 3.6496710526315787e-06,
|
| 625 |
+
"loss": 0.1233,
|
| 626 |
+
"step": 2075
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 1.0396,
|
| 630 |
+
"grad_norm": 3.3947739601135254,
|
| 631 |
+
"learning_rate": 3.638157894736842e-06,
|
| 632 |
+
"loss": 0.1188,
|
| 633 |
+
"step": 2100
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 1.0421,
|
| 637 |
+
"grad_norm": 3.4743120670318604,
|
| 638 |
+
"learning_rate": 3.626644736842105e-06,
|
| 639 |
+
"loss": 0.1318,
|
| 640 |
+
"step": 2125
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 1.0446,
|
| 644 |
+
"grad_norm": 2.9288718700408936,
|
| 645 |
+
"learning_rate": 3.615131578947368e-06,
|
| 646 |
+
"loss": 0.1224,
|
| 647 |
+
"step": 2150
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 1.0471,
|
| 651 |
+
"grad_norm": 2.6081368923187256,
|
| 652 |
+
"learning_rate": 3.603618421052631e-06,
|
| 653 |
+
"loss": 0.1232,
|
| 654 |
+
"step": 2175
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 1.0496,
|
| 658 |
+
"grad_norm": 2.4068429470062256,
|
| 659 |
+
"learning_rate": 3.5921052631578946e-06,
|
| 660 |
+
"loss": 0.1073,
|
| 661 |
+
"step": 2200
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 1.0521,
|
| 665 |
+
"grad_norm": 3.049074411392212,
|
| 666 |
+
"learning_rate": 3.5805921052631576e-06,
|
| 667 |
+
"loss": 0.1071,
|
| 668 |
+
"step": 2225
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 1.0546,
|
| 672 |
+
"grad_norm": 2.0809032917022705,
|
| 673 |
+
"learning_rate": 3.5690789473684206e-06,
|
| 674 |
+
"loss": 0.1217,
|
| 675 |
+
"step": 2250
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 1.0571,
|
| 679 |
+
"grad_norm": 3.0854332447052,
|
| 680 |
+
"learning_rate": 3.5575657894736836e-06,
|
| 681 |
+
"loss": 0.1332,
|
| 682 |
+
"step": 2275
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 1.0596,
|
| 686 |
+
"grad_norm": 3.580145835876465,
|
| 687 |
+
"learning_rate": 3.546052631578947e-06,
|
| 688 |
+
"loss": 0.131,
|
| 689 |
+
"step": 2300
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 1.0621,
|
| 693 |
+
"grad_norm": 3.8924479484558105,
|
| 694 |
+
"learning_rate": 3.53453947368421e-06,
|
| 695 |
+
"loss": 0.136,
|
| 696 |
+
"step": 2325
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 1.0646,
|
| 700 |
+
"grad_norm": 2.8398871421813965,
|
| 701 |
+
"learning_rate": 3.523026315789473e-06,
|
| 702 |
+
"loss": 0.1081,
|
| 703 |
+
"step": 2350
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 1.0671,
|
| 707 |
+
"grad_norm": 3.007026195526123,
|
| 708 |
+
"learning_rate": 3.511513157894737e-06,
|
| 709 |
+
"loss": 0.1115,
|
| 710 |
+
"step": 2375
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 1.0695999999999999,
|
| 714 |
+
"grad_norm": 1.5712552070617676,
|
| 715 |
+
"learning_rate": 3.5e-06,
|
| 716 |
+
"loss": 0.1183,
|
| 717 |
+
"step": 2400
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 1.0721,
|
| 721 |
+
"grad_norm": 3.844963312149048,
|
| 722 |
+
"learning_rate": 3.488486842105263e-06,
|
| 723 |
+
"loss": 0.113,
|
| 724 |
+
"step": 2425
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 1.0746,
|
| 728 |
+
"grad_norm": 2.8939759731292725,
|
| 729 |
+
"learning_rate": 3.476973684210526e-06,
|
| 730 |
+
"loss": 0.1115,
|
| 731 |
+
"step": 2450
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 1.0771,
|
| 735 |
+
"grad_norm": 1.8150537014007568,
|
| 736 |
+
"learning_rate": 3.4654605263157894e-06,
|
| 737 |
+
"loss": 0.1117,
|
| 738 |
+
"step": 2475
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 1.0796000000000001,
|
| 742 |
+
"grad_norm": 2.839418649673462,
|
| 743 |
+
"learning_rate": 3.4539473684210524e-06,
|
| 744 |
+
"loss": 0.1065,
|
| 745 |
+
"step": 2500
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 1.0796000000000001,
|
| 749 |
+
"eval_loss": 0.1456422209739685,
|
| 750 |
+
"eval_runtime": 4133.4016,
|
| 751 |
+
"eval_samples_per_second": 3.298,
|
| 752 |
+
"eval_steps_per_second": 0.412,
|
| 753 |
+
"eval_wer": 9.869361281102277,
|
| 754 |
+
"step": 2500
|
| 755 |
+
},
|
| 756 |
+
{
|
| 757 |
+
"epoch": 1.0821,
|
| 758 |
+
"grad_norm": 3.4274985790252686,
|
| 759 |
+
"learning_rate": 3.4424342105263154e-06,
|
| 760 |
+
"loss": 0.1067,
|
| 761 |
+
"step": 2525
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 1.0846,
|
| 765 |
+
"grad_norm": 2.2946057319641113,
|
| 766 |
+
"learning_rate": 3.4309210526315784e-06,
|
| 767 |
+
"loss": 0.1038,
|
| 768 |
+
"step": 2550
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 1.0871,
|
| 772 |
+
"grad_norm": 2.5364551544189453,
|
| 773 |
+
"learning_rate": 3.419407894736842e-06,
|
| 774 |
+
"loss": 0.1073,
|
| 775 |
+
"step": 2575
|
| 776 |
+
},
|
| 777 |
+
{
|
| 778 |
+
"epoch": 1.0896,
|
| 779 |
+
"grad_norm": 2.9779515266418457,
|
| 780 |
+
"learning_rate": 3.4083552631578944e-06,
|
| 781 |
+
"loss": 0.1067,
|
| 782 |
+
"step": 2600
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 1.0921,
|
| 786 |
+
"grad_norm": 2.502685308456421,
|
| 787 |
+
"learning_rate": 3.3968421052631574e-06,
|
| 788 |
+
"loss": 0.1229,
|
| 789 |
+
"step": 2625
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 1.0946,
|
| 793 |
+
"grad_norm": 2.181756019592285,
|
| 794 |
+
"learning_rate": 3.3853289473684205e-06,
|
| 795 |
+
"loss": 0.1071,
|
| 796 |
+
"step": 2650
|
| 797 |
+
},
|
| 798 |
+
{
|
| 799 |
+
"epoch": 1.0971,
|
| 800 |
+
"grad_norm": 2.428738594055176,
|
| 801 |
+
"learning_rate": 3.3738157894736843e-06,
|
| 802 |
+
"loss": 0.101,
|
| 803 |
+
"step": 2675
|
| 804 |
+
},
|
| 805 |
+
{
|
| 806 |
+
"epoch": 1.0996,
|
| 807 |
+
"grad_norm": 3.797952651977539,
|
| 808 |
+
"learning_rate": 3.3623026315789473e-06,
|
| 809 |
+
"loss": 0.1198,
|
| 810 |
+
"step": 2700
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 1.1021,
|
| 814 |
+
"grad_norm": 2.9902758598327637,
|
| 815 |
+
"learning_rate": 3.3507894736842103e-06,
|
| 816 |
+
"loss": 0.1013,
|
| 817 |
+
"step": 2725
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"epoch": 1.1046,
|
| 821 |
+
"grad_norm": 3.0514307022094727,
|
| 822 |
+
"learning_rate": 3.3392763157894734e-06,
|
| 823 |
+
"loss": 0.1075,
|
| 824 |
+
"step": 2750
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 1.1071,
|
| 828 |
+
"grad_norm": 3.2877554893493652,
|
| 829 |
+
"learning_rate": 3.327763157894737e-06,
|
| 830 |
+
"loss": 0.1059,
|
| 831 |
+
"step": 2775
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 1.1096,
|
| 835 |
+
"grad_norm": 2.3952691555023193,
|
| 836 |
+
"learning_rate": 3.31625e-06,
|
| 837 |
+
"loss": 0.0926,
|
| 838 |
+
"step": 2800
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 1.1121,
|
| 842 |
+
"grad_norm": 2.2840464115142822,
|
| 843 |
+
"learning_rate": 3.304736842105263e-06,
|
| 844 |
+
"loss": 0.1048,
|
| 845 |
+
"step": 2825
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"epoch": 1.1146,
|
| 849 |
+
"grad_norm": 2.7062416076660156,
|
| 850 |
+
"learning_rate": 3.293223684210526e-06,
|
| 851 |
+
"loss": 0.1049,
|
| 852 |
+
"step": 2850
|
| 853 |
+
},
|
| 854 |
+
{
|
| 855 |
+
"epoch": 1.1171,
|
| 856 |
+
"grad_norm": 2.971315860748291,
|
| 857 |
+
"learning_rate": 3.2817105263157893e-06,
|
| 858 |
+
"loss": 0.1073,
|
| 859 |
+
"step": 2875
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 1.1196,
|
| 863 |
+
"grad_norm": 2.8689844608306885,
|
| 864 |
+
"learning_rate": 3.2701973684210523e-06,
|
| 865 |
+
"loss": 0.1141,
|
| 866 |
+
"step": 2900
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 1.1221,
|
| 870 |
+
"grad_norm": 3.6150734424591064,
|
| 871 |
+
"learning_rate": 3.2586842105263153e-06,
|
| 872 |
+
"loss": 0.1066,
|
| 873 |
+
"step": 2925
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 1.1246,
|
| 877 |
+
"grad_norm": 2.3004024028778076,
|
| 878 |
+
"learning_rate": 3.2471710526315783e-06,
|
| 879 |
+
"loss": 0.1248,
|
| 880 |
+
"step": 2950
|
| 881 |
+
},
|
| 882 |
+
{
|
| 883 |
+
"epoch": 1.1271,
|
| 884 |
+
"grad_norm": 2.5995240211486816,
|
| 885 |
+
"learning_rate": 3.2356578947368417e-06,
|
| 886 |
+
"loss": 0.0972,
|
| 887 |
+
"step": 2975
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 1.1296,
|
| 891 |
+
"grad_norm": 2.957960367202759,
|
| 892 |
+
"learning_rate": 3.224144736842105e-06,
|
| 893 |
+
"loss": 0.106,
|
| 894 |
+
"step": 3000
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 1.1296,
|
| 898 |
+
"eval_loss": 0.13624447584152222,
|
| 899 |
+
"eval_runtime": 4123.4662,
|
| 900 |
+
"eval_samples_per_second": 3.305,
|
| 901 |
+
"eval_steps_per_second": 0.413,
|
| 902 |
+
"eval_wer": 9.09249148008355,
|
| 903 |
+
"step": 3000
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 1.1320999999999999,
|
| 907 |
+
"grad_norm": 2.653007984161377,
|
| 908 |
+
"learning_rate": 3.212631578947368e-06,
|
| 909 |
+
"loss": 0.1083,
|
| 910 |
+
"step": 3025
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 1.1346,
|
| 914 |
+
"grad_norm": 2.6895744800567627,
|
| 915 |
+
"learning_rate": 3.2011184210526316e-06,
|
| 916 |
+
"loss": 0.1119,
|
| 917 |
+
"step": 3050
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"epoch": 1.1371,
|
| 921 |
+
"grad_norm": 2.1507463455200195,
|
| 922 |
+
"learning_rate": 3.1896052631578946e-06,
|
| 923 |
+
"loss": 0.0944,
|
| 924 |
+
"step": 3075
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"epoch": 1.1396,
|
| 928 |
+
"grad_norm": 3.61063289642334,
|
| 929 |
+
"learning_rate": 3.1780921052631576e-06,
|
| 930 |
+
"loss": 0.095,
|
| 931 |
+
"step": 3100
|
| 932 |
+
},
|
| 933 |
+
{
|
| 934 |
+
"epoch": 1.1421000000000001,
|
| 935 |
+
"grad_norm": 2.570584774017334,
|
| 936 |
+
"learning_rate": 3.1665789473684206e-06,
|
| 937 |
+
"loss": 0.1076,
|
| 938 |
+
"step": 3125
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 1.1446,
|
| 942 |
+
"grad_norm": 3.05507230758667,
|
| 943 |
+
"learning_rate": 3.155065789473684e-06,
|
| 944 |
+
"loss": 0.1175,
|
| 945 |
+
"step": 3150
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 1.1471,
|
| 949 |
+
"grad_norm": 2.82817006111145,
|
| 950 |
+
"learning_rate": 3.143552631578947e-06,
|
| 951 |
+
"loss": 0.0965,
|
| 952 |
+
"step": 3175
|
| 953 |
+
},
|
| 954 |
+
{
|
| 955 |
+
"epoch": 1.1496,
|
| 956 |
+
"grad_norm": 2.336517572402954,
|
| 957 |
+
"learning_rate": 3.13203947368421e-06,
|
| 958 |
+
"loss": 0.0955,
|
| 959 |
+
"step": 3200
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 1.1521,
|
| 963 |
+
"grad_norm": 3.8640036582946777,
|
| 964 |
+
"learning_rate": 3.120526315789473e-06,
|
| 965 |
+
"loss": 0.1044,
|
| 966 |
+
"step": 3225
|
| 967 |
+
},
|
| 968 |
+
{
|
| 969 |
+
"epoch": 1.1546,
|
| 970 |
+
"grad_norm": 3.7205588817596436,
|
| 971 |
+
"learning_rate": 3.1090131578947366e-06,
|
| 972 |
+
"loss": 0.1013,
|
| 973 |
+
"step": 3250
|
| 974 |
+
},
|
| 975 |
+
{
|
| 976 |
+
"epoch": 1.1571,
|
| 977 |
+
"grad_norm": 2.1962900161743164,
|
| 978 |
+
"learning_rate": 3.0974999999999996e-06,
|
| 979 |
+
"loss": 0.0978,
|
| 980 |
+
"step": 3275
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 1.1596,
|
| 984 |
+
"grad_norm": 3.3310599327087402,
|
| 985 |
+
"learning_rate": 3.0859868421052626e-06,
|
| 986 |
+
"loss": 0.1089,
|
| 987 |
+
"step": 3300
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 1.1621,
|
| 991 |
+
"grad_norm": 2.699566602706909,
|
| 992 |
+
"learning_rate": 3.074473684210526e-06,
|
| 993 |
+
"loss": 0.1078,
|
| 994 |
+
"step": 3325
|
| 995 |
+
},
|
| 996 |
+
{
|
| 997 |
+
"epoch": 1.1646,
|
| 998 |
+
"grad_norm": 3.79370379447937,
|
| 999 |
+
"learning_rate": 3.0629605263157894e-06,
|
| 1000 |
+
"loss": 0.1118,
|
| 1001 |
+
"step": 3350
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 1.1671,
|
| 1005 |
+
"grad_norm": 1.9741384983062744,
|
| 1006 |
+
"learning_rate": 3.0514473684210525e-06,
|
| 1007 |
+
"loss": 0.1119,
|
| 1008 |
+
"step": 3375
|
| 1009 |
+
},
|
| 1010 |
+
{
|
| 1011 |
+
"epoch": 1.1696,
|
| 1012 |
+
"grad_norm": 2.29034686088562,
|
| 1013 |
+
"learning_rate": 3.0399342105263155e-06,
|
| 1014 |
+
"loss": 0.1015,
|
| 1015 |
+
"step": 3400
|
| 1016 |
+
},
|
| 1017 |
+
{
|
| 1018 |
+
"epoch": 2.0017,
|
| 1019 |
+
"grad_norm": 2.011443853378296,
|
| 1020 |
+
"learning_rate": 3.028421052631579e-06,
|
| 1021 |
+
"loss": 0.0708,
|
| 1022 |
+
"step": 3425
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 2.0042,
|
| 1026 |
+
"grad_norm": 1.2196134328842163,
|
| 1027 |
+
"learning_rate": 3.016907894736842e-06,
|
| 1028 |
+
"loss": 0.0668,
|
| 1029 |
+
"step": 3450
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 2.0067,
|
| 1033 |
+
"grad_norm": 2.863933563232422,
|
| 1034 |
+
"learning_rate": 3.005394736842105e-06,
|
| 1035 |
+
"loss": 0.0673,
|
| 1036 |
+
"step": 3475
|
| 1037 |
+
},
|
| 1038 |
+
{
|
| 1039 |
+
"epoch": 2.0092,
|
| 1040 |
+
"grad_norm": 1.9341013431549072,
|
| 1041 |
+
"learning_rate": 2.9938815789473684e-06,
|
| 1042 |
+
"loss": 0.0718,
|
| 1043 |
+
"step": 3500
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 2.0092,
|
| 1047 |
+
"eval_loss": 0.13255682587623596,
|
| 1048 |
+
"eval_runtime": 4133.4892,
|
| 1049 |
+
"eval_samples_per_second": 3.297,
|
| 1050 |
+
"eval_steps_per_second": 0.412,
|
| 1051 |
+
"eval_wer": 8.542819451060867,
|
| 1052 |
+
"step": 3500
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 2.0117,
|
| 1056 |
+
"grad_norm": 2.795734405517578,
|
| 1057 |
+
"learning_rate": 2.9823684210526314e-06,
|
| 1058 |
+
"loss": 0.071,
|
| 1059 |
+
"step": 3525
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 2.0142,
|
| 1063 |
+
"grad_norm": 1.982479214668274,
|
| 1064 |
+
"learning_rate": 2.9708552631578944e-06,
|
| 1065 |
+
"loss": 0.0629,
|
| 1066 |
+
"step": 3550
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 2.0167,
|
| 1070 |
+
"grad_norm": 3.168161630630493,
|
| 1071 |
+
"learning_rate": 2.9593421052631574e-06,
|
| 1072 |
+
"loss": 0.0593,
|
| 1073 |
+
"step": 3575
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 2.0192,
|
| 1077 |
+
"grad_norm": 2.259500741958618,
|
| 1078 |
+
"learning_rate": 2.947828947368421e-06,
|
| 1079 |
+
"loss": 0.0696,
|
| 1080 |
+
"step": 3600
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 2.0217,
|
| 1084 |
+
"grad_norm": 2.1626062393188477,
|
| 1085 |
+
"learning_rate": 2.936315789473684e-06,
|
| 1086 |
+
"loss": 0.0687,
|
| 1087 |
+
"step": 3625
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 2.0242,
|
| 1091 |
+
"grad_norm": 2.4419946670532227,
|
| 1092 |
+
"learning_rate": 2.924802631578947e-06,
|
| 1093 |
+
"loss": 0.0686,
|
| 1094 |
+
"step": 3650
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 2.0267,
|
| 1098 |
+
"grad_norm": 2.445758819580078,
|
| 1099 |
+
"learning_rate": 2.9132894736842103e-06,
|
| 1100 |
+
"loss": 0.0631,
|
| 1101 |
+
"step": 3675
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 2.0292,
|
| 1105 |
+
"grad_norm": 2.614476442337036,
|
| 1106 |
+
"learning_rate": 2.9017763157894737e-06,
|
| 1107 |
+
"loss": 0.0647,
|
| 1108 |
+
"step": 3700
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 2.0317,
|
| 1112 |
+
"grad_norm": 1.4166672229766846,
|
| 1113 |
+
"learning_rate": 2.8902631578947367e-06,
|
| 1114 |
+
"loss": 0.0653,
|
| 1115 |
+
"step": 3725
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 2.0342,
|
| 1119 |
+
"grad_norm": 1.8435245752334595,
|
| 1120 |
+
"learning_rate": 2.8787499999999998e-06,
|
| 1121 |
+
"loss": 0.0567,
|
| 1122 |
+
"step": 3750
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 2.0367,
|
| 1126 |
+
"grad_norm": 1.8179950714111328,
|
| 1127 |
+
"learning_rate": 2.867236842105263e-06,
|
| 1128 |
+
"loss": 0.0636,
|
| 1129 |
+
"step": 3775
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 2.0392,
|
| 1133 |
+
"grad_norm": 1.487122893333435,
|
| 1134 |
+
"learning_rate": 2.855723684210526e-06,
|
| 1135 |
+
"loss": 0.0598,
|
| 1136 |
+
"step": 3800
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 2.0417,
|
| 1140 |
+
"grad_norm": 2.9211690425872803,
|
| 1141 |
+
"learning_rate": 2.8442105263157892e-06,
|
| 1142 |
+
"loss": 0.0599,
|
| 1143 |
+
"step": 3825
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 2.0442,
|
| 1147 |
+
"grad_norm": 2.5018093585968018,
|
| 1148 |
+
"learning_rate": 2.8326973684210522e-06,
|
| 1149 |
+
"loss": 0.055,
|
| 1150 |
+
"step": 3850
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 2.0467,
|
| 1154 |
+
"grad_norm": 2.186502456665039,
|
| 1155 |
+
"learning_rate": 2.8211842105263157e-06,
|
| 1156 |
+
"loss": 0.0533,
|
| 1157 |
+
"step": 3875
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 2.0492,
|
| 1161 |
+
"grad_norm": 1.039233922958374,
|
| 1162 |
+
"learning_rate": 2.8096710526315787e-06,
|
| 1163 |
+
"loss": 0.0514,
|
| 1164 |
+
"step": 3900
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 2.0517,
|
| 1168 |
+
"grad_norm": 1.871267557144165,
|
| 1169 |
+
"learning_rate": 2.7981578947368417e-06,
|
| 1170 |
+
"loss": 0.0512,
|
| 1171 |
+
"step": 3925
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 2.0542,
|
| 1175 |
+
"grad_norm": 2.0849483013153076,
|
| 1176 |
+
"learning_rate": 2.7866447368421047e-06,
|
| 1177 |
+
"loss": 0.0579,
|
| 1178 |
+
"step": 3950
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 2.0567,
|
| 1182 |
+
"grad_norm": 1.6887531280517578,
|
| 1183 |
+
"learning_rate": 2.775131578947368e-06,
|
| 1184 |
+
"loss": 0.0575,
|
| 1185 |
+
"step": 3975
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 2.0592,
|
| 1189 |
+
"grad_norm": 1.88097083568573,
|
| 1190 |
+
"learning_rate": 2.763618421052631e-06,
|
| 1191 |
+
"loss": 0.0683,
|
| 1192 |
+
"step": 4000
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 2.0592,
|
| 1196 |
+
"eval_loss": 0.1342601627111435,
|
| 1197 |
+
"eval_runtime": 4125.8373,
|
| 1198 |
+
"eval_samples_per_second": 3.304,
|
| 1199 |
+
"eval_steps_per_second": 0.413,
|
| 1200 |
+
"eval_wer": 8.485103888013485,
|
| 1201 |
+
"step": 4000
|
| 1202 |
+
},
|
| 1203 |
+
{
|
| 1204 |
+
"epoch": 2.0617,
|
| 1205 |
+
"grad_norm": 2.1877427101135254,
|
| 1206 |
+
"learning_rate": 2.7521052631578946e-06,
|
| 1207 |
+
"loss": 0.0614,
|
| 1208 |
+
"step": 4025
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 2.0642,
|
| 1212 |
+
"grad_norm": 1.4176368713378906,
|
| 1213 |
+
"learning_rate": 2.740592105263158e-06,
|
| 1214 |
+
"loss": 0.0559,
|
| 1215 |
+
"step": 4050
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 2.0667,
|
| 1219 |
+
"grad_norm": 2.4362101554870605,
|
| 1220 |
+
"learning_rate": 2.729078947368421e-06,
|
| 1221 |
+
"loss": 0.0593,
|
| 1222 |
+
"step": 4075
|
| 1223 |
+
},
|
| 1224 |
+
{
|
| 1225 |
+
"epoch": 2.0692,
|
| 1226 |
+
"grad_norm": 1.8663033246994019,
|
| 1227 |
+
"learning_rate": 2.717565789473684e-06,
|
| 1228 |
+
"loss": 0.0591,
|
| 1229 |
+
"step": 4100
|
| 1230 |
+
},
|
| 1231 |
+
{
|
| 1232 |
+
"epoch": 2.0717,
|
| 1233 |
+
"grad_norm": 1.627626657485962,
|
| 1234 |
+
"learning_rate": 2.706052631578947e-06,
|
| 1235 |
+
"loss": 0.0637,
|
| 1236 |
+
"step": 4125
|
| 1237 |
+
},
|
| 1238 |
+
{
|
| 1239 |
+
"epoch": 2.0742,
|
| 1240 |
+
"grad_norm": 2.2072463035583496,
|
| 1241 |
+
"learning_rate": 2.6945394736842105e-06,
|
| 1242 |
+
"loss": 0.0571,
|
| 1243 |
+
"step": 4150
|
| 1244 |
+
},
|
| 1245 |
+
{
|
| 1246 |
+
"epoch": 2.0767,
|
| 1247 |
+
"grad_norm": 1.7411611080169678,
|
| 1248 |
+
"learning_rate": 2.6830263157894735e-06,
|
| 1249 |
+
"loss": 0.0588,
|
| 1250 |
+
"step": 4175
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 2.0792,
|
| 1254 |
+
"grad_norm": 1.324000358581543,
|
| 1255 |
+
"learning_rate": 2.6715131578947365e-06,
|
| 1256 |
+
"loss": 0.0482,
|
| 1257 |
+
"step": 4200
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 2.0817,
|
| 1261 |
+
"grad_norm": 1.4138795137405396,
|
| 1262 |
+
"learning_rate": 2.6599999999999995e-06,
|
| 1263 |
+
"loss": 0.0477,
|
| 1264 |
+
"step": 4225
|
| 1265 |
+
},
|
| 1266 |
+
{
|
| 1267 |
+
"epoch": 2.0842,
|
| 1268 |
+
"grad_norm": 2.403547763824463,
|
| 1269 |
+
"learning_rate": 2.648486842105263e-06,
|
| 1270 |
+
"loss": 0.0558,
|
| 1271 |
+
"step": 4250
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"epoch": 2.0867,
|
| 1275 |
+
"grad_norm": 1.3718703985214233,
|
| 1276 |
+
"learning_rate": 2.636973684210526e-06,
|
| 1277 |
+
"loss": 0.0546,
|
| 1278 |
+
"step": 4275
|
| 1279 |
+
},
|
| 1280 |
+
{
|
| 1281 |
+
"epoch": 2.0892,
|
| 1282 |
+
"grad_norm": 2.296445369720459,
|
| 1283 |
+
"learning_rate": 2.625460526315789e-06,
|
| 1284 |
+
"loss": 0.0554,
|
| 1285 |
+
"step": 4300
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"epoch": 2.0917,
|
| 1289 |
+
"grad_norm": 2.3471312522888184,
|
| 1290 |
+
"learning_rate": 2.613947368421052e-06,
|
| 1291 |
+
"loss": 0.051,
|
| 1292 |
+
"step": 4325
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 2.0942,
|
| 1296 |
+
"grad_norm": 1.6061975955963135,
|
| 1297 |
+
"learning_rate": 2.602434210526316e-06,
|
| 1298 |
+
"loss": 0.0548,
|
| 1299 |
+
"step": 4350
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 2.0967,
|
| 1303 |
+
"grad_norm": 2.979126453399658,
|
| 1304 |
+
"learning_rate": 2.590921052631579e-06,
|
| 1305 |
+
"loss": 0.0492,
|
| 1306 |
+
"step": 4375
|
| 1307 |
+
},
|
| 1308 |
+
{
|
| 1309 |
+
"epoch": 2.0992,
|
| 1310 |
+
"grad_norm": 1.7963169813156128,
|
| 1311 |
+
"learning_rate": 2.579407894736842e-06,
|
| 1312 |
+
"loss": 0.0514,
|
| 1313 |
+
"step": 4400
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 2.1017,
|
| 1317 |
+
"grad_norm": 2.4996039867401123,
|
| 1318 |
+
"learning_rate": 2.5678947368421053e-06,
|
| 1319 |
+
"loss": 0.0399,
|
| 1320 |
+
"step": 4425
|
| 1321 |
+
},
|
| 1322 |
+
{
|
| 1323 |
+
"epoch": 2.1042,
|
| 1324 |
+
"grad_norm": 1.7498191595077515,
|
| 1325 |
+
"learning_rate": 2.5563815789473683e-06,
|
| 1326 |
+
"loss": 0.0522,
|
| 1327 |
+
"step": 4450
|
| 1328 |
+
},
|
| 1329 |
+
{
|
| 1330 |
+
"epoch": 2.1067,
|
| 1331 |
+
"grad_norm": 1.413889765739441,
|
| 1332 |
+
"learning_rate": 2.5448684210526313e-06,
|
| 1333 |
+
"loss": 0.0517,
|
| 1334 |
+
"step": 4475
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 2.1092,
|
| 1338 |
+
"grad_norm": 2.0956978797912598,
|
| 1339 |
+
"learning_rate": 2.5333552631578943e-06,
|
| 1340 |
+
"loss": 0.0482,
|
| 1341 |
+
"step": 4500
|
| 1342 |
+
},
|
| 1343 |
+
{
|
| 1344 |
+
"epoch": 2.1092,
|
| 1345 |
+
"eval_loss": 0.1336347758769989,
|
| 1346 |
+
"eval_runtime": 4119.9162,
|
| 1347 |
+
"eval_samples_per_second": 3.308,
|
| 1348 |
+
"eval_steps_per_second": 0.414,
|
| 1349 |
+
"eval_wer": 8.104914067939463,
|
| 1350 |
+
"step": 4500
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 2.1117,
|
| 1354 |
+
"grad_norm": 3.138298749923706,
|
| 1355 |
+
"learning_rate": 2.5218421052631578e-06,
|
| 1356 |
+
"loss": 0.0568,
|
| 1357 |
+
"step": 4525
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 2.1142,
|
| 1361 |
+
"grad_norm": 1.4262772798538208,
|
| 1362 |
+
"learning_rate": 2.510328947368421e-06,
|
| 1363 |
+
"loss": 0.0475,
|
| 1364 |
+
"step": 4550
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 2.1167,
|
| 1368 |
+
"grad_norm": 3.3500139713287354,
|
| 1369 |
+
"learning_rate": 2.498815789473684e-06,
|
| 1370 |
+
"loss": 0.0474,
|
| 1371 |
+
"step": 4575
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 2.1192,
|
| 1375 |
+
"grad_norm": 4.509912014007568,
|
| 1376 |
+
"learning_rate": 2.4873026315789472e-06,
|
| 1377 |
+
"loss": 0.0586,
|
| 1378 |
+
"step": 4600
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 2.1217,
|
| 1382 |
+
"grad_norm": 2.1386468410491943,
|
| 1383 |
+
"learning_rate": 2.4757894736842102e-06,
|
| 1384 |
+
"loss": 0.062,
|
| 1385 |
+
"step": 4625
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 2.1242,
|
| 1389 |
+
"grad_norm": 1.1121129989624023,
|
| 1390 |
+
"learning_rate": 2.4642763157894733e-06,
|
| 1391 |
+
"loss": 0.0563,
|
| 1392 |
+
"step": 4650
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 2.1267,
|
| 1396 |
+
"grad_norm": 1.677538514137268,
|
| 1397 |
+
"learning_rate": 2.4527631578947363e-06,
|
| 1398 |
+
"loss": 0.0519,
|
| 1399 |
+
"step": 4675
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 2.1292,
|
| 1403 |
+
"grad_norm": 1.579513430595398,
|
| 1404 |
+
"learning_rate": 2.44125e-06,
|
| 1405 |
+
"loss": 0.0544,
|
| 1406 |
+
"step": 4700
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 2.1317,
|
| 1410 |
+
"grad_norm": 2.1100914478302,
|
| 1411 |
+
"learning_rate": 2.429736842105263e-06,
|
| 1412 |
+
"loss": 0.0578,
|
| 1413 |
+
"step": 4725
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 2.1342,
|
| 1417 |
+
"grad_norm": 1.779682993888855,
|
| 1418 |
+
"learning_rate": 2.418223684210526e-06,
|
| 1419 |
+
"loss": 0.0486,
|
| 1420 |
+
"step": 4750
|
| 1421 |
+
},
|
| 1422 |
+
{
|
| 1423 |
+
"epoch": 2.1367,
|
| 1424 |
+
"grad_norm": 1.7443439960479736,
|
| 1425 |
+
"learning_rate": 2.4067105263157896e-06,
|
| 1426 |
+
"loss": 0.0534,
|
| 1427 |
+
"step": 4775
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 2.1391999999999998,
|
| 1431 |
+
"grad_norm": 1.9388935565948486,
|
| 1432 |
+
"learning_rate": 2.3951973684210526e-06,
|
| 1433 |
+
"loss": 0.0516,
|
| 1434 |
+
"step": 4800
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 2.1417,
|
| 1438 |
+
"grad_norm": 1.82517409324646,
|
| 1439 |
+
"learning_rate": 2.3836842105263156e-06,
|
| 1440 |
+
"loss": 0.0451,
|
| 1441 |
+
"step": 4825
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 2.1442,
|
| 1445 |
+
"grad_norm": 1.9101967811584473,
|
| 1446 |
+
"learning_rate": 2.3721710526315786e-06,
|
| 1447 |
+
"loss": 0.0546,
|
| 1448 |
+
"step": 4850
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 2.1467,
|
| 1452 |
+
"grad_norm": 1.7242915630340576,
|
| 1453 |
+
"learning_rate": 2.360657894736842e-06,
|
| 1454 |
+
"loss": 0.0495,
|
| 1455 |
+
"step": 4875
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 2.1492,
|
| 1459 |
+
"grad_norm": 1.9127079248428345,
|
| 1460 |
+
"learning_rate": 2.349144736842105e-06,
|
| 1461 |
+
"loss": 0.0465,
|
| 1462 |
+
"step": 4900
|
| 1463 |
+
},
|
| 1464 |
+
{
|
| 1465 |
+
"epoch": 2.1517,
|
| 1466 |
+
"grad_norm": 2.7716519832611084,
|
| 1467 |
+
"learning_rate": 2.337631578947368e-06,
|
| 1468 |
+
"loss": 0.0493,
|
| 1469 |
+
"step": 4925
|
| 1470 |
+
},
|
| 1471 |
+
{
|
| 1472 |
+
"epoch": 2.1542,
|
| 1473 |
+
"grad_norm": 3.141706705093384,
|
| 1474 |
+
"learning_rate": 2.326118421052631e-06,
|
| 1475 |
+
"loss": 0.046,
|
| 1476 |
+
"step": 4950
|
| 1477 |
+
},
|
| 1478 |
+
{
|
| 1479 |
+
"epoch": 2.1567,
|
| 1480 |
+
"grad_norm": 2.2624270915985107,
|
| 1481 |
+
"learning_rate": 2.3146052631578945e-06,
|
| 1482 |
+
"loss": 0.0522,
|
| 1483 |
+
"step": 4975
|
| 1484 |
+
},
|
| 1485 |
+
{
|
| 1486 |
+
"epoch": 2.1592000000000002,
|
| 1487 |
+
"grad_norm": 1.2777652740478516,
|
| 1488 |
+
"learning_rate": 2.3030921052631575e-06,
|
| 1489 |
+
"loss": 0.0548,
|
| 1490 |
+
"step": 5000
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 2.1592000000000002,
|
| 1494 |
+
"eval_loss": 0.13162237405776978,
|
| 1495 |
+
"eval_runtime": 4127.2085,
|
| 1496 |
+
"eval_samples_per_second": 3.302,
|
| 1497 |
+
"eval_steps_per_second": 0.413,
|
| 1498 |
+
"eval_wer": 7.9244384184103485,
|
| 1499 |
+
"step": 5000
|
| 1500 |
+
},
|
| 1501 |
+
{
|
| 1502 |
+
"epoch": 2.1617,
|
| 1503 |
+
"grad_norm": 2.106818675994873,
|
| 1504 |
+
"learning_rate": 2.2915789473684206e-06,
|
| 1505 |
+
"loss": 0.0527,
|
| 1506 |
+
"step": 5025
|
| 1507 |
+
},
|
| 1508 |
+
{
|
| 1509 |
+
"epoch": 2.1642,
|
| 1510 |
+
"grad_norm": 2.2705554962158203,
|
| 1511 |
+
"learning_rate": 2.2800657894736844e-06,
|
| 1512 |
+
"loss": 0.0483,
|
| 1513 |
+
"step": 5050
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"epoch": 2.1667,
|
| 1517 |
+
"grad_norm": 1.5468271970748901,
|
| 1518 |
+
"learning_rate": 2.2685526315789474e-06,
|
| 1519 |
+
"loss": 0.0516,
|
| 1520 |
+
"step": 5075
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 2.1692,
|
| 1524 |
+
"grad_norm": 2.0331270694732666,
|
| 1525 |
+
"learning_rate": 2.2570394736842104e-06,
|
| 1526 |
+
"loss": 0.0551,
|
| 1527 |
+
"step": 5100
|
| 1528 |
+
},
|
| 1529 |
+
{
|
| 1530 |
+
"epoch": 3.0013,
|
| 1531 |
+
"grad_norm": 1.107423186302185,
|
| 1532 |
+
"learning_rate": 2.2455263157894734e-06,
|
| 1533 |
+
"loss": 0.0434,
|
| 1534 |
+
"step": 5125
|
| 1535 |
+
},
|
| 1536 |
+
{
|
| 1537 |
+
"epoch": 3.0038,
|
| 1538 |
+
"grad_norm": 3.9103100299835205,
|
| 1539 |
+
"learning_rate": 2.234013157894737e-06,
|
| 1540 |
+
"loss": 0.0362,
|
| 1541 |
+
"step": 5150
|
| 1542 |
+
},
|
| 1543 |
+
{
|
| 1544 |
+
"epoch": 3.0063,
|
| 1545 |
+
"grad_norm": 1.193088173866272,
|
| 1546 |
+
"learning_rate": 2.2225e-06,
|
| 1547 |
+
"loss": 0.0327,
|
| 1548 |
+
"step": 5175
|
| 1549 |
+
},
|
| 1550 |
+
{
|
| 1551 |
+
"epoch": 3.0088,
|
| 1552 |
+
"grad_norm": 1.0432852506637573,
|
| 1553 |
+
"learning_rate": 2.210986842105263e-06,
|
| 1554 |
+
"loss": 0.0326,
|
| 1555 |
+
"step": 5200
|
| 1556 |
+
},
|
| 1557 |
+
{
|
| 1558 |
+
"epoch": 3.0113,
|
| 1559 |
+
"grad_norm": 0.7116020917892456,
|
| 1560 |
+
"learning_rate": 2.199473684210526e-06,
|
| 1561 |
+
"loss": 0.0296,
|
| 1562 |
+
"step": 5225
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 3.0138,
|
| 1566 |
+
"grad_norm": 2.009617805480957,
|
| 1567 |
+
"learning_rate": 2.1879605263157894e-06,
|
| 1568 |
+
"loss": 0.0367,
|
| 1569 |
+
"step": 5250
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 3.0163,
|
| 1573 |
+
"grad_norm": 1.9047244787216187,
|
| 1574 |
+
"learning_rate": 2.1764473684210524e-06,
|
| 1575 |
+
"loss": 0.0347,
|
| 1576 |
+
"step": 5275
|
| 1577 |
+
},
|
| 1578 |
+
{
|
| 1579 |
+
"epoch": 3.0188,
|
| 1580 |
+
"grad_norm": 1.630439043045044,
|
| 1581 |
+
"learning_rate": 2.164934210526316e-06,
|
| 1582 |
+
"loss": 0.0291,
|
| 1583 |
+
"step": 5300
|
| 1584 |
+
},
|
| 1585 |
+
{
|
| 1586 |
+
"epoch": 3.0213,
|
| 1587 |
+
"grad_norm": 1.4158824682235718,
|
| 1588 |
+
"learning_rate": 2.153421052631579e-06,
|
| 1589 |
+
"loss": 0.0321,
|
| 1590 |
+
"step": 5325
|
| 1591 |
+
},
|
| 1592 |
+
{
|
| 1593 |
+
"epoch": 3.0238,
|
| 1594 |
+
"grad_norm": 1.2792794704437256,
|
| 1595 |
+
"learning_rate": 2.141907894736842e-06,
|
| 1596 |
+
"loss": 0.0338,
|
| 1597 |
+
"step": 5350
|
| 1598 |
+
},
|
| 1599 |
+
{
|
| 1600 |
+
"epoch": 3.0263,
|
| 1601 |
+
"grad_norm": 1.6505346298217773,
|
| 1602 |
+
"learning_rate": 2.1303947368421053e-06,
|
| 1603 |
+
"loss": 0.0348,
|
| 1604 |
+
"step": 5375
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 3.0288,
|
| 1608 |
+
"grad_norm": 1.5343618392944336,
|
| 1609 |
+
"learning_rate": 2.1188815789473683e-06,
|
| 1610 |
+
"loss": 0.0318,
|
| 1611 |
+
"step": 5400
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 3.0313,
|
| 1615 |
+
"grad_norm": 1.8325493335723877,
|
| 1616 |
+
"learning_rate": 2.1073684210526313e-06,
|
| 1617 |
+
"loss": 0.0333,
|
| 1618 |
+
"step": 5425
|
| 1619 |
+
},
|
| 1620 |
+
{
|
| 1621 |
+
"epoch": 3.0338,
|
| 1622 |
+
"grad_norm": 1.7224900722503662,
|
| 1623 |
+
"learning_rate": 2.0958552631578943e-06,
|
| 1624 |
+
"loss": 0.0322,
|
| 1625 |
+
"step": 5450
|
| 1626 |
+
},
|
| 1627 |
+
{
|
| 1628 |
+
"epoch": 3.0362999999999998,
|
| 1629 |
+
"grad_norm": 1.3443737030029297,
|
| 1630 |
+
"learning_rate": 2.0843421052631577e-06,
|
| 1631 |
+
"loss": 0.0304,
|
| 1632 |
+
"step": 5475
|
| 1633 |
+
},
|
| 1634 |
+
{
|
| 1635 |
+
"epoch": 3.0388,
|
| 1636 |
+
"grad_norm": 1.3260679244995117,
|
| 1637 |
+
"learning_rate": 2.0728289473684207e-06,
|
| 1638 |
+
"loss": 0.0282,
|
| 1639 |
+
"step": 5500
|
| 1640 |
+
},
|
| 1641 |
+
{
|
| 1642 |
+
"epoch": 3.0388,
|
| 1643 |
+
"eval_loss": 0.13909843564033508,
|
| 1644 |
+
"eval_runtime": 4135.2147,
|
| 1645 |
+
"eval_samples_per_second": 3.296,
|
| 1646 |
+
"eval_steps_per_second": 0.412,
|
| 1647 |
+
"eval_wer": 7.8181684927992965,
|
| 1648 |
+
"step": 5500
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 3.0413,
|
| 1652 |
+
"grad_norm": 1.0075204372406006,
|
| 1653 |
+
"learning_rate": 2.061315789473684e-06,
|
| 1654 |
+
"loss": 0.0308,
|
| 1655 |
+
"step": 5525
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 3.0438,
|
| 1659 |
+
"grad_norm": 1.0206842422485352,
|
| 1660 |
+
"learning_rate": 2.049802631578947e-06,
|
| 1661 |
+
"loss": 0.0306,
|
| 1662 |
+
"step": 5550
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 3.0463,
|
| 1666 |
+
"grad_norm": 1.411301851272583,
|
| 1667 |
+
"learning_rate": 2.03828947368421e-06,
|
| 1668 |
+
"loss": 0.0243,
|
| 1669 |
+
"step": 5575
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 3.0488,
|
| 1673 |
+
"grad_norm": 0.959862470626831,
|
| 1674 |
+
"learning_rate": 2.0267763157894732e-06,
|
| 1675 |
+
"loss": 0.0272,
|
| 1676 |
+
"step": 5600
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 3.0513,
|
| 1680 |
+
"grad_norm": 2.2999842166900635,
|
| 1681 |
+
"learning_rate": 2.0152631578947367e-06,
|
| 1682 |
+
"loss": 0.0246,
|
| 1683 |
+
"step": 5625
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 3.0538,
|
| 1687 |
+
"grad_norm": 2.890066146850586,
|
| 1688 |
+
"learning_rate": 2.00375e-06,
|
| 1689 |
+
"loss": 0.0299,
|
| 1690 |
+
"step": 5650
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 3.0563,
|
| 1694 |
+
"grad_norm": 1.7101376056671143,
|
| 1695 |
+
"learning_rate": 1.992236842105263e-06,
|
| 1696 |
+
"loss": 0.0322,
|
| 1697 |
+
"step": 5675
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 3.0588,
|
| 1701 |
+
"grad_norm": 1.531943917274475,
|
| 1702 |
+
"learning_rate": 1.980723684210526e-06,
|
| 1703 |
+
"loss": 0.0345,
|
| 1704 |
+
"step": 5700
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 3.0613,
|
| 1708 |
+
"grad_norm": 1.6334413290023804,
|
| 1709 |
+
"learning_rate": 1.969210526315789e-06,
|
| 1710 |
+
"loss": 0.032,
|
| 1711 |
+
"step": 5725
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 3.0638,
|
| 1715 |
+
"grad_norm": 2.112278461456299,
|
| 1716 |
+
"learning_rate": 1.9576973684210526e-06,
|
| 1717 |
+
"loss": 0.0304,
|
| 1718 |
+
"step": 5750
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 3.0663,
|
| 1722 |
+
"grad_norm": 1.7582517862319946,
|
| 1723 |
+
"learning_rate": 1.9461842105263156e-06,
|
| 1724 |
+
"loss": 0.0254,
|
| 1725 |
+
"step": 5775
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 3.0688,
|
| 1729 |
+
"grad_norm": 1.3391777276992798,
|
| 1730 |
+
"learning_rate": 1.934671052631579e-06,
|
| 1731 |
+
"loss": 0.0316,
|
| 1732 |
+
"step": 5800
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 3.0713,
|
| 1736 |
+
"grad_norm": 0.8350562453269958,
|
| 1737 |
+
"learning_rate": 1.923157894736842e-06,
|
| 1738 |
+
"loss": 0.0329,
|
| 1739 |
+
"step": 5825
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 3.0738,
|
| 1743 |
+
"grad_norm": 0.7084619402885437,
|
| 1744 |
+
"learning_rate": 1.911644736842105e-06,
|
| 1745 |
+
"loss": 0.0325,
|
| 1746 |
+
"step": 5850
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 3.0763,
|
| 1750 |
+
"grad_norm": 1.2961277961730957,
|
| 1751 |
+
"learning_rate": 1.9001315789473683e-06,
|
| 1752 |
+
"loss": 0.0313,
|
| 1753 |
+
"step": 5875
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 3.0788,
|
| 1757 |
+
"grad_norm": 1.032840371131897,
|
| 1758 |
+
"learning_rate": 1.8886184210526315e-06,
|
| 1759 |
+
"loss": 0.0224,
|
| 1760 |
+
"step": 5900
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 3.0813,
|
| 1764 |
+
"grad_norm": 1.2073044776916504,
|
| 1765 |
+
"learning_rate": 1.8771052631578945e-06,
|
| 1766 |
+
"loss": 0.0215,
|
| 1767 |
+
"step": 5925
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 3.0838,
|
| 1771 |
+
"grad_norm": 0.8210967779159546,
|
| 1772 |
+
"learning_rate": 1.8655921052631577e-06,
|
| 1773 |
+
"loss": 0.0258,
|
| 1774 |
+
"step": 5950
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 3.0863,
|
| 1778 |
+
"grad_norm": 1.5273653268814087,
|
| 1779 |
+
"learning_rate": 1.854078947368421e-06,
|
| 1780 |
+
"loss": 0.0254,
|
| 1781 |
+
"step": 5975
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 3.0888,
|
| 1785 |
+
"grad_norm": 3.194197177886963,
|
| 1786 |
+
"learning_rate": 1.8425657894736842e-06,
|
| 1787 |
+
"loss": 0.025,
|
| 1788 |
+
"step": 6000
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 3.0888,
|
| 1792 |
+
"eval_loss": 0.14247554540634155,
|
| 1793 |
+
"eval_runtime": 4123.5746,
|
| 1794 |
+
"eval_samples_per_second": 3.305,
|
| 1795 |
+
"eval_steps_per_second": 0.413,
|
| 1796 |
+
"eval_wer": 7.940928579281029,
|
| 1797 |
+
"step": 6000
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 3.0913,
|
| 1801 |
+
"grad_norm": 2.1373400688171387,
|
| 1802 |
+
"learning_rate": 1.8310526315789472e-06,
|
| 1803 |
+
"loss": 0.031,
|
| 1804 |
+
"step": 6025
|
| 1805 |
+
},
|
| 1806 |
+
{
|
| 1807 |
+
"epoch": 3.0938,
|
| 1808 |
+
"grad_norm": 1.0779415369033813,
|
| 1809 |
+
"learning_rate": 1.8195394736842104e-06,
|
| 1810 |
+
"loss": 0.024,
|
| 1811 |
+
"step": 6050
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"epoch": 3.0963,
|
| 1815 |
+
"grad_norm": 0.9637121558189392,
|
| 1816 |
+
"learning_rate": 1.8080263157894734e-06,
|
| 1817 |
+
"loss": 0.0282,
|
| 1818 |
+
"step": 6075
|
| 1819 |
+
},
|
| 1820 |
+
{
|
| 1821 |
+
"epoch": 3.0987999999999998,
|
| 1822 |
+
"grad_norm": 1.1645703315734863,
|
| 1823 |
+
"learning_rate": 1.7965131578947366e-06,
|
| 1824 |
+
"loss": 0.0278,
|
| 1825 |
+
"step": 6100
|
| 1826 |
+
},
|
| 1827 |
+
{
|
| 1828 |
+
"epoch": 3.1013,
|
| 1829 |
+
"grad_norm": 1.2814173698425293,
|
| 1830 |
+
"learning_rate": 1.7849999999999996e-06,
|
| 1831 |
+
"loss": 0.0199,
|
| 1832 |
+
"step": 6125
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 3.1038,
|
| 1836 |
+
"grad_norm": 1.458809494972229,
|
| 1837 |
+
"learning_rate": 1.773486842105263e-06,
|
| 1838 |
+
"loss": 0.0264,
|
| 1839 |
+
"step": 6150
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 3.1063,
|
| 1843 |
+
"grad_norm": 1.6669671535491943,
|
| 1844 |
+
"learning_rate": 1.7619736842105263e-06,
|
| 1845 |
+
"loss": 0.0272,
|
| 1846 |
+
"step": 6175
|
| 1847 |
+
},
|
| 1848 |
+
{
|
| 1849 |
+
"epoch": 3.1088,
|
| 1850 |
+
"grad_norm": 1.5049173831939697,
|
| 1851 |
+
"learning_rate": 1.7504605263157893e-06,
|
| 1852 |
+
"loss": 0.0243,
|
| 1853 |
+
"step": 6200
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 3.1113,
|
| 1857 |
+
"grad_norm": 0.861107587814331,
|
| 1858 |
+
"learning_rate": 1.7389473684210525e-06,
|
| 1859 |
+
"loss": 0.0274,
|
| 1860 |
+
"step": 6225
|
| 1861 |
+
},
|
| 1862 |
+
{
|
| 1863 |
+
"epoch": 3.1138,
|
| 1864 |
+
"grad_norm": 1.0454998016357422,
|
| 1865 |
+
"learning_rate": 1.7274342105263155e-06,
|
| 1866 |
+
"loss": 0.0258,
|
| 1867 |
+
"step": 6250
|
| 1868 |
+
},
|
| 1869 |
+
{
|
| 1870 |
+
"epoch": 3.1163,
|
| 1871 |
+
"grad_norm": 1.7108014822006226,
|
| 1872 |
+
"learning_rate": 1.7159210526315788e-06,
|
| 1873 |
+
"loss": 0.0259,
|
| 1874 |
+
"step": 6275
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 3.1188,
|
| 1878 |
+
"grad_norm": 0.8804712295532227,
|
| 1879 |
+
"learning_rate": 1.704407894736842e-06,
|
| 1880 |
+
"loss": 0.0255,
|
| 1881 |
+
"step": 6300
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 3.1213,
|
| 1885 |
+
"grad_norm": 2.0050883293151855,
|
| 1886 |
+
"learning_rate": 1.6928947368421052e-06,
|
| 1887 |
+
"loss": 0.0304,
|
| 1888 |
+
"step": 6325
|
| 1889 |
+
},
|
| 1890 |
+
{
|
| 1891 |
+
"epoch": 3.1238,
|
| 1892 |
+
"grad_norm": 1.4400875568389893,
|
| 1893 |
+
"learning_rate": 1.6813815789473682e-06,
|
| 1894 |
+
"loss": 0.0333,
|
| 1895 |
+
"step": 6350
|
| 1896 |
+
},
|
| 1897 |
+
{
|
| 1898 |
+
"epoch": 3.1263,
|
| 1899 |
+
"grad_norm": 1.4423948526382446,
|
| 1900 |
+
"learning_rate": 1.6698684210526315e-06,
|
| 1901 |
+
"loss": 0.0279,
|
| 1902 |
+
"step": 6375
|
| 1903 |
+
},
|
| 1904 |
+
{
|
| 1905 |
+
"epoch": 3.1288,
|
| 1906 |
+
"grad_norm": 1.3972327709197998,
|
| 1907 |
+
"learning_rate": 1.6583552631578947e-06,
|
| 1908 |
+
"loss": 0.0255,
|
| 1909 |
+
"step": 6400
|
| 1910 |
+
},
|
| 1911 |
+
{
|
| 1912 |
+
"epoch": 3.1313,
|
| 1913 |
+
"grad_norm": 1.6908966302871704,
|
| 1914 |
+
"learning_rate": 1.6468421052631577e-06,
|
| 1915 |
+
"loss": 0.0267,
|
| 1916 |
+
"step": 6425
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 3.1338,
|
| 1920 |
+
"grad_norm": 0.9540082216262817,
|
| 1921 |
+
"learning_rate": 1.635328947368421e-06,
|
| 1922 |
+
"loss": 0.0265,
|
| 1923 |
+
"step": 6450
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 3.1363,
|
| 1927 |
+
"grad_norm": 1.41488778591156,
|
| 1928 |
+
"learning_rate": 1.6238157894736841e-06,
|
| 1929 |
+
"loss": 0.0224,
|
| 1930 |
+
"step": 6475
|
| 1931 |
+
},
|
| 1932 |
+
{
|
| 1933 |
+
"epoch": 3.1388,
|
| 1934 |
+
"grad_norm": 0.4790860116481781,
|
| 1935 |
+
"learning_rate": 1.6123026315789474e-06,
|
| 1936 |
+
"loss": 0.0274,
|
| 1937 |
+
"step": 6500
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 3.1388,
|
| 1941 |
+
"eval_loss": 0.13914132118225098,
|
| 1942 |
+
"eval_runtime": 4133.8202,
|
| 1943 |
+
"eval_samples_per_second": 3.297,
|
| 1944 |
+
"eval_steps_per_second": 0.412,
|
| 1945 |
+
"eval_wer": 7.731137088204039,
|
| 1946 |
+
"step": 6500
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 3.1413,
|
| 1950 |
+
"grad_norm": 2.5638585090637207,
|
| 1951 |
+
"learning_rate": 1.6007894736842104e-06,
|
| 1952 |
+
"loss": 0.025,
|
| 1953 |
+
"step": 6525
|
| 1954 |
+
},
|
| 1955 |
+
{
|
| 1956 |
+
"epoch": 3.1438,
|
| 1957 |
+
"grad_norm": 1.8847306966781616,
|
| 1958 |
+
"learning_rate": 1.5892763157894736e-06,
|
| 1959 |
+
"loss": 0.0294,
|
| 1960 |
+
"step": 6550
|
| 1961 |
+
},
|
| 1962 |
+
{
|
| 1963 |
+
"epoch": 3.1463,
|
| 1964 |
+
"grad_norm": 1.0196236371994019,
|
| 1965 |
+
"learning_rate": 1.5777631578947366e-06,
|
| 1966 |
+
"loss": 0.0255,
|
| 1967 |
+
"step": 6575
|
| 1968 |
+
},
|
| 1969 |
+
{
|
| 1970 |
+
"epoch": 3.1488,
|
| 1971 |
+
"grad_norm": 1.0703202486038208,
|
| 1972 |
+
"learning_rate": 1.5662499999999998e-06,
|
| 1973 |
+
"loss": 0.0246,
|
| 1974 |
+
"step": 6600
|
| 1975 |
+
},
|
| 1976 |
+
{
|
| 1977 |
+
"epoch": 3.1513,
|
| 1978 |
+
"grad_norm": 2.646519422531128,
|
| 1979 |
+
"learning_rate": 1.5547368421052628e-06,
|
| 1980 |
+
"loss": 0.0213,
|
| 1981 |
+
"step": 6625
|
| 1982 |
+
},
|
| 1983 |
+
{
|
| 1984 |
+
"epoch": 3.1538,
|
| 1985 |
+
"grad_norm": 1.7430530786514282,
|
| 1986 |
+
"learning_rate": 1.5432236842105263e-06,
|
| 1987 |
+
"loss": 0.0267,
|
| 1988 |
+
"step": 6650
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"epoch": 3.1563,
|
| 1992 |
+
"grad_norm": 1.0606240034103394,
|
| 1993 |
+
"learning_rate": 1.5317105263157895e-06,
|
| 1994 |
+
"loss": 0.0269,
|
| 1995 |
+
"step": 6675
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 3.1588,
|
| 1999 |
+
"grad_norm": 1.4670476913452148,
|
| 2000 |
+
"learning_rate": 1.5201973684210525e-06,
|
| 2001 |
+
"loss": 0.0271,
|
| 2002 |
+
"step": 6700
|
| 2003 |
+
},
|
| 2004 |
+
{
|
| 2005 |
+
"epoch": 3.1612999999999998,
|
| 2006 |
+
"grad_norm": 2.345014810562134,
|
| 2007 |
+
"learning_rate": 1.5086842105263157e-06,
|
| 2008 |
+
"loss": 0.0252,
|
| 2009 |
+
"step": 6725
|
| 2010 |
+
},
|
| 2011 |
+
{
|
| 2012 |
+
"epoch": 3.1638,
|
| 2013 |
+
"grad_norm": 2.9098987579345703,
|
| 2014 |
+
"learning_rate": 1.4971710526315787e-06,
|
| 2015 |
+
"loss": 0.0272,
|
| 2016 |
+
"step": 6750
|
| 2017 |
+
},
|
| 2018 |
+
{
|
| 2019 |
+
"epoch": 3.1663,
|
| 2020 |
+
"grad_norm": 0.5682694911956787,
|
| 2021 |
+
"learning_rate": 1.485657894736842e-06,
|
| 2022 |
+
"loss": 0.0237,
|
| 2023 |
+
"step": 6775
|
| 2024 |
+
},
|
| 2025 |
+
{
|
| 2026 |
+
"epoch": 3.1688,
|
| 2027 |
+
"grad_norm": 1.4645904302597046,
|
| 2028 |
+
"learning_rate": 1.4746052631578947e-06,
|
| 2029 |
+
"loss": 0.0303,
|
| 2030 |
+
"step": 6800
|
| 2031 |
+
},
|
| 2032 |
+
{
|
| 2033 |
+
"epoch": 4.0009,
|
| 2034 |
+
"grad_norm": 1.3764489889144897,
|
| 2035 |
+
"learning_rate": 1.4630921052631578e-06,
|
| 2036 |
+
"loss": 0.0242,
|
| 2037 |
+
"step": 6825
|
| 2038 |
+
},
|
| 2039 |
+
{
|
| 2040 |
+
"epoch": 4.0034,
|
| 2041 |
+
"grad_norm": 0.8848748803138733,
|
| 2042 |
+
"learning_rate": 1.451578947368421e-06,
|
| 2043 |
+
"loss": 0.0163,
|
| 2044 |
+
"step": 6850
|
| 2045 |
+
},
|
| 2046 |
+
{
|
| 2047 |
+
"epoch": 4.0059,
|
| 2048 |
+
"grad_norm": 0.619125485420227,
|
| 2049 |
+
"learning_rate": 1.440065789473684e-06,
|
| 2050 |
+
"loss": 0.0188,
|
| 2051 |
+
"step": 6875
|
| 2052 |
+
},
|
| 2053 |
+
{
|
| 2054 |
+
"epoch": 4.0084,
|
| 2055 |
+
"grad_norm": 0.9328649044036865,
|
| 2056 |
+
"learning_rate": 1.4285526315789472e-06,
|
| 2057 |
+
"loss": 0.0173,
|
| 2058 |
+
"step": 6900
|
| 2059 |
+
},
|
| 2060 |
+
{
|
| 2061 |
+
"epoch": 4.0109,
|
| 2062 |
+
"grad_norm": 1.77474045753479,
|
| 2063 |
+
"learning_rate": 1.4170394736842104e-06,
|
| 2064 |
+
"loss": 0.0146,
|
| 2065 |
+
"step": 6925
|
| 2066 |
+
},
|
| 2067 |
+
{
|
| 2068 |
+
"epoch": 4.0134,
|
| 2069 |
+
"grad_norm": 1.3934537172317505,
|
| 2070 |
+
"learning_rate": 1.4055263157894737e-06,
|
| 2071 |
+
"loss": 0.0156,
|
| 2072 |
+
"step": 6950
|
| 2073 |
+
},
|
| 2074 |
+
{
|
| 2075 |
+
"epoch": 4.0159,
|
| 2076 |
+
"grad_norm": 1.2856354713439941,
|
| 2077 |
+
"learning_rate": 1.3940131578947367e-06,
|
| 2078 |
+
"loss": 0.0173,
|
| 2079 |
+
"step": 6975
|
| 2080 |
+
},
|
| 2081 |
+
{
|
| 2082 |
+
"epoch": 4.0184,
|
| 2083 |
+
"grad_norm": 2.1229758262634277,
|
| 2084 |
+
"learning_rate": 1.3824999999999999e-06,
|
| 2085 |
+
"loss": 0.0155,
|
| 2086 |
+
"step": 7000
|
| 2087 |
+
},
|
| 2088 |
+
{
|
| 2089 |
+
"epoch": 4.0184,
|
| 2090 |
+
"eval_loss": 0.14916160702705383,
|
| 2091 |
+
"eval_runtime": 4128.7355,
|
| 2092 |
+
"eval_samples_per_second": 3.301,
|
| 2093 |
+
"eval_steps_per_second": 0.413,
|
| 2094 |
+
"eval_wer": 7.697240646414307,
|
| 2095 |
+
"step": 7000
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 4.0209,
|
| 2099 |
+
"grad_norm": 0.44512999057769775,
|
| 2100 |
+
"learning_rate": 1.3709868421052631e-06,
|
| 2101 |
+
"loss": 0.0153,
|
| 2102 |
+
"step": 7025
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 4.0234,
|
| 2106 |
+
"grad_norm": 1.8791674375534058,
|
| 2107 |
+
"learning_rate": 1.3594736842105261e-06,
|
| 2108 |
+
"loss": 0.0165,
|
| 2109 |
+
"step": 7050
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 4.0259,
|
| 2113 |
+
"grad_norm": 5.244405746459961,
|
| 2114 |
+
"learning_rate": 1.3479605263157894e-06,
|
| 2115 |
+
"loss": 0.0179,
|
| 2116 |
+
"step": 7075
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 4.0284,
|
| 2120 |
+
"grad_norm": 1.1926153898239136,
|
| 2121 |
+
"learning_rate": 1.3364473684210526e-06,
|
| 2122 |
+
"loss": 0.0161,
|
| 2123 |
+
"step": 7100
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 4.0309,
|
| 2127 |
+
"grad_norm": 1.1147819757461548,
|
| 2128 |
+
"learning_rate": 1.3249342105263158e-06,
|
| 2129 |
+
"loss": 0.015,
|
| 2130 |
+
"step": 7125
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 4.0334,
|
| 2134 |
+
"grad_norm": 1.9370721578598022,
|
| 2135 |
+
"learning_rate": 1.3134210526315788e-06,
|
| 2136 |
+
"loss": 0.0142,
|
| 2137 |
+
"step": 7150
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 4.0359,
|
| 2141 |
+
"grad_norm": 0.49344903230667114,
|
| 2142 |
+
"learning_rate": 1.301907894736842e-06,
|
| 2143 |
+
"loss": 0.0134,
|
| 2144 |
+
"step": 7175
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 4.0384,
|
| 2148 |
+
"grad_norm": 1.8190902471542358,
|
| 2149 |
+
"learning_rate": 1.290394736842105e-06,
|
| 2150 |
+
"loss": 0.0168,
|
| 2151 |
+
"step": 7200
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 4.0409,
|
| 2155 |
+
"grad_norm": 0.7560425400733948,
|
| 2156 |
+
"learning_rate": 1.2788815789473683e-06,
|
| 2157 |
+
"loss": 0.0143,
|
| 2158 |
+
"step": 7225
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 4.0434,
|
| 2162 |
+
"grad_norm": 1.0451087951660156,
|
| 2163 |
+
"learning_rate": 1.2673684210526313e-06,
|
| 2164 |
+
"loss": 0.0149,
|
| 2165 |
+
"step": 7250
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 4.0459,
|
| 2169 |
+
"grad_norm": 1.0334726572036743,
|
| 2170 |
+
"learning_rate": 1.2558552631578947e-06,
|
| 2171 |
+
"loss": 0.0136,
|
| 2172 |
+
"step": 7275
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 4.0484,
|
| 2176 |
+
"grad_norm": 0.6531663537025452,
|
| 2177 |
+
"learning_rate": 1.244342105263158e-06,
|
| 2178 |
+
"loss": 0.0137,
|
| 2179 |
+
"step": 7300
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 4.0509,
|
| 2183 |
+
"grad_norm": 0.8954887986183167,
|
| 2184 |
+
"learning_rate": 1.232828947368421e-06,
|
| 2185 |
+
"loss": 0.0118,
|
| 2186 |
+
"step": 7325
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 4.0534,
|
| 2190 |
+
"grad_norm": 1.0640511512756348,
|
| 2191 |
+
"learning_rate": 1.2213157894736842e-06,
|
| 2192 |
+
"loss": 0.0126,
|
| 2193 |
+
"step": 7350
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 4.0559,
|
| 2197 |
+
"grad_norm": 0.2824617922306061,
|
| 2198 |
+
"learning_rate": 1.2098026315789472e-06,
|
| 2199 |
+
"loss": 0.0139,
|
| 2200 |
+
"step": 7375
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 4.0584,
|
| 2204 |
+
"grad_norm": 1.0095443725585938,
|
| 2205 |
+
"learning_rate": 1.1982894736842104e-06,
|
| 2206 |
+
"loss": 0.018,
|
| 2207 |
+
"step": 7400
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 4.0609,
|
| 2211 |
+
"grad_norm": 1.1475225687026978,
|
| 2212 |
+
"learning_rate": 1.1867763157894734e-06,
|
| 2213 |
+
"loss": 0.0133,
|
| 2214 |
+
"step": 7425
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 4.0634,
|
| 2218 |
+
"grad_norm": 1.5951991081237793,
|
| 2219 |
+
"learning_rate": 1.1752631578947369e-06,
|
| 2220 |
+
"loss": 0.013,
|
| 2221 |
+
"step": 7450
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 4.0659,
|
| 2225 |
+
"grad_norm": 0.3482917249202728,
|
| 2226 |
+
"learning_rate": 1.1637499999999999e-06,
|
| 2227 |
+
"loss": 0.0154,
|
| 2228 |
+
"step": 7475
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 4.0684,
|
| 2232 |
+
"grad_norm": 1.1572391986846924,
|
| 2233 |
+
"learning_rate": 1.152236842105263e-06,
|
| 2234 |
+
"loss": 0.0189,
|
| 2235 |
+
"step": 7500
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 4.0684,
|
| 2239 |
+
"eval_loss": 0.15172211825847626,
|
| 2240 |
+
"eval_runtime": 4117.5679,
|
| 2241 |
+
"eval_samples_per_second": 3.31,
|
| 2242 |
+
"eval_steps_per_second": 0.414,
|
| 2243 |
+
"eval_wer": 7.656931364285977,
|
| 2244 |
+
"step": 7500
|
| 2245 |
+
},
|
| 2246 |
+
{
|
| 2247 |
+
"epoch": 4.0709,
|
| 2248 |
+
"grad_norm": 1.3942557573318481,
|
| 2249 |
+
"learning_rate": 1.140723684210526e-06,
|
| 2250 |
+
"loss": 0.0143,
|
| 2251 |
+
"step": 7525
|
| 2252 |
+
},
|
| 2253 |
+
{
|
| 2254 |
+
"epoch": 4.0734,
|
| 2255 |
+
"grad_norm": 0.8097572326660156,
|
| 2256 |
+
"learning_rate": 1.1292105263157893e-06,
|
| 2257 |
+
"loss": 0.0127,
|
| 2258 |
+
"step": 7550
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 4.0759,
|
| 2262 |
+
"grad_norm": 0.740375816822052,
|
| 2263 |
+
"learning_rate": 1.1176973684210526e-06,
|
| 2264 |
+
"loss": 0.0124,
|
| 2265 |
+
"step": 7575
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 4.0784,
|
| 2269 |
+
"grad_norm": 0.8702480792999268,
|
| 2270 |
+
"learning_rate": 1.1061842105263156e-06,
|
| 2271 |
+
"loss": 0.0137,
|
| 2272 |
+
"step": 7600
|
| 2273 |
+
},
|
| 2274 |
+
{
|
| 2275 |
+
"epoch": 4.0809,
|
| 2276 |
+
"grad_norm": 1.223105788230896,
|
| 2277 |
+
"learning_rate": 1.094671052631579e-06,
|
| 2278 |
+
"loss": 0.0137,
|
| 2279 |
+
"step": 7625
|
| 2280 |
+
},
|
| 2281 |
+
{
|
| 2282 |
+
"epoch": 4.0834,
|
| 2283 |
+
"grad_norm": 0.43614983558654785,
|
| 2284 |
+
"learning_rate": 1.083157894736842e-06,
|
| 2285 |
+
"loss": 0.0109,
|
| 2286 |
+
"step": 7650
|
| 2287 |
+
},
|
| 2288 |
+
{
|
| 2289 |
+
"epoch": 4.0859,
|
| 2290 |
+
"grad_norm": 1.0974986553192139,
|
| 2291 |
+
"learning_rate": 1.0716447368421052e-06,
|
| 2292 |
+
"loss": 0.0118,
|
| 2293 |
+
"step": 7675
|
| 2294 |
+
},
|
| 2295 |
+
{
|
| 2296 |
+
"epoch": 4.0884,
|
| 2297 |
+
"grad_norm": 0.7234652042388916,
|
| 2298 |
+
"learning_rate": 1.0601315789473682e-06,
|
| 2299 |
+
"loss": 0.0125,
|
| 2300 |
+
"step": 7700
|
| 2301 |
+
},
|
| 2302 |
+
{
|
| 2303 |
+
"epoch": 4.0909,
|
| 2304 |
+
"grad_norm": 0.7752431035041809,
|
| 2305 |
+
"learning_rate": 1.0486184210526315e-06,
|
| 2306 |
+
"loss": 0.0135,
|
| 2307 |
+
"step": 7725
|
| 2308 |
+
},
|
| 2309 |
+
{
|
| 2310 |
+
"epoch": 4.0934,
|
| 2311 |
+
"grad_norm": 0.8796952366828918,
|
| 2312 |
+
"learning_rate": 1.0371052631578947e-06,
|
| 2313 |
+
"loss": 0.0158,
|
| 2314 |
+
"step": 7750
|
| 2315 |
+
},
|
| 2316 |
+
{
|
| 2317 |
+
"epoch": 4.0959,
|
| 2318 |
+
"grad_norm": 3.9135661125183105,
|
| 2319 |
+
"learning_rate": 1.0255921052631577e-06,
|
| 2320 |
+
"loss": 0.0139,
|
| 2321 |
+
"step": 7775
|
| 2322 |
+
},
|
| 2323 |
+
{
|
| 2324 |
+
"epoch": 4.0984,
|
| 2325 |
+
"grad_norm": 0.4837290942668915,
|
| 2326 |
+
"learning_rate": 1.014078947368421e-06,
|
| 2327 |
+
"loss": 0.0103,
|
| 2328 |
+
"step": 7800
|
| 2329 |
+
},
|
| 2330 |
+
{
|
| 2331 |
+
"epoch": 4.1009,
|
| 2332 |
+
"grad_norm": 1.1155998706817627,
|
| 2333 |
+
"learning_rate": 1.0025657894736842e-06,
|
| 2334 |
+
"loss": 0.0106,
|
| 2335 |
+
"step": 7825
|
| 2336 |
+
},
|
| 2337 |
+
{
|
| 2338 |
+
"epoch": 4.1034,
|
| 2339 |
+
"grad_norm": 2.628676652908325,
|
| 2340 |
+
"learning_rate": 9.910526315789474e-07,
|
| 2341 |
+
"loss": 0.0089,
|
| 2342 |
+
"step": 7850
|
| 2343 |
+
},
|
| 2344 |
+
{
|
| 2345 |
+
"epoch": 4.1059,
|
| 2346 |
+
"grad_norm": 1.716665506362915,
|
| 2347 |
+
"learning_rate": 9.795394736842104e-07,
|
| 2348 |
+
"loss": 0.0132,
|
| 2349 |
+
"step": 7875
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 4.1084,
|
| 2353 |
+
"grad_norm": 1.6751716136932373,
|
| 2354 |
+
"learning_rate": 9.680263157894736e-07,
|
| 2355 |
+
"loss": 0.0137,
|
| 2356 |
+
"step": 7900
|
| 2357 |
+
},
|
| 2358 |
+
{
|
| 2359 |
+
"epoch": 4.1109,
|
| 2360 |
+
"grad_norm": 0.9773244261741638,
|
| 2361 |
+
"learning_rate": 9.565131578947368e-07,
|
| 2362 |
+
"loss": 0.0111,
|
| 2363 |
+
"step": 7925
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 4.1134,
|
| 2367 |
+
"grad_norm": 1.44219172000885,
|
| 2368 |
+
"learning_rate": 9.45e-07,
|
| 2369 |
+
"loss": 0.0139,
|
| 2370 |
+
"step": 7950
|
| 2371 |
+
},
|
| 2372 |
+
{
|
| 2373 |
+
"epoch": 4.1159,
|
| 2374 |
+
"grad_norm": 0.8723123073577881,
|
| 2375 |
+
"learning_rate": 9.334868421052631e-07,
|
| 2376 |
+
"loss": 0.0117,
|
| 2377 |
+
"step": 7975
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 4.1184,
|
| 2381 |
+
"grad_norm": 0.6484673023223877,
|
| 2382 |
+
"learning_rate": 9.219736842105263e-07,
|
| 2383 |
+
"loss": 0.0139,
|
| 2384 |
+
"step": 8000
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 4.1184,
|
| 2388 |
+
"eval_loss": 0.15393850207328796,
|
| 2389 |
+
"eval_runtime": 4128.9341,
|
| 2390 |
+
"eval_samples_per_second": 3.301,
|
| 2391 |
+
"eval_steps_per_second": 0.413,
|
| 2392 |
+
"eval_wer": 7.626699402689728,
|
| 2393 |
+
"step": 8000
|
| 2394 |
+
},
|
| 2395 |
+
{
|
| 2396 |
+
"epoch": 4.1209,
|
| 2397 |
+
"grad_norm": 1.3702197074890137,
|
| 2398 |
+
"learning_rate": 9.104605263157894e-07,
|
| 2399 |
+
"loss": 0.0158,
|
| 2400 |
+
"step": 8025
|
| 2401 |
+
},
|
| 2402 |
+
{
|
| 2403 |
+
"epoch": 4.1234,
|
| 2404 |
+
"grad_norm": 1.425645351409912,
|
| 2405 |
+
"learning_rate": 8.989473684210525e-07,
|
| 2406 |
+
"loss": 0.0117,
|
| 2407 |
+
"step": 8050
|
| 2408 |
+
},
|
| 2409 |
+
{
|
| 2410 |
+
"epoch": 4.1259,
|
| 2411 |
+
"grad_norm": 1.4255399703979492,
|
| 2412 |
+
"learning_rate": 8.874342105263158e-07,
|
| 2413 |
+
"loss": 0.015,
|
| 2414 |
+
"step": 8075
|
| 2415 |
+
},
|
| 2416 |
+
{
|
| 2417 |
+
"epoch": 4.1284,
|
| 2418 |
+
"grad_norm": 0.6988621950149536,
|
| 2419 |
+
"learning_rate": 8.759210526315789e-07,
|
| 2420 |
+
"loss": 0.0141,
|
| 2421 |
+
"step": 8100
|
| 2422 |
+
},
|
| 2423 |
+
{
|
| 2424 |
+
"epoch": 4.1309,
|
| 2425 |
+
"grad_norm": 1.1563546657562256,
|
| 2426 |
+
"learning_rate": 8.64407894736842e-07,
|
| 2427 |
+
"loss": 0.0122,
|
| 2428 |
+
"step": 8125
|
| 2429 |
+
},
|
| 2430 |
+
{
|
| 2431 |
+
"epoch": 4.1334,
|
| 2432 |
+
"grad_norm": 1.2023714780807495,
|
| 2433 |
+
"learning_rate": 8.528947368421051e-07,
|
| 2434 |
+
"loss": 0.013,
|
| 2435 |
+
"step": 8150
|
| 2436 |
+
},
|
| 2437 |
+
{
|
| 2438 |
+
"epoch": 4.1359,
|
| 2439 |
+
"grad_norm": 0.9450110197067261,
|
| 2440 |
+
"learning_rate": 8.413815789473683e-07,
|
| 2441 |
+
"loss": 0.0123,
|
| 2442 |
+
"step": 8175
|
| 2443 |
+
},
|
| 2444 |
+
{
|
| 2445 |
+
"epoch": 4.1384,
|
| 2446 |
+
"grad_norm": 0.9265995621681213,
|
| 2447 |
+
"learning_rate": 8.298684210526316e-07,
|
| 2448 |
+
"loss": 0.0114,
|
| 2449 |
+
"step": 8200
|
| 2450 |
+
},
|
| 2451 |
+
{
|
| 2452 |
+
"epoch": 4.1409,
|
| 2453 |
+
"grad_norm": 0.4234980046749115,
|
| 2454 |
+
"learning_rate": 8.183552631578947e-07,
|
| 2455 |
+
"loss": 0.0085,
|
| 2456 |
+
"step": 8225
|
| 2457 |
+
},
|
| 2458 |
+
{
|
| 2459 |
+
"epoch": 4.1434,
|
| 2460 |
+
"grad_norm": 1.3323073387145996,
|
| 2461 |
+
"learning_rate": 8.068421052631579e-07,
|
| 2462 |
+
"loss": 0.014,
|
| 2463 |
+
"step": 8250
|
| 2464 |
+
},
|
| 2465 |
+
{
|
| 2466 |
+
"epoch": 4.1459,
|
| 2467 |
+
"grad_norm": 1.2050007581710815,
|
| 2468 |
+
"learning_rate": 7.95328947368421e-07,
|
| 2469 |
+
"loss": 0.0106,
|
| 2470 |
+
"step": 8275
|
| 2471 |
+
},
|
| 2472 |
+
{
|
| 2473 |
+
"epoch": 4.1484,
|
| 2474 |
+
"grad_norm": 1.261042594909668,
|
| 2475 |
+
"learning_rate": 7.838157894736841e-07,
|
| 2476 |
+
"loss": 0.0107,
|
| 2477 |
+
"step": 8300
|
| 2478 |
+
},
|
| 2479 |
+
{
|
| 2480 |
+
"epoch": 4.1509,
|
| 2481 |
+
"grad_norm": 1.2892303466796875,
|
| 2482 |
+
"learning_rate": 7.723026315789474e-07,
|
| 2483 |
+
"loss": 0.0145,
|
| 2484 |
+
"step": 8325
|
| 2485 |
+
},
|
| 2486 |
+
{
|
| 2487 |
+
"epoch": 4.1534,
|
| 2488 |
+
"grad_norm": 1.1626112461090088,
|
| 2489 |
+
"learning_rate": 7.607894736842105e-07,
|
| 2490 |
+
"loss": 0.0139,
|
| 2491 |
+
"step": 8350
|
| 2492 |
+
},
|
| 2493 |
+
{
|
| 2494 |
+
"epoch": 4.1559,
|
| 2495 |
+
"grad_norm": 1.0547322034835815,
|
| 2496 |
+
"learning_rate": 7.492763157894736e-07,
|
| 2497 |
+
"loss": 0.0154,
|
| 2498 |
+
"step": 8375
|
| 2499 |
+
},
|
| 2500 |
+
{
|
| 2501 |
+
"epoch": 4.1584,
|
| 2502 |
+
"grad_norm": 0.44805532693862915,
|
| 2503 |
+
"learning_rate": 7.377631578947367e-07,
|
| 2504 |
+
"loss": 0.0109,
|
| 2505 |
+
"step": 8400
|
| 2506 |
+
},
|
| 2507 |
+
{
|
| 2508 |
+
"epoch": 4.1609,
|
| 2509 |
+
"grad_norm": 0.7095866203308105,
|
| 2510 |
+
"learning_rate": 7.262499999999999e-07,
|
| 2511 |
+
"loss": 0.0114,
|
| 2512 |
+
"step": 8425
|
| 2513 |
+
},
|
| 2514 |
+
{
|
| 2515 |
+
"epoch": 4.1634,
|
| 2516 |
+
"grad_norm": 1.4220194816589355,
|
| 2517 |
+
"learning_rate": 7.14736842105263e-07,
|
| 2518 |
+
"loss": 0.0134,
|
| 2519 |
+
"step": 8450
|
| 2520 |
+
},
|
| 2521 |
+
{
|
| 2522 |
+
"epoch": 4.1659,
|
| 2523 |
+
"grad_norm": 1.0814168453216553,
|
| 2524 |
+
"learning_rate": 7.032236842105263e-07,
|
| 2525 |
+
"loss": 0.0142,
|
| 2526 |
+
"step": 8475
|
| 2527 |
+
},
|
| 2528 |
+
{
|
| 2529 |
+
"epoch": 4.1684,
|
| 2530 |
+
"grad_norm": 0.7026916146278381,
|
| 2531 |
+
"learning_rate": 6.917105263157895e-07,
|
| 2532 |
+
"loss": 0.0141,
|
| 2533 |
+
"step": 8500
|
| 2534 |
+
},
|
| 2535 |
+
{
|
| 2536 |
+
"epoch": 4.1684,
|
| 2537 |
+
"eval_loss": 0.15496784448623657,
|
| 2538 |
+
"eval_runtime": 4124.1829,
|
| 2539 |
+
"eval_samples_per_second": 3.305,
|
| 2540 |
+
"eval_steps_per_second": 0.413,
|
| 2541 |
+
"eval_wer": 7.542416358239584,
|
| 2542 |
+
"step": 8500
|
| 2543 |
+
},
|
| 2544 |
+
{
|
| 2545 |
+
"epoch": 5.0005,
|
| 2546 |
+
"grad_norm": 4.648550033569336,
|
| 2547 |
+
"learning_rate": 6.801973684210526e-07,
|
| 2548 |
+
"loss": 0.0285,
|
| 2549 |
+
"step": 8525
|
| 2550 |
+
},
|
| 2551 |
+
{
|
| 2552 |
+
"epoch": 5.003,
|
| 2553 |
+
"grad_norm": 1.9204503297805786,
|
| 2554 |
+
"learning_rate": 6.691447368421053e-07,
|
| 2555 |
+
"loss": 0.0761,
|
| 2556 |
+
"step": 8550
|
| 2557 |
+
},
|
| 2558 |
+
{
|
| 2559 |
+
"epoch": 5.0055,
|
| 2560 |
+
"grad_norm": 1.7285746335983276,
|
| 2561 |
+
"learning_rate": 6.576315789473684e-07,
|
| 2562 |
+
"loss": 0.0602,
|
| 2563 |
+
"step": 8575
|
| 2564 |
+
},
|
| 2565 |
+
{
|
| 2566 |
+
"epoch": 5.008,
|
| 2567 |
+
"grad_norm": 1.1516830921173096,
|
| 2568 |
+
"learning_rate": 6.461184210526315e-07,
|
| 2569 |
+
"loss": 0.0585,
|
| 2570 |
+
"step": 8600
|
| 2571 |
+
},
|
| 2572 |
+
{
|
| 2573 |
+
"epoch": 5.0105,
|
| 2574 |
+
"grad_norm": 3.3867828845977783,
|
| 2575 |
+
"learning_rate": 6.346052631578947e-07,
|
| 2576 |
+
"loss": 0.0656,
|
| 2577 |
+
"step": 8625
|
| 2578 |
+
},
|
| 2579 |
+
{
|
| 2580 |
+
"epoch": 5.013,
|
| 2581 |
+
"grad_norm": 4.064920902252197,
|
| 2582 |
+
"learning_rate": 6.230921052631579e-07,
|
| 2583 |
+
"loss": 0.0683,
|
| 2584 |
+
"step": 8650
|
| 2585 |
+
},
|
| 2586 |
+
{
|
| 2587 |
+
"epoch": 5.0155,
|
| 2588 |
+
"grad_norm": 3.695047378540039,
|
| 2589 |
+
"learning_rate": 6.11578947368421e-07,
|
| 2590 |
+
"loss": 0.0659,
|
| 2591 |
+
"step": 8675
|
| 2592 |
+
},
|
| 2593 |
+
{
|
| 2594 |
+
"epoch": 5.018,
|
| 2595 |
+
"grad_norm": 2.9087939262390137,
|
| 2596 |
+
"learning_rate": 6.000657894736842e-07,
|
| 2597 |
+
"loss": 0.0611,
|
| 2598 |
+
"step": 8700
|
| 2599 |
+
},
|
| 2600 |
+
{
|
| 2601 |
+
"epoch": 5.0205,
|
| 2602 |
+
"grad_norm": 3.368290424346924,
|
| 2603 |
+
"learning_rate": 5.885526315789473e-07,
|
| 2604 |
+
"loss": 0.0603,
|
| 2605 |
+
"step": 8725
|
| 2606 |
+
},
|
| 2607 |
+
{
|
| 2608 |
+
"epoch": 5.023,
|
| 2609 |
+
"grad_norm": 3.7565319538116455,
|
| 2610 |
+
"learning_rate": 5.770394736842104e-07,
|
| 2611 |
+
"loss": 0.0614,
|
| 2612 |
+
"step": 8750
|
| 2613 |
+
},
|
| 2614 |
+
{
|
| 2615 |
+
"epoch": 5.0255,
|
| 2616 |
+
"grad_norm": 2.4887771606445312,
|
| 2617 |
+
"learning_rate": 5.655263157894735e-07,
|
| 2618 |
+
"loss": 0.0497,
|
| 2619 |
+
"step": 8775
|
| 2620 |
+
},
|
| 2621 |
+
{
|
| 2622 |
+
"epoch": 5.028,
|
| 2623 |
+
"grad_norm": 2.1670076847076416,
|
| 2624 |
+
"learning_rate": 5.540131578947369e-07,
|
| 2625 |
+
"loss": 0.0662,
|
| 2626 |
+
"step": 8800
|
| 2627 |
+
},
|
| 2628 |
+
{
|
| 2629 |
+
"epoch": 5.0305,
|
| 2630 |
+
"grad_norm": 1.3746148347854614,
|
| 2631 |
+
"learning_rate": 5.425e-07,
|
| 2632 |
+
"loss": 0.0507,
|
| 2633 |
+
"step": 8825
|
| 2634 |
+
},
|
| 2635 |
+
{
|
| 2636 |
+
"epoch": 5.033,
|
| 2637 |
+
"grad_norm": 1.8274154663085938,
|
| 2638 |
+
"learning_rate": 5.309868421052631e-07,
|
| 2639 |
+
"loss": 0.0449,
|
| 2640 |
+
"step": 8850
|
| 2641 |
+
},
|
| 2642 |
+
{
|
| 2643 |
+
"epoch": 5.0355,
|
| 2644 |
+
"grad_norm": 2.9424078464508057,
|
| 2645 |
+
"learning_rate": 5.194736842105262e-07,
|
| 2646 |
+
"loss": 0.0529,
|
| 2647 |
+
"step": 8875
|
| 2648 |
+
},
|
| 2649 |
+
{
|
| 2650 |
+
"epoch": 5.038,
|
| 2651 |
+
"grad_norm": 2.457754611968994,
|
| 2652 |
+
"learning_rate": 5.079605263157895e-07,
|
| 2653 |
+
"loss": 0.042,
|
| 2654 |
+
"step": 8900
|
| 2655 |
+
},
|
| 2656 |
+
{
|
| 2657 |
+
"epoch": 5.0405,
|
| 2658 |
+
"grad_norm": 2.208768606185913,
|
| 2659 |
+
"learning_rate": 4.964473684210526e-07,
|
| 2660 |
+
"loss": 0.0407,
|
| 2661 |
+
"step": 8925
|
| 2662 |
+
},
|
| 2663 |
+
{
|
| 2664 |
+
"epoch": 5.043,
|
| 2665 |
+
"grad_norm": 1.9554438591003418,
|
| 2666 |
+
"learning_rate": 4.849342105263158e-07,
|
| 2667 |
+
"loss": 0.0465,
|
| 2668 |
+
"step": 8950
|
| 2669 |
+
},
|
| 2670 |
+
{
|
| 2671 |
+
"epoch": 5.0455,
|
| 2672 |
+
"grad_norm": 1.1464567184448242,
|
| 2673 |
+
"learning_rate": 4.734210526315789e-07,
|
| 2674 |
+
"loss": 0.0537,
|
| 2675 |
+
"step": 8975
|
| 2676 |
+
},
|
| 2677 |
+
{
|
| 2678 |
+
"epoch": 5.048,
|
| 2679 |
+
"grad_norm": 3.1216509342193604,
|
| 2680 |
+
"learning_rate": 4.6190789473684203e-07,
|
| 2681 |
+
"loss": 0.0368,
|
| 2682 |
+
"step": 9000
|
| 2683 |
+
},
|
| 2684 |
+
{
|
| 2685 |
+
"epoch": 5.048,
|
| 2686 |
+
"eval_loss": 0.12588092684745789,
|
| 2687 |
+
"eval_runtime": 4149.257,
|
| 2688 |
+
"eval_samples_per_second": 3.285,
|
| 2689 |
+
"eval_steps_per_second": 0.411,
|
| 2690 |
+
"eval_wer": 7.215361500971087,
|
| 2691 |
+
"step": 9000
|
| 2692 |
+
},
|
| 2693 |
+
{
|
| 2694 |
+
"epoch": 5.048,
|
| 2695 |
+
"step": 9000,
|
| 2696 |
+
"total_flos": 4.891718061785088e+20,
|
| 2697 |
+
"train_loss": 0.0,
|
| 2698 |
+
"train_runtime": 289.8068,
|
| 2699 |
+
"train_samples_per_second": 552.092,
|
| 2700 |
+
"train_steps_per_second": 34.506
|
| 2701 |
+
}
|
| 2702 |
+
],
|
| 2703 |
+
"logging_steps": 25,
|
| 2704 |
+
"max_steps": 10000,
|
| 2705 |
+
"num_input_tokens_seen": 0,
|
| 2706 |
+
"num_train_epochs": 9223372036854775807,
|
| 2707 |
+
"save_steps": 1000,
|
| 2708 |
+
"stateful_callbacks": {
|
| 2709 |
+
"TrainerControl": {
|
| 2710 |
+
"args": {
|
| 2711 |
+
"should_epoch_stop": false,
|
| 2712 |
+
"should_evaluate": false,
|
| 2713 |
+
"should_log": false,
|
| 2714 |
+
"should_save": true,
|
| 2715 |
+
"should_training_stop": false
|
| 2716 |
+
},
|
| 2717 |
+
"attributes": {}
|
| 2718 |
+
}
|
| 2719 |
+
},
|
| 2720 |
+
"total_flos": 4.891718061785088e+20,
|
| 2721 |
+
"train_batch_size": 16,
|
| 2722 |
+
"trial_name": null,
|
| 2723 |
+
"trial_params": null
|
| 2724 |
+
}
|
wandb/run-20241007_131849-0rbzerob/files/output.log
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
wandb/run-20241007_131849-0rbzerob/run-0rbzerob.wandb
CHANGED
|
Binary files a/wandb/run-20241007_131849-0rbzerob/run-0rbzerob.wandb and b/wandb/run-20241007_131849-0rbzerob/run-0rbzerob.wandb differ
|
|
|