End of training
Browse files- README.md +9 -4
- all_results.json +15 -0
- runs/May01_23-52-56_dgx-a100/events.out.tfevents.1746160531.dgx-a100.2099488.1 +3 -0
- train_results.json +9 -0
- trainer_state.json +923 -0
- val_results.json +9 -0
README.md
CHANGED
|
@@ -1,22 +1,27 @@
|
|
| 1 |
---
|
| 2 |
library_name: transformers
|
|
|
|
|
|
|
| 3 |
license: apache-2.0
|
| 4 |
-
base_model: openai/whisper-
|
| 5 |
tags:
|
|
|
|
|
|
|
|
|
|
| 6 |
- generated_from_trainer
|
| 7 |
metrics:
|
| 8 |
- wer
|
| 9 |
model-index:
|
| 10 |
-
- name:
|
| 11 |
results: []
|
| 12 |
---
|
| 13 |
|
| 14 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 15 |
should probably proofread and complete it, then remove this comment. -->
|
| 16 |
|
| 17 |
-
#
|
| 18 |
|
| 19 |
-
This model is a fine-tuned version of [openai/whisper-
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
- Loss: 0.1112
|
| 22 |
- Wer: 0.0612
|
|
|
|
| 1 |
---
|
| 2 |
library_name: transformers
|
| 3 |
+
language:
|
| 4 |
+
- jv
|
| 5 |
license: apache-2.0
|
| 6 |
+
base_model: openai/whisper-tiny
|
| 7 |
tags:
|
| 8 |
+
- whisper
|
| 9 |
+
- javanese
|
| 10 |
+
- asr
|
| 11 |
- generated_from_trainer
|
| 12 |
metrics:
|
| 13 |
- wer
|
| 14 |
model-index:
|
| 15 |
+
- name: Whisper-Large-v2-Java-v2
|
| 16 |
results: []
|
| 17 |
---
|
| 18 |
|
| 19 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 20 |
should probably proofread and complete it, then remove this comment. -->
|
| 21 |
|
| 22 |
+
# Whisper-Large-v2-Java-v2
|
| 23 |
|
| 24 |
+
This model is a fine-tuned version of [openai/whisper-tiny](https://huggingface.co/openai/whisper-tiny) on the None dataset.
|
| 25 |
It achieves the following results on the evaluation set:
|
| 26 |
- Loss: 0.1112
|
| 27 |
- Wer: 0.0612
|
all_results.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 8.0,
|
| 3 |
+
"total_flos": 1.698366962958336e+20,
|
| 4 |
+
"train_loss": 0.16200368287563324,
|
| 5 |
+
"train_runtime": 40923.5162,
|
| 6 |
+
"train_samples": 2000,
|
| 7 |
+
"train_samples_per_second": 1.955,
|
| 8 |
+
"train_steps_per_second": 0.489,
|
| 9 |
+
"val_loss": 0.11124598234891891,
|
| 10 |
+
"val_runtime": 893.4451,
|
| 11 |
+
"val_samples": 250,
|
| 12 |
+
"val_samples_per_second": 1.399,
|
| 13 |
+
"val_steps_per_second": 0.7,
|
| 14 |
+
"val_wer": 0.06115037298782882
|
| 15 |
+
}
|
runs/May01_23-52-56_dgx-a100/events.out.tfevents.1746160531.dgx-a100.2099488.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:136ba24cefbc3d727301c4814fb7406377ba0010934aa9fcb61641e5b0629334
|
| 3 |
+
size 437
|
train_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 8.0,
|
| 3 |
+
"total_flos": 1.698366962958336e+20,
|
| 4 |
+
"train_loss": 0.16200368287563324,
|
| 5 |
+
"train_runtime": 40923.5162,
|
| 6 |
+
"train_samples": 2000,
|
| 7 |
+
"train_samples_per_second": 1.955,
|
| 8 |
+
"train_steps_per_second": 0.489
|
| 9 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,923 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 20000,
|
| 3 |
+
"best_metric": 0.06115037298782882,
|
| 4 |
+
"best_model_checkpoint": "/home/cluster-dgxa100/slp01/bagas-fine-tune-whisper/whisper-large-v2-javanese-openslr-v2/checkpoint-20000",
|
| 5 |
+
"epoch": 8.0,
|
| 6 |
+
"eval_steps": 1000,
|
| 7 |
+
"global_step": 20000,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.08,
|
| 14 |
+
"grad_norm": 19.06084442138672,
|
| 15 |
+
"learning_rate": 1.97e-06,
|
| 16 |
+
"loss": 1.8907,
|
| 17 |
+
"step": 200
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.16,
|
| 21 |
+
"grad_norm": 16.973052978515625,
|
| 22 |
+
"learning_rate": 3.97e-06,
|
| 23 |
+
"loss": 0.6026,
|
| 24 |
+
"step": 400
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.24,
|
| 28 |
+
"grad_norm": 8.261700630187988,
|
| 29 |
+
"learning_rate": 5.9700000000000004e-06,
|
| 30 |
+
"loss": 0.5317,
|
| 31 |
+
"step": 600
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.32,
|
| 35 |
+
"grad_norm": 7.87921667098999,
|
| 36 |
+
"learning_rate": 7.970000000000002e-06,
|
| 37 |
+
"loss": 0.491,
|
| 38 |
+
"step": 800
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.4,
|
| 42 |
+
"grad_norm": 7.55085563659668,
|
| 43 |
+
"learning_rate": 9.970000000000001e-06,
|
| 44 |
+
"loss": 0.4821,
|
| 45 |
+
"step": 1000
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.4,
|
| 49 |
+
"eval_loss": 0.3569938838481903,
|
| 50 |
+
"eval_runtime": 880.4774,
|
| 51 |
+
"eval_samples_per_second": 1.42,
|
| 52 |
+
"eval_steps_per_second": 0.71,
|
| 53 |
+
"eval_wer": 0.22683549273655282,
|
| 54 |
+
"step": 1000
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.48,
|
| 58 |
+
"grad_norm": 5.795778751373291,
|
| 59 |
+
"learning_rate": 1.1970000000000002e-05,
|
| 60 |
+
"loss": 0.4608,
|
| 61 |
+
"step": 1200
|
| 62 |
+
},
|
| 63 |
+
{
|
| 64 |
+
"epoch": 0.56,
|
| 65 |
+
"grad_norm": 8.08930492401123,
|
| 66 |
+
"learning_rate": 1.3970000000000002e-05,
|
| 67 |
+
"loss": 0.4761,
|
| 68 |
+
"step": 1400
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.64,
|
| 72 |
+
"grad_norm": 7.31475305557251,
|
| 73 |
+
"learning_rate": 1.597e-05,
|
| 74 |
+
"loss": 0.4942,
|
| 75 |
+
"step": 1600
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.72,
|
| 79 |
+
"grad_norm": 7.746321201324463,
|
| 80 |
+
"learning_rate": 1.796e-05,
|
| 81 |
+
"loss": 0.4758,
|
| 82 |
+
"step": 1800
|
| 83 |
+
},
|
| 84 |
+
{
|
| 85 |
+
"epoch": 0.8,
|
| 86 |
+
"grad_norm": 5.482341289520264,
|
| 87 |
+
"learning_rate": 1.9960000000000002e-05,
|
| 88 |
+
"loss": 0.4761,
|
| 89 |
+
"step": 2000
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 0.8,
|
| 93 |
+
"eval_loss": 0.34740468859672546,
|
| 94 |
+
"eval_runtime": 901.2547,
|
| 95 |
+
"eval_samples_per_second": 1.387,
|
| 96 |
+
"eval_steps_per_second": 0.693,
|
| 97 |
+
"eval_wer": 0.2396937573616019,
|
| 98 |
+
"step": 2000
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.88,
|
| 102 |
+
"grad_norm": 7.196951866149902,
|
| 103 |
+
"learning_rate": 1.9782222222222226e-05,
|
| 104 |
+
"loss": 0.4924,
|
| 105 |
+
"step": 2200
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.96,
|
| 109 |
+
"grad_norm": 6.221120834350586,
|
| 110 |
+
"learning_rate": 1.9560000000000002e-05,
|
| 111 |
+
"loss": 0.4779,
|
| 112 |
+
"step": 2400
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.04,
|
| 116 |
+
"grad_norm": 6.380108833312988,
|
| 117 |
+
"learning_rate": 1.933777777777778e-05,
|
| 118 |
+
"loss": 0.3806,
|
| 119 |
+
"step": 2600
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.12,
|
| 123 |
+
"grad_norm": 5.491684913635254,
|
| 124 |
+
"learning_rate": 1.9115555555555555e-05,
|
| 125 |
+
"loss": 0.3465,
|
| 126 |
+
"step": 2800
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 1.2,
|
| 130 |
+
"grad_norm": 4.4581379890441895,
|
| 131 |
+
"learning_rate": 1.8893333333333334e-05,
|
| 132 |
+
"loss": 0.3099,
|
| 133 |
+
"step": 3000
|
| 134 |
+
},
|
| 135 |
+
{
|
| 136 |
+
"epoch": 1.2,
|
| 137 |
+
"eval_loss": 0.30754217505455017,
|
| 138 |
+
"eval_runtime": 891.4205,
|
| 139 |
+
"eval_samples_per_second": 1.402,
|
| 140 |
+
"eval_steps_per_second": 0.701,
|
| 141 |
+
"eval_wer": 0.20259128386336867,
|
| 142 |
+
"step": 3000
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 1.28,
|
| 146 |
+
"grad_norm": 7.791854381561279,
|
| 147 |
+
"learning_rate": 1.8671111111111114e-05,
|
| 148 |
+
"loss": 0.3275,
|
| 149 |
+
"step": 3200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 1.3599999999999999,
|
| 153 |
+
"grad_norm": 5.212815761566162,
|
| 154 |
+
"learning_rate": 1.844888888888889e-05,
|
| 155 |
+
"loss": 0.3296,
|
| 156 |
+
"step": 3400
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 1.44,
|
| 160 |
+
"grad_norm": 9.706353187561035,
|
| 161 |
+
"learning_rate": 1.822666666666667e-05,
|
| 162 |
+
"loss": 0.3142,
|
| 163 |
+
"step": 3600
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 1.52,
|
| 167 |
+
"grad_norm": 5.222635269165039,
|
| 168 |
+
"learning_rate": 1.8004444444444446e-05,
|
| 169 |
+
"loss": 0.3097,
|
| 170 |
+
"step": 3800
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 1.6,
|
| 174 |
+
"grad_norm": 3.670520067214966,
|
| 175 |
+
"learning_rate": 1.7782222222222222e-05,
|
| 176 |
+
"loss": 0.3108,
|
| 177 |
+
"step": 4000
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 1.6,
|
| 181 |
+
"eval_loss": 0.26660874485969543,
|
| 182 |
+
"eval_runtime": 896.6083,
|
| 183 |
+
"eval_samples_per_second": 1.394,
|
| 184 |
+
"eval_steps_per_second": 0.697,
|
| 185 |
+
"eval_wer": 0.19709462112288967,
|
| 186 |
+
"step": 4000
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 1.6800000000000002,
|
| 190 |
+
"grad_norm": 5.417144298553467,
|
| 191 |
+
"learning_rate": 1.756e-05,
|
| 192 |
+
"loss": 0.3096,
|
| 193 |
+
"step": 4200
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 1.76,
|
| 197 |
+
"grad_norm": 5.865799427032471,
|
| 198 |
+
"learning_rate": 1.733777777777778e-05,
|
| 199 |
+
"loss": 0.3266,
|
| 200 |
+
"step": 4400
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 1.8399999999999999,
|
| 204 |
+
"grad_norm": 3.2357375621795654,
|
| 205 |
+
"learning_rate": 1.7115555555555557e-05,
|
| 206 |
+
"loss": 0.3134,
|
| 207 |
+
"step": 4600
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 1.92,
|
| 211 |
+
"grad_norm": 6.072234630584717,
|
| 212 |
+
"learning_rate": 1.6893333333333336e-05,
|
| 213 |
+
"loss": 0.289,
|
| 214 |
+
"step": 4800
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"epoch": 2.0,
|
| 218 |
+
"grad_norm": 3.143366813659668,
|
| 219 |
+
"learning_rate": 1.6671111111111113e-05,
|
| 220 |
+
"loss": 0.3131,
|
| 221 |
+
"step": 5000
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 2.0,
|
| 225 |
+
"eval_loss": 0.21725231409072876,
|
| 226 |
+
"eval_runtime": 885.5009,
|
| 227 |
+
"eval_samples_per_second": 1.412,
|
| 228 |
+
"eval_steps_per_second": 0.706,
|
| 229 |
+
"eval_wer": 0.15312131919905772,
|
| 230 |
+
"step": 5000
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 2.08,
|
| 234 |
+
"grad_norm": 4.550586700439453,
|
| 235 |
+
"learning_rate": 1.644888888888889e-05,
|
| 236 |
+
"loss": 0.1525,
|
| 237 |
+
"step": 5200
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 2.16,
|
| 241 |
+
"grad_norm": 3.604379892349243,
|
| 242 |
+
"learning_rate": 1.6226666666666668e-05,
|
| 243 |
+
"loss": 0.1656,
|
| 244 |
+
"step": 5400
|
| 245 |
+
},
|
| 246 |
+
{
|
| 247 |
+
"epoch": 2.24,
|
| 248 |
+
"grad_norm": 3.664323568344116,
|
| 249 |
+
"learning_rate": 1.6004444444444444e-05,
|
| 250 |
+
"loss": 0.1743,
|
| 251 |
+
"step": 5600
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 2.32,
|
| 255 |
+
"grad_norm": 2.5190961360931396,
|
| 256 |
+
"learning_rate": 1.5782222222222224e-05,
|
| 257 |
+
"loss": 0.175,
|
| 258 |
+
"step": 5800
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 2.4,
|
| 262 |
+
"grad_norm": 3.1211793422698975,
|
| 263 |
+
"learning_rate": 1.556e-05,
|
| 264 |
+
"loss": 0.1796,
|
| 265 |
+
"step": 6000
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 2.4,
|
| 269 |
+
"eval_loss": 0.2026291936635971,
|
| 270 |
+
"eval_runtime": 894.7957,
|
| 271 |
+
"eval_samples_per_second": 1.397,
|
| 272 |
+
"eval_steps_per_second": 0.698,
|
| 273 |
+
"eval_wer": 0.13839811542991756,
|
| 274 |
+
"step": 6000
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 2.48,
|
| 278 |
+
"grad_norm": 3.203092336654663,
|
| 279 |
+
"learning_rate": 1.533777777777778e-05,
|
| 280 |
+
"loss": 0.1721,
|
| 281 |
+
"step": 6200
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 2.56,
|
| 285 |
+
"grad_norm": 1.1959407329559326,
|
| 286 |
+
"learning_rate": 1.5115555555555557e-05,
|
| 287 |
+
"loss": 0.1722,
|
| 288 |
+
"step": 6400
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 2.64,
|
| 292 |
+
"grad_norm": 4.8345160484313965,
|
| 293 |
+
"learning_rate": 1.4893333333333335e-05,
|
| 294 |
+
"loss": 0.17,
|
| 295 |
+
"step": 6600
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 2.7199999999999998,
|
| 299 |
+
"grad_norm": 5.333383560180664,
|
| 300 |
+
"learning_rate": 1.4671111111111111e-05,
|
| 301 |
+
"loss": 0.1751,
|
| 302 |
+
"step": 6800
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 2.8,
|
| 306 |
+
"grad_norm": 1.5021018981933594,
|
| 307 |
+
"learning_rate": 1.444888888888889e-05,
|
| 308 |
+
"loss": 0.17,
|
| 309 |
+
"step": 7000
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 2.8,
|
| 313 |
+
"eval_loss": 0.19221611320972443,
|
| 314 |
+
"eval_runtime": 891.3109,
|
| 315 |
+
"eval_samples_per_second": 1.402,
|
| 316 |
+
"eval_steps_per_second": 0.701,
|
| 317 |
+
"eval_wer": 0.13780918727915195,
|
| 318 |
+
"step": 7000
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 2.88,
|
| 322 |
+
"grad_norm": 3.0855863094329834,
|
| 323 |
+
"learning_rate": 1.4226666666666669e-05,
|
| 324 |
+
"loss": 0.1672,
|
| 325 |
+
"step": 7200
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 2.96,
|
| 329 |
+
"grad_norm": 4.673288822174072,
|
| 330 |
+
"learning_rate": 1.4004444444444445e-05,
|
| 331 |
+
"loss": 0.1724,
|
| 332 |
+
"step": 7400
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 3.04,
|
| 336 |
+
"grad_norm": 2.8060922622680664,
|
| 337 |
+
"learning_rate": 1.3782222222222223e-05,
|
| 338 |
+
"loss": 0.1262,
|
| 339 |
+
"step": 7600
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 3.12,
|
| 343 |
+
"grad_norm": 1.5213590860366821,
|
| 344 |
+
"learning_rate": 1.3560000000000002e-05,
|
| 345 |
+
"loss": 0.094,
|
| 346 |
+
"step": 7800
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 3.2,
|
| 350 |
+
"grad_norm": 5.938966751098633,
|
| 351 |
+
"learning_rate": 1.3337777777777778e-05,
|
| 352 |
+
"loss": 0.0995,
|
| 353 |
+
"step": 8000
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 3.2,
|
| 357 |
+
"eval_loss": 0.17917467653751373,
|
| 358 |
+
"eval_runtime": 900.0369,
|
| 359 |
+
"eval_samples_per_second": 1.389,
|
| 360 |
+
"eval_steps_per_second": 0.694,
|
| 361 |
+
"eval_wer": 0.12063211621515509,
|
| 362 |
+
"step": 8000
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 3.2800000000000002,
|
| 366 |
+
"grad_norm": 3.5296432971954346,
|
| 367 |
+
"learning_rate": 1.3115555555555556e-05,
|
| 368 |
+
"loss": 0.1002,
|
| 369 |
+
"step": 8200
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 3.36,
|
| 373 |
+
"grad_norm": 5.334218502044678,
|
| 374 |
+
"learning_rate": 1.2893333333333336e-05,
|
| 375 |
+
"loss": 0.1061,
|
| 376 |
+
"step": 8400
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 3.44,
|
| 380 |
+
"grad_norm": 5.724509239196777,
|
| 381 |
+
"learning_rate": 1.2671111111111112e-05,
|
| 382 |
+
"loss": 0.1012,
|
| 383 |
+
"step": 8600
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 3.52,
|
| 387 |
+
"grad_norm": 2.5498409271240234,
|
| 388 |
+
"learning_rate": 1.244888888888889e-05,
|
| 389 |
+
"loss": 0.099,
|
| 390 |
+
"step": 8800
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 3.6,
|
| 394 |
+
"grad_norm": 6.009134769439697,
|
| 395 |
+
"learning_rate": 1.2226666666666666e-05,
|
| 396 |
+
"loss": 0.0972,
|
| 397 |
+
"step": 9000
|
| 398 |
+
},
|
| 399 |
+
{
|
| 400 |
+
"epoch": 3.6,
|
| 401 |
+
"eval_loss": 0.16701579093933105,
|
| 402 |
+
"eval_runtime": 899.6835,
|
| 403 |
+
"eval_samples_per_second": 1.389,
|
| 404 |
+
"eval_steps_per_second": 0.695,
|
| 405 |
+
"eval_wer": 0.11493914409108755,
|
| 406 |
+
"step": 9000
|
| 407 |
+
},
|
| 408 |
+
{
|
| 409 |
+
"epoch": 3.68,
|
| 410 |
+
"grad_norm": 2.472468137741089,
|
| 411 |
+
"learning_rate": 1.2004444444444445e-05,
|
| 412 |
+
"loss": 0.0933,
|
| 413 |
+
"step": 9200
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 3.76,
|
| 417 |
+
"grad_norm": 2.989001750946045,
|
| 418 |
+
"learning_rate": 1.1782222222222223e-05,
|
| 419 |
+
"loss": 0.0999,
|
| 420 |
+
"step": 9400
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 3.84,
|
| 424 |
+
"grad_norm": 3.3118479251861572,
|
| 425 |
+
"learning_rate": 1.156e-05,
|
| 426 |
+
"loss": 0.0949,
|
| 427 |
+
"step": 9600
|
| 428 |
+
},
|
| 429 |
+
{
|
| 430 |
+
"epoch": 3.92,
|
| 431 |
+
"grad_norm": 4.284938812255859,
|
| 432 |
+
"learning_rate": 1.1337777777777779e-05,
|
| 433 |
+
"loss": 0.0964,
|
| 434 |
+
"step": 9800
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 4.0,
|
| 438 |
+
"grad_norm": 3.4380695819854736,
|
| 439 |
+
"learning_rate": 1.1115555555555557e-05,
|
| 440 |
+
"loss": 0.097,
|
| 441 |
+
"step": 10000
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 4.0,
|
| 445 |
+
"eval_loss": 0.1545259952545166,
|
| 446 |
+
"eval_runtime": 903.5859,
|
| 447 |
+
"eval_samples_per_second": 1.383,
|
| 448 |
+
"eval_steps_per_second": 0.692,
|
| 449 |
+
"eval_wer": 0.10963879073419709,
|
| 450 |
+
"step": 10000
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 4.08,
|
| 454 |
+
"grad_norm": 1.92518949508667,
|
| 455 |
+
"learning_rate": 1.0893333333333333e-05,
|
| 456 |
+
"loss": 0.0524,
|
| 457 |
+
"step": 10200
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 4.16,
|
| 461 |
+
"grad_norm": 2.2278802394866943,
|
| 462 |
+
"learning_rate": 1.0671111111111112e-05,
|
| 463 |
+
"loss": 0.0523,
|
| 464 |
+
"step": 10400
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 4.24,
|
| 468 |
+
"grad_norm": 0.5611541271209717,
|
| 469 |
+
"learning_rate": 1.044888888888889e-05,
|
| 470 |
+
"loss": 0.0521,
|
| 471 |
+
"step": 10600
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 4.32,
|
| 475 |
+
"grad_norm": 1.7205854654312134,
|
| 476 |
+
"learning_rate": 1.0226666666666666e-05,
|
| 477 |
+
"loss": 0.0499,
|
| 478 |
+
"step": 10800
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 4.4,
|
| 482 |
+
"grad_norm": 2.563293218612671,
|
| 483 |
+
"learning_rate": 1.0004444444444446e-05,
|
| 484 |
+
"loss": 0.0553,
|
| 485 |
+
"step": 11000
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 4.4,
|
| 489 |
+
"eval_loss": 0.1574954092502594,
|
| 490 |
+
"eval_runtime": 893.7485,
|
| 491 |
+
"eval_samples_per_second": 1.399,
|
| 492 |
+
"eval_steps_per_second": 0.699,
|
| 493 |
+
"eval_wer": 0.10296427169218689,
|
| 494 |
+
"step": 11000
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 4.48,
|
| 498 |
+
"grad_norm": 2.2215042114257812,
|
| 499 |
+
"learning_rate": 9.783333333333335e-06,
|
| 500 |
+
"loss": 0.0537,
|
| 501 |
+
"step": 11200
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 4.5600000000000005,
|
| 505 |
+
"grad_norm": 1.360498070716858,
|
| 506 |
+
"learning_rate": 9.561111111111113e-06,
|
| 507 |
+
"loss": 0.0588,
|
| 508 |
+
"step": 11400
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"epoch": 4.64,
|
| 512 |
+
"grad_norm": 4.086482048034668,
|
| 513 |
+
"learning_rate": 9.340000000000002e-06,
|
| 514 |
+
"loss": 0.0537,
|
| 515 |
+
"step": 11600
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 4.72,
|
| 519 |
+
"grad_norm": 2.1945817470550537,
|
| 520 |
+
"learning_rate": 9.117777777777778e-06,
|
| 521 |
+
"loss": 0.0599,
|
| 522 |
+
"step": 11800
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 4.8,
|
| 526 |
+
"grad_norm": 2.7617948055267334,
|
| 527 |
+
"learning_rate": 8.895555555555556e-06,
|
| 528 |
+
"loss": 0.0526,
|
| 529 |
+
"step": 12000
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"epoch": 4.8,
|
| 533 |
+
"eval_loss": 0.14308780431747437,
|
| 534 |
+
"eval_runtime": 905.6641,
|
| 535 |
+
"eval_samples_per_second": 1.38,
|
| 536 |
+
"eval_steps_per_second": 0.69,
|
| 537 |
+
"eval_wer": 0.08902630545740087,
|
| 538 |
+
"step": 12000
|
| 539 |
+
},
|
| 540 |
+
{
|
| 541 |
+
"epoch": 4.88,
|
| 542 |
+
"grad_norm": 0.3309612572193146,
|
| 543 |
+
"learning_rate": 8.673333333333334e-06,
|
| 544 |
+
"loss": 0.0527,
|
| 545 |
+
"step": 12200
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 4.96,
|
| 549 |
+
"grad_norm": 0.3702127933502197,
|
| 550 |
+
"learning_rate": 8.451111111111112e-06,
|
| 551 |
+
"loss": 0.0574,
|
| 552 |
+
"step": 12400
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 5.04,
|
| 556 |
+
"grad_norm": 1.9310526847839355,
|
| 557 |
+
"learning_rate": 8.22888888888889e-06,
|
| 558 |
+
"loss": 0.0385,
|
| 559 |
+
"step": 12600
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 5.12,
|
| 563 |
+
"grad_norm": 0.8932788372039795,
|
| 564 |
+
"learning_rate": 8.006666666666667e-06,
|
| 565 |
+
"loss": 0.0369,
|
| 566 |
+
"step": 12800
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 5.2,
|
| 570 |
+
"grad_norm": 2.246431589126587,
|
| 571 |
+
"learning_rate": 7.784444444444445e-06,
|
| 572 |
+
"loss": 0.0299,
|
| 573 |
+
"step": 13000
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 5.2,
|
| 577 |
+
"eval_loss": 0.13719478249549866,
|
| 578 |
+
"eval_runtime": 896.1487,
|
| 579 |
+
"eval_samples_per_second": 1.395,
|
| 580 |
+
"eval_steps_per_second": 0.697,
|
| 581 |
+
"eval_wer": 0.08333333333333333,
|
| 582 |
+
"step": 13000
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 5.28,
|
| 586 |
+
"grad_norm": 2.8669216632843018,
|
| 587 |
+
"learning_rate": 7.562222222222223e-06,
|
| 588 |
+
"loss": 0.0276,
|
| 589 |
+
"step": 13200
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 5.36,
|
| 593 |
+
"grad_norm": 0.38674354553222656,
|
| 594 |
+
"learning_rate": 7.341111111111112e-06,
|
| 595 |
+
"loss": 0.0334,
|
| 596 |
+
"step": 13400
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 5.44,
|
| 600 |
+
"grad_norm": 4.218255996704102,
|
| 601 |
+
"learning_rate": 7.11888888888889e-06,
|
| 602 |
+
"loss": 0.0253,
|
| 603 |
+
"step": 13600
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 5.52,
|
| 607 |
+
"grad_norm": 2.424956798553467,
|
| 608 |
+
"learning_rate": 6.896666666666667e-06,
|
| 609 |
+
"loss": 0.0287,
|
| 610 |
+
"step": 13800
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"epoch": 5.6,
|
| 614 |
+
"grad_norm": 1.3083245754241943,
|
| 615 |
+
"learning_rate": 6.674444444444445e-06,
|
| 616 |
+
"loss": 0.0311,
|
| 617 |
+
"step": 14000
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 5.6,
|
| 621 |
+
"eval_loss": 0.12584801018238068,
|
| 622 |
+
"eval_runtime": 892.859,
|
| 623 |
+
"eval_samples_per_second": 1.4,
|
| 624 |
+
"eval_steps_per_second": 0.7,
|
| 625 |
+
"eval_wer": 0.07803297997644287,
|
| 626 |
+
"step": 14000
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 5.68,
|
| 630 |
+
"grad_norm": 1.7961013317108154,
|
| 631 |
+
"learning_rate": 6.452222222222223e-06,
|
| 632 |
+
"loss": 0.0291,
|
| 633 |
+
"step": 14200
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 5.76,
|
| 637 |
+
"grad_norm": 0.16184474527835846,
|
| 638 |
+
"learning_rate": 6.2300000000000005e-06,
|
| 639 |
+
"loss": 0.0287,
|
| 640 |
+
"step": 14400
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 5.84,
|
| 644 |
+
"grad_norm": 0.13447080552577972,
|
| 645 |
+
"learning_rate": 6.007777777777778e-06,
|
| 646 |
+
"loss": 0.0303,
|
| 647 |
+
"step": 14600
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 5.92,
|
| 651 |
+
"grad_norm": 2.1226162910461426,
|
| 652 |
+
"learning_rate": 5.785555555555556e-06,
|
| 653 |
+
"loss": 0.0261,
|
| 654 |
+
"step": 14800
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 6.0,
|
| 658 |
+
"grad_norm": 1.8055452108383179,
|
| 659 |
+
"learning_rate": 5.563333333333334e-06,
|
| 660 |
+
"loss": 0.0295,
|
| 661 |
+
"step": 15000
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 6.0,
|
| 665 |
+
"eval_loss": 0.12012948095798492,
|
| 666 |
+
"eval_runtime": 888.9428,
|
| 667 |
+
"eval_samples_per_second": 1.406,
|
| 668 |
+
"eval_steps_per_second": 0.703,
|
| 669 |
+
"eval_wer": 0.07253631723596388,
|
| 670 |
+
"step": 15000
|
| 671 |
+
},
|
| 672 |
+
{
|
| 673 |
+
"epoch": 6.08,
|
| 674 |
+
"grad_norm": 0.1165083572268486,
|
| 675 |
+
"learning_rate": 5.341111111111111e-06,
|
| 676 |
+
"loss": 0.0143,
|
| 677 |
+
"step": 15200
|
| 678 |
+
},
|
| 679 |
+
{
|
| 680 |
+
"epoch": 6.16,
|
| 681 |
+
"grad_norm": 0.12827950716018677,
|
| 682 |
+
"learning_rate": 5.118888888888889e-06,
|
| 683 |
+
"loss": 0.0169,
|
| 684 |
+
"step": 15400
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 6.24,
|
| 688 |
+
"grad_norm": 3.075695276260376,
|
| 689 |
+
"learning_rate": 4.896666666666667e-06,
|
| 690 |
+
"loss": 0.0144,
|
| 691 |
+
"step": 15600
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 6.32,
|
| 695 |
+
"grad_norm": 0.4469108581542969,
|
| 696 |
+
"learning_rate": 4.6744444444444445e-06,
|
| 697 |
+
"loss": 0.0124,
|
| 698 |
+
"step": 15800
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 6.4,
|
| 702 |
+
"grad_norm": 0.264642596244812,
|
| 703 |
+
"learning_rate": 4.452222222222223e-06,
|
| 704 |
+
"loss": 0.0151,
|
| 705 |
+
"step": 16000
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 6.4,
|
| 709 |
+
"eval_loss": 0.12285350263118744,
|
| 710 |
+
"eval_runtime": 883.7262,
|
| 711 |
+
"eval_samples_per_second": 1.414,
|
| 712 |
+
"eval_steps_per_second": 0.707,
|
| 713 |
+
"eval_wer": 0.07420494699646643,
|
| 714 |
+
"step": 16000
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 6.48,
|
| 718 |
+
"grad_norm": 3.0525002479553223,
|
| 719 |
+
"learning_rate": 4.23e-06,
|
| 720 |
+
"loss": 0.0122,
|
| 721 |
+
"step": 16200
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 6.5600000000000005,
|
| 725 |
+
"grad_norm": 1.7435777187347412,
|
| 726 |
+
"learning_rate": 4.007777777777778e-06,
|
| 727 |
+
"loss": 0.0134,
|
| 728 |
+
"step": 16400
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 6.64,
|
| 732 |
+
"grad_norm": 0.03412042185664177,
|
| 733 |
+
"learning_rate": 3.785555555555556e-06,
|
| 734 |
+
"loss": 0.0146,
|
| 735 |
+
"step": 16600
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 6.72,
|
| 739 |
+
"grad_norm": 4.467521667480469,
|
| 740 |
+
"learning_rate": 3.5633333333333337e-06,
|
| 741 |
+
"loss": 0.0139,
|
| 742 |
+
"step": 16800
|
| 743 |
+
},
|
| 744 |
+
{
|
| 745 |
+
"epoch": 6.8,
|
| 746 |
+
"grad_norm": 1.6740899085998535,
|
| 747 |
+
"learning_rate": 3.3411111111111115e-06,
|
| 748 |
+
"loss": 0.0163,
|
| 749 |
+
"step": 17000
|
| 750 |
+
},
|
| 751 |
+
{
|
| 752 |
+
"epoch": 6.8,
|
| 753 |
+
"eval_loss": 0.1136574074625969,
|
| 754 |
+
"eval_runtime": 890.0494,
|
| 755 |
+
"eval_samples_per_second": 1.404,
|
| 756 |
+
"eval_steps_per_second": 0.702,
|
| 757 |
+
"eval_wer": 0.06576364350215941,
|
| 758 |
+
"step": 17000
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 6.88,
|
| 762 |
+
"grad_norm": 0.25668439269065857,
|
| 763 |
+
"learning_rate": 3.118888888888889e-06,
|
| 764 |
+
"loss": 0.0143,
|
| 765 |
+
"step": 17200
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 6.96,
|
| 769 |
+
"grad_norm": 1.1925643682479858,
|
| 770 |
+
"learning_rate": 2.896666666666667e-06,
|
| 771 |
+
"loss": 0.0132,
|
| 772 |
+
"step": 17400
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 7.04,
|
| 776 |
+
"grad_norm": 0.45678913593292236,
|
| 777 |
+
"learning_rate": 2.6744444444444446e-06,
|
| 778 |
+
"loss": 0.0103,
|
| 779 |
+
"step": 17600
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 7.12,
|
| 783 |
+
"grad_norm": 0.25028079748153687,
|
| 784 |
+
"learning_rate": 2.4522222222222224e-06,
|
| 785 |
+
"loss": 0.0089,
|
| 786 |
+
"step": 17800
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 7.2,
|
| 790 |
+
"grad_norm": 0.04460795596241951,
|
| 791 |
+
"learning_rate": 2.2300000000000002e-06,
|
| 792 |
+
"loss": 0.0082,
|
| 793 |
+
"step": 18000
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 7.2,
|
| 797 |
+
"eval_loss": 0.1142345443367958,
|
| 798 |
+
"eval_runtime": 885.5644,
|
| 799 |
+
"eval_samples_per_second": 1.412,
|
| 800 |
+
"eval_steps_per_second": 0.706,
|
| 801 |
+
"eval_wer": 0.06389870435806831,
|
| 802 |
+
"step": 18000
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 7.28,
|
| 806 |
+
"grad_norm": 6.106534957885742,
|
| 807 |
+
"learning_rate": 2.007777777777778e-06,
|
| 808 |
+
"loss": 0.0063,
|
| 809 |
+
"step": 18200
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 7.36,
|
| 813 |
+
"grad_norm": 0.5640923380851746,
|
| 814 |
+
"learning_rate": 1.7855555555555557e-06,
|
| 815 |
+
"loss": 0.0088,
|
| 816 |
+
"step": 18400
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 7.44,
|
| 820 |
+
"grad_norm": 0.09705500304698944,
|
| 821 |
+
"learning_rate": 1.5633333333333333e-06,
|
| 822 |
+
"loss": 0.0091,
|
| 823 |
+
"step": 18600
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 7.52,
|
| 827 |
+
"grad_norm": 0.042120128870010376,
|
| 828 |
+
"learning_rate": 1.3411111111111112e-06,
|
| 829 |
+
"loss": 0.0066,
|
| 830 |
+
"step": 18800
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 7.6,
|
| 834 |
+
"grad_norm": 0.024996856227517128,
|
| 835 |
+
"learning_rate": 1.118888888888889e-06,
|
| 836 |
+
"loss": 0.0092,
|
| 837 |
+
"step": 19000
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 7.6,
|
| 841 |
+
"eval_loss": 0.11208222806453705,
|
| 842 |
+
"eval_runtime": 886.2816,
|
| 843 |
+
"eval_samples_per_second": 1.41,
|
| 844 |
+
"eval_steps_per_second": 0.705,
|
| 845 |
+
"eval_wer": 0.0627208480565371,
|
| 846 |
+
"step": 19000
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 7.68,
|
| 850 |
+
"grad_norm": 0.38041412830352783,
|
| 851 |
+
"learning_rate": 8.966666666666668e-07,
|
| 852 |
+
"loss": 0.0068,
|
| 853 |
+
"step": 19200
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 7.76,
|
| 857 |
+
"grad_norm": 3.13415265083313,
|
| 858 |
+
"learning_rate": 6.744444444444446e-07,
|
| 859 |
+
"loss": 0.0062,
|
| 860 |
+
"step": 19400
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 7.84,
|
| 864 |
+
"grad_norm": 0.7191887497901917,
|
| 865 |
+
"learning_rate": 4.5222222222222224e-07,
|
| 866 |
+
"loss": 0.0067,
|
| 867 |
+
"step": 19600
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 7.92,
|
| 871 |
+
"grad_norm": 0.09878556430339813,
|
| 872 |
+
"learning_rate": 2.3000000000000002e-07,
|
| 873 |
+
"loss": 0.0077,
|
| 874 |
+
"step": 19800
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 8.0,
|
| 878 |
+
"grad_norm": 0.007985732518136501,
|
| 879 |
+
"learning_rate": 7.777777777777778e-09,
|
| 880 |
+
"loss": 0.006,
|
| 881 |
+
"step": 20000
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 8.0,
|
| 885 |
+
"eval_loss": 0.11124598234891891,
|
| 886 |
+
"eval_runtime": 889.6592,
|
| 887 |
+
"eval_samples_per_second": 1.405,
|
| 888 |
+
"eval_steps_per_second": 0.703,
|
| 889 |
+
"eval_wer": 0.06115037298782882,
|
| 890 |
+
"step": 20000
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 8.0,
|
| 894 |
+
"step": 20000,
|
| 895 |
+
"total_flos": 1.698366962958336e+20,
|
| 896 |
+
"train_loss": 0.16200368287563324,
|
| 897 |
+
"train_runtime": 40923.5162,
|
| 898 |
+
"train_samples_per_second": 1.955,
|
| 899 |
+
"train_steps_per_second": 0.489
|
| 900 |
+
}
|
| 901 |
+
],
|
| 902 |
+
"logging_steps": 200,
|
| 903 |
+
"max_steps": 20000,
|
| 904 |
+
"num_input_tokens_seen": 0,
|
| 905 |
+
"num_train_epochs": 8,
|
| 906 |
+
"save_steps": 1000,
|
| 907 |
+
"stateful_callbacks": {
|
| 908 |
+
"TrainerControl": {
|
| 909 |
+
"args": {
|
| 910 |
+
"should_epoch_stop": false,
|
| 911 |
+
"should_evaluate": false,
|
| 912 |
+
"should_log": false,
|
| 913 |
+
"should_save": true,
|
| 914 |
+
"should_training_stop": true
|
| 915 |
+
},
|
| 916 |
+
"attributes": {}
|
| 917 |
+
}
|
| 918 |
+
},
|
| 919 |
+
"total_flos": 1.698366962958336e+20,
|
| 920 |
+
"train_batch_size": 2,
|
| 921 |
+
"trial_name": null,
|
| 922 |
+
"trial_params": null
|
| 923 |
+
}
|
val_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 8.0,
|
| 3 |
+
"val_loss": 0.11124598234891891,
|
| 4 |
+
"val_runtime": 893.4451,
|
| 5 |
+
"val_samples": 250,
|
| 6 |
+
"val_samples_per_second": 1.399,
|
| 7 |
+
"val_steps_per_second": 0.7,
|
| 8 |
+
"val_wer": 0.06115037298782882
|
| 9 |
+
}
|