gary109 commited on
Commit
ac61c5b
1 Parent(s): 7372ac0

End of training

Browse files
all_results.json ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 10.0,
3
+ "eval_loss": 0.43008479475975037,
4
+ "eval_runtime": 246.4992,
5
+ "eval_samples": 1965,
6
+ "eval_samples_per_second": 7.972,
7
+ "eval_steps_per_second": 3.988,
8
+ "eval_wer": 0.1633294313350995,
9
+ "train_loss": 0.11096900397020837,
10
+ "train_runtime": 38440.8489,
11
+ "train_samples": 17685,
12
+ "train_samples_per_second": 4.601,
13
+ "train_steps_per_second": 0.144
14
+ }
eval_results.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 10.0,
3
+ "eval_loss": 0.43008479475975037,
4
+ "eval_runtime": 246.4992,
5
+ "eval_samples": 1965,
6
+ "eval_samples_per_second": 7.972,
7
+ "eval_steps_per_second": 3.988,
8
+ "eval_wer": 0.1633294313350995
9
+ }
runs/Jun19_11-52-21_e8a2ffe351ca/events.out.tfevents.1655679568.e8a2ffe351ca.4100.2 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:69dbb2f84cb927bc1bd20bab4acd5c3265df70f62ec5ba35fc895af51d6c23e8
3
+ size 358
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 10.0,
3
+ "train_loss": 0.11096900397020837,
4
+ "train_runtime": 38440.8489,
5
+ "train_samples": 17685,
6
+ "train_samples_per_second": 4.601,
7
+ "train_steps_per_second": 0.144
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,3427 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.43008479475975037,
3
+ "best_model_checkpoint": "ai-light-dance_singing_ft_wav2vec2-large-xlsr-53-5gram-v3/checkpoint-552",
4
+ "epoch": 9.998756078253987,
5
+ "global_step": 5520,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.02,
12
+ "learning_rate": 2.0000000000000002e-07,
13
+ "loss": 0.144,
14
+ "step": 10
15
+ },
16
+ {
17
+ "epoch": 0.04,
18
+ "learning_rate": 4.0000000000000003e-07,
19
+ "loss": 0.1475,
20
+ "step": 20
21
+ },
22
+ {
23
+ "epoch": 0.05,
24
+ "learning_rate": 6.000000000000001e-07,
25
+ "loss": 0.1668,
26
+ "step": 30
27
+ },
28
+ {
29
+ "epoch": 0.07,
30
+ "learning_rate": 8.000000000000001e-07,
31
+ "loss": 0.1406,
32
+ "step": 40
33
+ },
34
+ {
35
+ "epoch": 0.09,
36
+ "learning_rate": 9.800000000000001e-07,
37
+ "loss": 0.165,
38
+ "step": 50
39
+ },
40
+ {
41
+ "epoch": 0.11,
42
+ "learning_rate": 1.1800000000000001e-06,
43
+ "loss": 0.1566,
44
+ "step": 60
45
+ },
46
+ {
47
+ "epoch": 0.13,
48
+ "learning_rate": 1.3800000000000001e-06,
49
+ "loss": 0.1449,
50
+ "step": 70
51
+ },
52
+ {
53
+ "epoch": 0.14,
54
+ "learning_rate": 1.5800000000000001e-06,
55
+ "loss": 0.1674,
56
+ "step": 80
57
+ },
58
+ {
59
+ "epoch": 0.16,
60
+ "learning_rate": 1.7800000000000001e-06,
61
+ "loss": 0.1563,
62
+ "step": 90
63
+ },
64
+ {
65
+ "epoch": 0.18,
66
+ "learning_rate": 1.98e-06,
67
+ "loss": 0.1726,
68
+ "step": 100
69
+ },
70
+ {
71
+ "epoch": 0.2,
72
+ "learning_rate": 2.1800000000000003e-06,
73
+ "loss": 0.1373,
74
+ "step": 110
75
+ },
76
+ {
77
+ "epoch": 0.22,
78
+ "learning_rate": 2.38e-06,
79
+ "loss": 0.1516,
80
+ "step": 120
81
+ },
82
+ {
83
+ "epoch": 0.24,
84
+ "learning_rate": 2.5800000000000003e-06,
85
+ "loss": 0.157,
86
+ "step": 130
87
+ },
88
+ {
89
+ "epoch": 0.25,
90
+ "learning_rate": 2.7800000000000005e-06,
91
+ "loss": 0.1434,
92
+ "step": 140
93
+ },
94
+ {
95
+ "epoch": 0.27,
96
+ "learning_rate": 2.9800000000000003e-06,
97
+ "loss": 0.157,
98
+ "step": 150
99
+ },
100
+ {
101
+ "epoch": 0.29,
102
+ "learning_rate": 3.1800000000000005e-06,
103
+ "loss": 0.153,
104
+ "step": 160
105
+ },
106
+ {
107
+ "epoch": 0.31,
108
+ "learning_rate": 3.3800000000000007e-06,
109
+ "loss": 0.1445,
110
+ "step": 170
111
+ },
112
+ {
113
+ "epoch": 0.33,
114
+ "learning_rate": 3.58e-06,
115
+ "loss": 0.1579,
116
+ "step": 180
117
+ },
118
+ {
119
+ "epoch": 0.34,
120
+ "learning_rate": 3.7800000000000002e-06,
121
+ "loss": 0.1357,
122
+ "step": 190
123
+ },
124
+ {
125
+ "epoch": 0.36,
126
+ "learning_rate": 3.980000000000001e-06,
127
+ "loss": 0.145,
128
+ "step": 200
129
+ },
130
+ {
131
+ "epoch": 0.38,
132
+ "learning_rate": 4.18e-06,
133
+ "loss": 0.1465,
134
+ "step": 210
135
+ },
136
+ {
137
+ "epoch": 0.4,
138
+ "learning_rate": 4.38e-06,
139
+ "loss": 0.1298,
140
+ "step": 220
141
+ },
142
+ {
143
+ "epoch": 0.42,
144
+ "learning_rate": 4.58e-06,
145
+ "loss": 0.1416,
146
+ "step": 230
147
+ },
148
+ {
149
+ "epoch": 0.43,
150
+ "learning_rate": 4.78e-06,
151
+ "loss": 0.1402,
152
+ "step": 240
153
+ },
154
+ {
155
+ "epoch": 0.45,
156
+ "learning_rate": 4.980000000000001e-06,
157
+ "loss": 0.1361,
158
+ "step": 250
159
+ },
160
+ {
161
+ "epoch": 0.47,
162
+ "learning_rate": 5.18e-06,
163
+ "loss": 0.1479,
164
+ "step": 260
165
+ },
166
+ {
167
+ "epoch": 0.49,
168
+ "learning_rate": 5.380000000000001e-06,
169
+ "loss": 0.1333,
170
+ "step": 270
171
+ },
172
+ {
173
+ "epoch": 0.51,
174
+ "learning_rate": 5.580000000000001e-06,
175
+ "loss": 0.1527,
176
+ "step": 280
177
+ },
178
+ {
179
+ "epoch": 0.52,
180
+ "learning_rate": 5.78e-06,
181
+ "loss": 0.1352,
182
+ "step": 290
183
+ },
184
+ {
185
+ "epoch": 0.54,
186
+ "learning_rate": 5.98e-06,
187
+ "loss": 0.1596,
188
+ "step": 300
189
+ },
190
+ {
191
+ "epoch": 0.56,
192
+ "learning_rate": 6.18e-06,
193
+ "loss": 0.1446,
194
+ "step": 310
195
+ },
196
+ {
197
+ "epoch": 0.58,
198
+ "learning_rate": 6.380000000000001e-06,
199
+ "loss": 0.146,
200
+ "step": 320
201
+ },
202
+ {
203
+ "epoch": 0.6,
204
+ "learning_rate": 6.5800000000000005e-06,
205
+ "loss": 0.1631,
206
+ "step": 330
207
+ },
208
+ {
209
+ "epoch": 0.62,
210
+ "learning_rate": 6.780000000000001e-06,
211
+ "loss": 0.1377,
212
+ "step": 340
213
+ },
214
+ {
215
+ "epoch": 0.63,
216
+ "learning_rate": 6.98e-06,
217
+ "loss": 0.1541,
218
+ "step": 350
219
+ },
220
+ {
221
+ "epoch": 0.65,
222
+ "learning_rate": 7.180000000000001e-06,
223
+ "loss": 0.12,
224
+ "step": 360
225
+ },
226
+ {
227
+ "epoch": 0.67,
228
+ "learning_rate": 7.3800000000000005e-06,
229
+ "loss": 0.1318,
230
+ "step": 370
231
+ },
232
+ {
233
+ "epoch": 0.69,
234
+ "learning_rate": 7.58e-06,
235
+ "loss": 0.1396,
236
+ "step": 380
237
+ },
238
+ {
239
+ "epoch": 0.71,
240
+ "learning_rate": 7.78e-06,
241
+ "loss": 0.138,
242
+ "step": 390
243
+ },
244
+ {
245
+ "epoch": 0.72,
246
+ "learning_rate": 7.980000000000002e-06,
247
+ "loss": 0.1683,
248
+ "step": 400
249
+ },
250
+ {
251
+ "epoch": 0.74,
252
+ "learning_rate": 8.18e-06,
253
+ "loss": 0.1427,
254
+ "step": 410
255
+ },
256
+ {
257
+ "epoch": 0.76,
258
+ "learning_rate": 8.380000000000001e-06,
259
+ "loss": 0.1552,
260
+ "step": 420
261
+ },
262
+ {
263
+ "epoch": 0.78,
264
+ "learning_rate": 8.580000000000001e-06,
265
+ "loss": 0.1406,
266
+ "step": 430
267
+ },
268
+ {
269
+ "epoch": 0.8,
270
+ "learning_rate": 8.78e-06,
271
+ "loss": 0.1287,
272
+ "step": 440
273
+ },
274
+ {
275
+ "epoch": 0.81,
276
+ "learning_rate": 8.96e-06,
277
+ "loss": 0.1341,
278
+ "step": 450
279
+ },
280
+ {
281
+ "epoch": 0.83,
282
+ "learning_rate": 9.16e-06,
283
+ "loss": 0.1316,
284
+ "step": 460
285
+ },
286
+ {
287
+ "epoch": 0.85,
288
+ "learning_rate": 9.360000000000002e-06,
289
+ "loss": 0.1339,
290
+ "step": 470
291
+ },
292
+ {
293
+ "epoch": 0.87,
294
+ "learning_rate": 9.56e-06,
295
+ "loss": 0.1129,
296
+ "step": 480
297
+ },
298
+ {
299
+ "epoch": 0.89,
300
+ "learning_rate": 9.760000000000001e-06,
301
+ "loss": 0.1334,
302
+ "step": 490
303
+ },
304
+ {
305
+ "epoch": 0.9,
306
+ "learning_rate": 9.960000000000001e-06,
307
+ "loss": 0.1608,
308
+ "step": 500
309
+ },
310
+ {
311
+ "epoch": 0.92,
312
+ "learning_rate": 9.999937336970615e-06,
313
+ "loss": 0.1567,
314
+ "step": 510
315
+ },
316
+ {
317
+ "epoch": 0.94,
318
+ "learning_rate": 9.999682771105642e-06,
319
+ "loss": 0.1204,
320
+ "step": 520
321
+ },
322
+ {
323
+ "epoch": 0.96,
324
+ "learning_rate": 9.99923239592798e-06,
325
+ "loss": 0.128,
326
+ "step": 530
327
+ },
328
+ {
329
+ "epoch": 0.98,
330
+ "learning_rate": 9.998586229076279e-06,
331
+ "loss": 0.1416,
332
+ "step": 540
333
+ },
334
+ {
335
+ "epoch": 1.0,
336
+ "learning_rate": 9.997744295857243e-06,
337
+ "loss": 0.1517,
338
+ "step": 550
339
+ },
340
+ {
341
+ "epoch": 1.0,
342
+ "eval_loss": 0.43008479475975037,
343
+ "eval_runtime": 248.9762,
344
+ "eval_samples_per_second": 7.892,
345
+ "eval_steps_per_second": 3.948,
346
+ "eval_wer": 0.1633294313350995,
347
+ "step": 552
348
+ },
349
+ {
350
+ "epoch": 1.01,
351
+ "learning_rate": 9.996706629244636e-06,
352
+ "loss": 0.1469,
353
+ "step": 560
354
+ },
355
+ {
356
+ "epoch": 1.03,
357
+ "learning_rate": 9.995473269877994e-06,
358
+ "loss": 0.1207,
359
+ "step": 570
360
+ },
361
+ {
362
+ "epoch": 1.05,
363
+ "learning_rate": 9.994044266061033e-06,
364
+ "loss": 0.1239,
365
+ "step": 580
366
+ },
367
+ {
368
+ "epoch": 1.07,
369
+ "learning_rate": 9.992419673759752e-06,
370
+ "loss": 0.1294,
371
+ "step": 590
372
+ },
373
+ {
374
+ "epoch": 1.09,
375
+ "learning_rate": 9.990599556600252e-06,
376
+ "loss": 0.1446,
377
+ "step": 600
378
+ },
379
+ {
380
+ "epoch": 1.1,
381
+ "learning_rate": 9.98858398586623e-06,
382
+ "loss": 0.1397,
383
+ "step": 610
384
+ },
385
+ {
386
+ "epoch": 1.12,
387
+ "learning_rate": 9.9863730404962e-06,
388
+ "loss": 0.149,
389
+ "step": 620
390
+ },
391
+ {
392
+ "epoch": 1.14,
393
+ "learning_rate": 9.983966807080394e-06,
394
+ "loss": 0.1448,
395
+ "step": 630
396
+ },
397
+ {
398
+ "epoch": 1.16,
399
+ "learning_rate": 9.981365379857373e-06,
400
+ "loss": 0.1295,
401
+ "step": 640
402
+ },
403
+ {
404
+ "epoch": 1.18,
405
+ "learning_rate": 9.97856886071034e-06,
406
+ "loss": 0.1278,
407
+ "step": 650
408
+ },
409
+ {
410
+ "epoch": 1.2,
411
+ "learning_rate": 9.975577359163137e-06,
412
+ "loss": 0.1421,
413
+ "step": 660
414
+ },
415
+ {
416
+ "epoch": 1.21,
417
+ "learning_rate": 9.972390992375972e-06,
418
+ "loss": 0.1156,
419
+ "step": 670
420
+ },
421
+ {
422
+ "epoch": 1.23,
423
+ "learning_rate": 9.969009885140821e-06,
424
+ "loss": 0.1355,
425
+ "step": 680
426
+ },
427
+ {
428
+ "epoch": 1.25,
429
+ "learning_rate": 9.96543416987654e-06,
430
+ "loss": 0.1135,
431
+ "step": 690
432
+ },
433
+ {
434
+ "epoch": 1.27,
435
+ "learning_rate": 9.961663986623688e-06,
436
+ "loss": 0.1506,
437
+ "step": 700
438
+ },
439
+ {
440
+ "epoch": 1.29,
441
+ "learning_rate": 9.957699483039027e-06,
442
+ "loss": 0.1439,
443
+ "step": 710
444
+ },
445
+ {
446
+ "epoch": 1.3,
447
+ "learning_rate": 9.953540814389754e-06,
448
+ "loss": 0.1195,
449
+ "step": 720
450
+ },
451
+ {
452
+ "epoch": 1.32,
453
+ "learning_rate": 9.949188143547414e-06,
454
+ "loss": 0.1406,
455
+ "step": 730
456
+ },
457
+ {
458
+ "epoch": 1.34,
459
+ "learning_rate": 9.944641640981519e-06,
460
+ "loss": 0.1145,
461
+ "step": 740
462
+ },
463
+ {
464
+ "epoch": 1.36,
465
+ "learning_rate": 9.939901484752874e-06,
466
+ "loss": 0.1502,
467
+ "step": 750
468
+ },
469
+ {
470
+ "epoch": 1.38,
471
+ "learning_rate": 9.934967860506604e-06,
472
+ "loss": 0.1428,
473
+ "step": 760
474
+ },
475
+ {
476
+ "epoch": 1.39,
477
+ "learning_rate": 9.929840961464883e-06,
478
+ "loss": 0.1237,
479
+ "step": 770
480
+ },
481
+ {
482
+ "epoch": 1.41,
483
+ "learning_rate": 9.924520988419366e-06,
484
+ "loss": 0.1519,
485
+ "step": 780
486
+ },
487
+ {
488
+ "epoch": 1.43,
489
+ "learning_rate": 9.919008149723326e-06,
490
+ "loss": 0.1247,
491
+ "step": 790
492
+ },
493
+ {
494
+ "epoch": 1.45,
495
+ "learning_rate": 9.91330266128349e-06,
496
+ "loss": 0.1406,
497
+ "step": 800
498
+ },
499
+ {
500
+ "epoch": 1.47,
501
+ "learning_rate": 9.90740474655159e-06,
502
+ "loss": 0.1437,
503
+ "step": 810
504
+ },
505
+ {
506
+ "epoch": 1.48,
507
+ "learning_rate": 9.901314636515612e-06,
508
+ "loss": 0.1324,
509
+ "step": 820
510
+ },
511
+ {
512
+ "epoch": 1.5,
513
+ "learning_rate": 9.895032569690732e-06,
514
+ "loss": 0.1244,
515
+ "step": 830
516
+ },
517
+ {
518
+ "epoch": 1.52,
519
+ "learning_rate": 9.888558792110005e-06,
520
+ "loss": 0.1244,
521
+ "step": 840
522
+ },
523
+ {
524
+ "epoch": 1.54,
525
+ "learning_rate": 9.881893557314701e-06,
526
+ "loss": 0.1293,
527
+ "step": 850
528
+ },
529
+ {
530
+ "epoch": 1.56,
531
+ "learning_rate": 9.875037126344394e-06,
532
+ "loss": 0.143,
533
+ "step": 860
534
+ },
535
+ {
536
+ "epoch": 1.58,
537
+ "learning_rate": 9.867989767726722e-06,
538
+ "loss": 0.1339,
539
+ "step": 870
540
+ },
541
+ {
542
+ "epoch": 1.59,
543
+ "learning_rate": 9.86075175746689e-06,
544
+ "loss": 0.14,
545
+ "step": 880
546
+ },
547
+ {
548
+ "epoch": 1.61,
549
+ "learning_rate": 9.853323379036848e-06,
550
+ "loss": 0.1424,
551
+ "step": 890
552
+ },
553
+ {
554
+ "epoch": 1.63,
555
+ "learning_rate": 9.845704923364186e-06,
556
+ "loss": 0.1209,
557
+ "step": 900
558
+ },
559
+ {
560
+ "epoch": 1.65,
561
+ "learning_rate": 9.837896688820747e-06,
562
+ "loss": 0.1518,
563
+ "step": 910
564
+ },
565
+ {
566
+ "epoch": 1.67,
567
+ "learning_rate": 9.829898981210944e-06,
568
+ "loss": 0.1168,
569
+ "step": 920
570
+ },
571
+ {
572
+ "epoch": 1.68,
573
+ "learning_rate": 9.821712113759773e-06,
574
+ "loss": 0.149,
575
+ "step": 930
576
+ },
577
+ {
578
+ "epoch": 1.7,
579
+ "learning_rate": 9.813336407100558e-06,
580
+ "loss": 0.1339,
581
+ "step": 940
582
+ },
583
+ {
584
+ "epoch": 1.72,
585
+ "learning_rate": 9.80477218926238e-06,
586
+ "loss": 0.1342,
587
+ "step": 950
588
+ },
589
+ {
590
+ "epoch": 1.74,
591
+ "learning_rate": 9.796019795657244e-06,
592
+ "loss": 0.1425,
593
+ "step": 960
594
+ },
595
+ {
596
+ "epoch": 1.76,
597
+ "learning_rate": 9.787079569066933e-06,
598
+ "loss": 0.1223,
599
+ "step": 970
600
+ },
601
+ {
602
+ "epoch": 1.77,
603
+ "learning_rate": 9.777951859629584e-06,
604
+ "loss": 0.1262,
605
+ "step": 980
606
+ },
607
+ {
608
+ "epoch": 1.79,
609
+ "learning_rate": 9.768637024825985e-06,
610
+ "loss": 0.1206,
611
+ "step": 990
612
+ },
613
+ {
614
+ "epoch": 1.81,
615
+ "learning_rate": 9.759135429465556e-06,
616
+ "loss": 0.1106,
617
+ "step": 1000
618
+ },
619
+ {
620
+ "epoch": 1.83,
621
+ "learning_rate": 9.749447445672081e-06,
622
+ "loss": 0.1214,
623
+ "step": 1010
624
+ },
625
+ {
626
+ "epoch": 1.85,
627
+ "learning_rate": 9.739573452869123e-06,
628
+ "loss": 0.119,
629
+ "step": 1020
630
+ },
631
+ {
632
+ "epoch": 1.86,
633
+ "learning_rate": 9.729513837765166e-06,
634
+ "loss": 0.1193,
635
+ "step": 1030
636
+ },
637
+ {
638
+ "epoch": 1.88,
639
+ "learning_rate": 9.719268994338466e-06,
640
+ "loss": 0.1407,
641
+ "step": 1040
642
+ },
643
+ {
644
+ "epoch": 1.9,
645
+ "learning_rate": 9.708839323821635e-06,
646
+ "loss": 0.1281,
647
+ "step": 1050
648
+ },
649
+ {
650
+ "epoch": 1.92,
651
+ "learning_rate": 9.698225234685909e-06,
652
+ "loss": 0.1402,
653
+ "step": 1060
654
+ },
655
+ {
656
+ "epoch": 1.94,
657
+ "learning_rate": 9.687427142625163e-06,
658
+ "loss": 0.1214,
659
+ "step": 1070
660
+ },
661
+ {
662
+ "epoch": 1.96,
663
+ "learning_rate": 9.676445470539624e-06,
664
+ "loss": 0.1193,
665
+ "step": 1080
666
+ },
667
+ {
668
+ "epoch": 1.97,
669
+ "learning_rate": 9.665280648519311e-06,
670
+ "loss": 0.154,
671
+ "step": 1090
672
+ },
673
+ {
674
+ "epoch": 1.99,
675
+ "learning_rate": 9.653933113827196e-06,
676
+ "loss": 0.1309,
677
+ "step": 1100
678
+ },
679
+ {
680
+ "epoch": 2.0,
681
+ "eval_loss": 0.43477532267570496,
682
+ "eval_runtime": 248.2664,
683
+ "eval_samples_per_second": 7.915,
684
+ "eval_steps_per_second": 3.959,
685
+ "eval_wer": 0.1628673525968825,
686
+ "step": 1104
687
+ },
688
+ {
689
+ "epoch": 2.01,
690
+ "learning_rate": 9.642403310882065e-06,
691
+ "loss": 0.146,
692
+ "step": 1110
693
+ },
694
+ {
695
+ "epoch": 2.03,
696
+ "learning_rate": 9.630691691241124e-06,
697
+ "loss": 0.1305,
698
+ "step": 1120
699
+ },
700
+ {
701
+ "epoch": 2.05,
702
+ "learning_rate": 9.61879871358232e-06,
703
+ "loss": 0.1247,
704
+ "step": 1130
705
+ },
706
+ {
707
+ "epoch": 2.07,
708
+ "learning_rate": 9.606724843686355e-06,
709
+ "loss": 0.1237,
710
+ "step": 1140
711
+ },
712
+ {
713
+ "epoch": 2.08,
714
+ "learning_rate": 9.594470554418464e-06,
715
+ "loss": 0.1182,
716
+ "step": 1150
717
+ },
718
+ {
719
+ "epoch": 2.1,
720
+ "learning_rate": 9.582036325709892e-06,
721
+ "loss": 0.1338,
722
+ "step": 1160
723
+ },
724
+ {
725
+ "epoch": 2.12,
726
+ "learning_rate": 9.569422644539086e-06,
727
+ "loss": 0.1144,
728
+ "step": 1170
729
+ },
730
+ {
731
+ "epoch": 2.14,
732
+ "learning_rate": 9.55663000491264e-06,
733
+ "loss": 0.11,
734
+ "step": 1180
735
+ },
736
+ {
737
+ "epoch": 2.16,
738
+ "learning_rate": 9.543658907845937e-06,
739
+ "loss": 0.1153,
740
+ "step": 1190
741
+ },
742
+ {
743
+ "epoch": 2.17,
744
+ "learning_rate": 9.530509861343524e-06,
745
+ "loss": 0.1178,
746
+ "step": 1200
747
+ },
748
+ {
749
+ "epoch": 2.19,
750
+ "learning_rate": 9.51718338037923e-06,
751
+ "loss": 0.13,
752
+ "step": 1210
753
+ },
754
+ {
755
+ "epoch": 2.21,
756
+ "learning_rate": 9.50367998687598e-06,
757
+ "loss": 0.1302,
758
+ "step": 1220
759
+ },
760
+ {
761
+ "epoch": 2.23,
762
+ "learning_rate": 9.49000020968537e-06,
763
+ "loss": 0.1282,
764
+ "step": 1230
765
+ },
766
+ {
767
+ "epoch": 2.25,
768
+ "learning_rate": 9.476144584566947e-06,
769
+ "loss": 0.1275,
770
+ "step": 1240
771
+ },
772
+ {
773
+ "epoch": 2.26,
774
+ "learning_rate": 9.462113654167222e-06,
775
+ "loss": 0.1282,
776
+ "step": 1250
777
+ },
778
+ {
779
+ "epoch": 2.28,
780
+ "learning_rate": 9.447907967998431e-06,
781
+ "loss": 0.12,
782
+ "step": 1260
783
+ },
784
+ {
785
+ "epoch": 2.3,
786
+ "learning_rate": 9.433528082416999e-06,
787
+ "loss": 0.1156,
788
+ "step": 1270
789
+ },
790
+ {
791
+ "epoch": 2.32,
792
+ "learning_rate": 9.418974560601761e-06,
793
+ "loss": 0.1073,
794
+ "step": 1280
795
+ },
796
+ {
797
+ "epoch": 2.34,
798
+ "learning_rate": 9.404247972531897e-06,
799
+ "loss": 0.1139,
800
+ "step": 1290
801
+ },
802
+ {
803
+ "epoch": 2.35,
804
+ "learning_rate": 9.389348894964624e-06,
805
+ "loss": 0.1262,
806
+ "step": 1300
807
+ },
808
+ {
809
+ "epoch": 2.37,
810
+ "learning_rate": 9.374277911412586e-06,
811
+ "loss": 0.1226,
812
+ "step": 1310
813
+ },
814
+ {
815
+ "epoch": 2.39,
816
+ "learning_rate": 9.359035612121018e-06,
817
+ "loss": 0.1027,
818
+ "step": 1320
819
+ },
820
+ {
821
+ "epoch": 2.41,
822
+ "learning_rate": 9.343622594044626e-06,
823
+ "loss": 0.121,
824
+ "step": 1330
825
+ },
826
+ {
827
+ "epoch": 2.43,
828
+ "learning_rate": 9.328039460824204e-06,
829
+ "loss": 0.1153,
830
+ "step": 1340
831
+ },
832
+ {
833
+ "epoch": 2.45,
834
+ "learning_rate": 9.31228682276299e-06,
835
+ "loss": 0.126,
836
+ "step": 1350
837
+ },
838
+ {
839
+ "epoch": 2.46,
840
+ "learning_rate": 9.296365296802778e-06,
841
+ "loss": 0.1183,
842
+ "step": 1360
843
+ },
844
+ {
845
+ "epoch": 2.48,
846
+ "learning_rate": 9.280275506499736e-06,
847
+ "loss": 0.1305,
848
+ "step": 1370
849
+ },
850
+ {
851
+ "epoch": 2.5,
852
+ "learning_rate": 9.264018082000007e-06,
853
+ "loss": 0.1311,
854
+ "step": 1380
855
+ },
856
+ {
857
+ "epoch": 2.52,
858
+ "learning_rate": 9.247593660015001e-06,
859
+ "loss": 0.1342,
860
+ "step": 1390
861
+ },
862
+ {
863
+ "epoch": 2.54,
864
+ "learning_rate": 9.231002883796495e-06,
865
+ "loss": 0.1285,
866
+ "step": 1400
867
+ },
868
+ {
869
+ "epoch": 2.55,
870
+ "learning_rate": 9.21424640311141e-06,
871
+ "loss": 0.136,
872
+ "step": 1410
873
+ },
874
+ {
875
+ "epoch": 2.57,
876
+ "learning_rate": 9.197324874216372e-06,
877
+ "loss": 0.1104,
878
+ "step": 1420
879
+ },
880
+ {
881
+ "epoch": 2.59,
882
+ "learning_rate": 9.18023895983202e-06,
883
+ "loss": 0.1212,
884
+ "step": 1430
885
+ },
886
+ {
887
+ "epoch": 2.61,
888
+ "learning_rate": 9.162989329117037e-06,
889
+ "loss": 0.1107,
890
+ "step": 1440
891
+ },
892
+ {
893
+ "epoch": 2.63,
894
+ "learning_rate": 9.145576657641954e-06,
895
+ "loss": 0.0921,
896
+ "step": 1450
897
+ },
898
+ {
899
+ "epoch": 2.64,
900
+ "learning_rate": 9.128001627362683e-06,
901
+ "loss": 0.1335,
902
+ "step": 1460
903
+ },
904
+ {
905
+ "epoch": 2.66,
906
+ "learning_rate": 9.110264926593817e-06,
907
+ "loss": 0.1136,
908
+ "step": 1470
909
+ },
910
+ {
911
+ "epoch": 2.68,
912
+ "learning_rate": 9.092367249981663e-06,
913
+ "loss": 0.1241,
914
+ "step": 1480
915
+ },
916
+ {
917
+ "epoch": 2.7,
918
+ "learning_rate": 9.074309298477048e-06,
919
+ "loss": 0.1268,
920
+ "step": 1490
921
+ },
922
+ {
923
+ "epoch": 2.72,
924
+ "learning_rate": 9.056091779307854e-06,
925
+ "loss": 0.1104,
926
+ "step": 1500
927
+ },
928
+ {
929
+ "epoch": 2.73,
930
+ "learning_rate": 9.037715405951332e-06,
931
+ "loss": 0.1357,
932
+ "step": 1510
933
+ },
934
+ {
935
+ "epoch": 2.75,
936
+ "learning_rate": 9.019180898106148e-06,
937
+ "loss": 0.1361,
938
+ "step": 1520
939
+ },
940
+ {
941
+ "epoch": 2.77,
942
+ "learning_rate": 9.00048898166421e-06,
943
+ "loss": 0.1138,
944
+ "step": 1530
945
+ },
946
+ {
947
+ "epoch": 2.79,
948
+ "learning_rate": 8.981640388682222e-06,
949
+ "loss": 0.122,
950
+ "step": 1540
951
+ },
952
+ {
953
+ "epoch": 2.81,
954
+ "learning_rate": 8.962635857353028e-06,
955
+ "loss": 0.1273,
956
+ "step": 1550
957
+ },
958
+ {
959
+ "epoch": 2.83,
960
+ "learning_rate": 8.943476131976694e-06,
961
+ "loss": 0.1248,
962
+ "step": 1560
963
+ },
964
+ {
965
+ "epoch": 2.84,
966
+ "learning_rate": 8.92416196293136e-06,
967
+ "loss": 0.1027,
968
+ "step": 1570
969
+ },
970
+ {
971
+ "epoch": 2.86,
972
+ "learning_rate": 8.904694106643852e-06,
973
+ "loss": 0.1076,
974
+ "step": 1580
975
+ },
976
+ {
977
+ "epoch": 2.88,
978
+ "learning_rate": 8.885073325560054e-06,
979
+ "loss": 0.1234,
980
+ "step": 1590
981
+ },
982
+ {
983
+ "epoch": 2.9,
984
+ "learning_rate": 8.865300388115053e-06,
985
+ "loss": 0.1206,
986
+ "step": 1600
987
+ },
988
+ {
989
+ "epoch": 2.92,
990
+ "learning_rate": 8.84537606870304e-06,
991
+ "loss": 0.1253,
992
+ "step": 1610
993
+ },
994
+ {
995
+ "epoch": 2.93,
996
+ "learning_rate": 8.825301147646978e-06,
997
+ "loss": 0.1199,
998
+ "step": 1620
999
+ },
1000
+ {
1001
+ "epoch": 2.95,
1002
+ "learning_rate": 8.805076411168054e-06,
1003
+ "loss": 0.1119,
1004
+ "step": 1630
1005
+ },
1006
+ {
1007
+ "epoch": 2.97,
1008
+ "learning_rate": 8.784702651354866e-06,
1009
+ "loss": 0.1058,
1010
+ "step": 1640
1011
+ },
1012
+ {
1013
+ "epoch": 2.99,
1014
+ "learning_rate": 8.764180666132424e-06,
1015
+ "loss": 0.1237,
1016
+ "step": 1650
1017
+ },
1018
+ {
1019
+ "epoch": 3.0,
1020
+ "eval_loss": 0.4611044228076935,
1021
+ "eval_runtime": 252.1565,
1022
+ "eval_samples_per_second": 7.793,
1023
+ "eval_steps_per_second": 3.898,
1024
+ "eval_wer": 0.1604337379089397,
1025
+ "step": 1656
1026
+ },
1027
+ {
1028
+ "epoch": 3.01,
1029
+ "learning_rate": 8.743511259230885e-06,
1030
+ "loss": 0.1404,
1031
+ "step": 1660
1032
+ },
1033
+ {
1034
+ "epoch": 3.03,
1035
+ "learning_rate": 8.72269524015408e-06,
1036
+ "loss": 0.1113,
1037
+ "step": 1670
1038
+ },
1039
+ {
1040
+ "epoch": 3.04,
1041
+ "learning_rate": 8.701733424147808e-06,
1042
+ "loss": 0.1023,
1043
+ "step": 1680
1044
+ },
1045
+ {
1046
+ "epoch": 3.06,
1047
+ "learning_rate": 8.68062663216792e-06,
1048
+ "loss": 0.1219,
1049
+ "step": 1690
1050
+ },
1051
+ {
1052
+ "epoch": 3.08,
1053
+ "learning_rate": 8.65937569084814e-06,
1054
+ "loss": 0.1072,
1055
+ "step": 1700
1056
+ },
1057
+ {
1058
+ "epoch": 3.1,
1059
+ "learning_rate": 8.637981432467719e-06,
1060
+ "loss": 0.1304,
1061
+ "step": 1710
1062
+ },
1063
+ {
1064
+ "epoch": 3.12,
1065
+ "learning_rate": 8.616444694918821e-06,
1066
+ "loss": 0.0999,
1067
+ "step": 1720
1068
+ },
1069
+ {
1070
+ "epoch": 3.13,
1071
+ "learning_rate": 8.594766321673723e-06,
1072
+ "loss": 0.113,
1073
+ "step": 1730
1074
+ },
1075
+ {
1076
+ "epoch": 3.15,
1077
+ "learning_rate": 8.572947161751757e-06,
1078
+ "loss": 0.1113,
1079
+ "step": 1740
1080
+ },
1081
+ {
1082
+ "epoch": 3.17,
1083
+ "learning_rate": 8.550988069686085e-06,
1084
+ "loss": 0.1135,
1085
+ "step": 1750
1086
+ },
1087
+ {
1088
+ "epoch": 3.19,
1089
+ "learning_rate": 8.528889905490217e-06,
1090
+ "loss": 0.1201,
1091
+ "step": 1760
1092
+ },
1093
+ {
1094
+ "epoch": 3.21,
1095
+ "learning_rate": 8.506653534624327e-06,
1096
+ "loss": 0.1119,
1097
+ "step": 1770
1098
+ },
1099
+ {
1100
+ "epoch": 3.22,
1101
+ "learning_rate": 8.484279827961372e-06,
1102
+ "loss": 0.1103,
1103
+ "step": 1780
1104
+ },
1105
+ {
1106
+ "epoch": 3.24,
1107
+ "learning_rate": 8.461769661752968e-06,
1108
+ "loss": 0.102,
1109
+ "step": 1790
1110
+ },
1111
+ {
1112
+ "epoch": 3.26,
1113
+ "learning_rate": 8.439123917595086e-06,
1114
+ "loss": 0.1319,
1115
+ "step": 1800
1116
+ },
1117
+ {
1118
+ "epoch": 3.28,
1119
+ "learning_rate": 8.416343482393514e-06,
1120
+ "loss": 0.1242,
1121
+ "step": 1810
1122
+ },
1123
+ {
1124
+ "epoch": 3.3,
1125
+ "learning_rate": 8.393429248329131e-06,
1126
+ "loss": 0.1322,
1127
+ "step": 1820
1128
+ },
1129
+ {
1130
+ "epoch": 3.31,
1131
+ "learning_rate": 8.370382112822957e-06,
1132
+ "loss": 0.1248,
1133
+ "step": 1830
1134
+ },
1135
+ {
1136
+ "epoch": 3.33,
1137
+ "learning_rate": 8.347202978501013e-06,
1138
+ "loss": 0.1116,
1139
+ "step": 1840
1140
+ },
1141
+ {
1142
+ "epoch": 3.35,
1143
+ "learning_rate": 8.323892753158969e-06,
1144
+ "loss": 0.123,
1145
+ "step": 1850
1146
+ },
1147
+ {
1148
+ "epoch": 3.37,
1149
+ "learning_rate": 8.300452349726585e-06,
1150
+ "loss": 0.1335,
1151
+ "step": 1860
1152
+ },
1153
+ {
1154
+ "epoch": 3.39,
1155
+ "learning_rate": 8.276882686231961e-06,
1156
+ "loss": 0.1048,
1157
+ "step": 1870
1158
+ },
1159
+ {
1160
+ "epoch": 3.41,
1161
+ "learning_rate": 8.253184685765591e-06,
1162
+ "loss": 0.1112,
1163
+ "step": 1880
1164
+ },
1165
+ {
1166
+ "epoch": 3.42,
1167
+ "learning_rate": 8.22935927644419e-06,
1168
+ "loss": 0.1093,
1169
+ "step": 1890
1170
+ },
1171
+ {
1172
+ "epoch": 3.44,
1173
+ "learning_rate": 8.205407391374372e-06,
1174
+ "loss": 0.1108,
1175
+ "step": 1900
1176
+ },
1177
+ {
1178
+ "epoch": 3.46,
1179
+ "learning_rate": 8.181329968616077e-06,
1180
+ "loss": 0.1261,
1181
+ "step": 1910
1182
+ },
1183
+ {
1184
+ "epoch": 3.48,
1185
+ "learning_rate": 8.157127951145855e-06,
1186
+ "loss": 0.1012,
1187
+ "step": 1920
1188
+ },
1189
+ {
1190
+ "epoch": 3.5,
1191
+ "learning_rate": 8.132802286819925e-06,
1192
+ "loss": 0.1117,
1193
+ "step": 1930
1194
+ },
1195
+ {
1196
+ "epoch": 3.51,
1197
+ "learning_rate": 8.108353928337056e-06,
1198
+ "loss": 0.0956,
1199
+ "step": 1940
1200
+ },
1201
+ {
1202
+ "epoch": 3.53,
1203
+ "learning_rate": 8.083783833201249e-06,
1204
+ "loss": 0.1005,
1205
+ "step": 1950
1206
+ },
1207
+ {
1208
+ "epoch": 3.55,
1209
+ "learning_rate": 8.059092963684246e-06,
1210
+ "loss": 0.1093,
1211
+ "step": 1960
1212
+ },
1213
+ {
1214
+ "epoch": 3.57,
1215
+ "learning_rate": 8.034282286787837e-06,
1216
+ "loss": 0.1173,
1217
+ "step": 1970
1218
+ },
1219
+ {
1220
+ "epoch": 3.59,
1221
+ "learning_rate": 8.00935277420599e-06,
1222
+ "loss": 0.1014,
1223
+ "step": 1980
1224
+ },
1225
+ {
1226
+ "epoch": 3.6,
1227
+ "learning_rate": 7.984305402286793e-06,
1228
+ "loss": 0.1186,
1229
+ "step": 1990
1230
+ },
1231
+ {
1232
+ "epoch": 3.62,
1233
+ "learning_rate": 7.959141151994223e-06,
1234
+ "loss": 0.1099,
1235
+ "step": 2000
1236
+ },
1237
+ {
1238
+ "epoch": 3.64,
1239
+ "learning_rate": 7.933861008869718e-06,
1240
+ "loss": 0.0937,
1241
+ "step": 2010
1242
+ },
1243
+ {
1244
+ "epoch": 3.66,
1245
+ "learning_rate": 7.908465962993584e-06,
1246
+ "loss": 0.1138,
1247
+ "step": 2020
1248
+ },
1249
+ {
1250
+ "epoch": 3.68,
1251
+ "learning_rate": 7.88295700894622e-06,
1252
+ "loss": 0.0977,
1253
+ "step": 2030
1254
+ },
1255
+ {
1256
+ "epoch": 3.69,
1257
+ "learning_rate": 7.85733514576916e-06,
1258
+ "loss": 0.094,
1259
+ "step": 2040
1260
+ },
1261
+ {
1262
+ "epoch": 3.71,
1263
+ "learning_rate": 7.831601376925957e-06,
1264
+ "loss": 0.1073,
1265
+ "step": 2050
1266
+ },
1267
+ {
1268
+ "epoch": 3.73,
1269
+ "learning_rate": 7.80575671026287e-06,
1270
+ "loss": 0.1082,
1271
+ "step": 2060
1272
+ },
1273
+ {
1274
+ "epoch": 3.75,
1275
+ "learning_rate": 7.779802157969402e-06,
1276
+ "loss": 0.0955,
1277
+ "step": 2070
1278
+ },
1279
+ {
1280
+ "epoch": 3.77,
1281
+ "learning_rate": 7.753738736538654e-06,
1282
+ "loss": 0.1182,
1283
+ "step": 2080
1284
+ },
1285
+ {
1286
+ "epoch": 3.79,
1287
+ "learning_rate": 7.727567466727517e-06,
1288
+ "loss": 0.1198,
1289
+ "step": 2090
1290
+ },
1291
+ {
1292
+ "epoch": 3.8,
1293
+ "learning_rate": 7.701289373516692e-06,
1294
+ "loss": 0.1086,
1295
+ "step": 2100
1296
+ },
1297
+ {
1298
+ "epoch": 3.82,
1299
+ "learning_rate": 7.67490548607055e-06,
1300
+ "loss": 0.1268,
1301
+ "step": 2110
1302
+ },
1303
+ {
1304
+ "epoch": 3.84,
1305
+ "learning_rate": 7.648416837696821e-06,
1306
+ "loss": 0.1117,
1307
+ "step": 2120
1308
+ },
1309
+ {
1310
+ "epoch": 3.86,
1311
+ "learning_rate": 7.621824465806134e-06,
1312
+ "loss": 0.1051,
1313
+ "step": 2130
1314
+ },
1315
+ {
1316
+ "epoch": 3.88,
1317
+ "learning_rate": 7.595129411871376e-06,
1318
+ "loss": 0.1251,
1319
+ "step": 2140
1320
+ },
1321
+ {
1322
+ "epoch": 3.89,
1323
+ "learning_rate": 7.568332721386913e-06,
1324
+ "loss": 0.1011,
1325
+ "step": 2150
1326
+ },
1327
+ {
1328
+ "epoch": 3.91,
1329
+ "learning_rate": 7.541435443827638e-06,
1330
+ "loss": 0.1136,
1331
+ "step": 2160
1332
+ },
1333
+ {
1334
+ "epoch": 3.93,
1335
+ "learning_rate": 7.5144386326078715e-06,
1336
+ "loss": 0.105,
1337
+ "step": 2170
1338
+ },
1339
+ {
1340
+ "epoch": 3.95,
1341
+ "learning_rate": 7.487343345040109e-06,
1342
+ "loss": 0.0933,
1343
+ "step": 2180
1344
+ },
1345
+ {
1346
+ "epoch": 3.97,
1347
+ "learning_rate": 7.460150642293601e-06,
1348
+ "loss": 0.1097,
1349
+ "step": 2190
1350
+ },
1351
+ {
1352
+ "epoch": 3.98,
1353
+ "learning_rate": 7.432861589352804e-06,
1354
+ "loss": 0.1056,
1355
+ "step": 2200
1356
+ },
1357
+ {
1358
+ "epoch": 4.0,
1359
+ "eval_loss": 0.454052597284317,
1360
+ "eval_runtime": 253.3273,
1361
+ "eval_samples_per_second": 7.757,
1362
+ "eval_steps_per_second": 3.88,
1363
+ "eval_wer": 0.15738401823670753,
1364
+ "step": 2208
1365
+ },
1366
+ {
1367
+ "epoch": 4.0,
1368
+ "learning_rate": 7.40547725497567e-06,
1369
+ "loss": 0.1269,
1370
+ "step": 2210
1371
+ },
1372
+ {
1373
+ "epoch": 4.02,
1374
+ "learning_rate": 7.3779987116517745e-06,
1375
+ "loss": 0.1024,
1376
+ "step": 2220
1377
+ },
1378
+ {
1379
+ "epoch": 4.04,
1380
+ "learning_rate": 7.35042703556034e-06,
1381
+ "loss": 0.094,
1382
+ "step": 2230
1383
+ },
1384
+ {
1385
+ "epoch": 4.06,
1386
+ "learning_rate": 7.322763306528063e-06,
1387
+ "loss": 0.0975,
1388
+ "step": 2240
1389
+ },
1390
+ {
1391
+ "epoch": 4.08,
1392
+ "learning_rate": 7.2950086079868375e-06,
1393
+ "loss": 0.1024,
1394
+ "step": 2250
1395
+ },
1396
+ {
1397
+ "epoch": 4.09,
1398
+ "learning_rate": 7.267164026931317e-06,
1399
+ "loss": 0.1076,
1400
+ "step": 2260
1401
+ },
1402
+ {
1403
+ "epoch": 4.11,
1404
+ "learning_rate": 7.239230653876345e-06,
1405
+ "loss": 0.1134,
1406
+ "step": 2270
1407
+ },
1408
+ {
1409
+ "epoch": 4.13,
1410
+ "learning_rate": 7.2112095828142484e-06,
1411
+ "loss": 0.0762,
1412
+ "step": 2280
1413
+ },
1414
+ {
1415
+ "epoch": 4.15,
1416
+ "learning_rate": 7.1831019111719855e-06,
1417
+ "loss": 0.0982,
1418
+ "step": 2290
1419
+ },
1420
+ {
1421
+ "epoch": 4.17,
1422
+ "learning_rate": 7.154908739768173e-06,
1423
+ "loss": 0.1,
1424
+ "step": 2300
1425
+ },
1426
+ {
1427
+ "epoch": 4.18,
1428
+ "learning_rate": 7.126631172769968e-06,
1429
+ "loss": 0.1076,
1430
+ "step": 2310
1431
+ },
1432
+ {
1433
+ "epoch": 4.2,
1434
+ "learning_rate": 7.098270317649827e-06,
1435
+ "loss": 0.0846,
1436
+ "step": 2320
1437
+ },
1438
+ {
1439
+ "epoch": 4.22,
1440
+ "learning_rate": 7.06982728514213e-06,
1441
+ "loss": 0.0915,
1442
+ "step": 2330
1443
+ },
1444
+ {
1445
+ "epoch": 4.24,
1446
+ "learning_rate": 7.041303189199684e-06,
1447
+ "loss": 0.0962,
1448
+ "step": 2340
1449
+ },
1450
+ {
1451
+ "epoch": 4.26,
1452
+ "learning_rate": 7.012699146950088e-06,
1453
+ "loss": 0.0853,
1454
+ "step": 2350
1455
+ },
1456
+ {
1457
+ "epoch": 4.28,
1458
+ "learning_rate": 6.98401627865199e-06,
1459
+ "loss": 0.0997,
1460
+ "step": 2360
1461
+ },
1462
+ {
1463
+ "epoch": 4.29,
1464
+ "learning_rate": 6.955255707651207e-06,
1465
+ "loss": 0.0927,
1466
+ "step": 2370
1467
+ },
1468
+ {
1469
+ "epoch": 4.31,
1470
+ "learning_rate": 6.926418560336731e-06,
1471
+ "loss": 0.0992,
1472
+ "step": 2380
1473
+ },
1474
+ {
1475
+ "epoch": 4.33,
1476
+ "learning_rate": 6.897505966096616e-06,
1477
+ "loss": 0.1307,
1478
+ "step": 2390
1479
+ },
1480
+ {
1481
+ "epoch": 4.35,
1482
+ "learning_rate": 6.868519057273744e-06,
1483
+ "loss": 0.1065,
1484
+ "step": 2400
1485
+ },
1486
+ {
1487
+ "epoch": 4.37,
1488
+ "learning_rate": 6.839458969121485e-06,
1489
+ "loss": 0.1049,
1490
+ "step": 2410
1491
+ },
1492
+ {
1493
+ "epoch": 4.38,
1494
+ "learning_rate": 6.810326839759221e-06,
1495
+ "loss": 0.1004,
1496
+ "step": 2420
1497
+ },
1498
+ {
1499
+ "epoch": 4.4,
1500
+ "learning_rate": 6.7811238101277876e-06,
1501
+ "loss": 0.0894,
1502
+ "step": 2430
1503
+ },
1504
+ {
1505
+ "epoch": 4.42,
1506
+ "learning_rate": 6.751851023944781e-06,
1507
+ "loss": 0.1119,
1508
+ "step": 2440
1509
+ },
1510
+ {
1511
+ "epoch": 4.44,
1512
+ "learning_rate": 6.722509627659768e-06,
1513
+ "loss": 0.0798,
1514
+ "step": 2450
1515
+ },
1516
+ {
1517
+ "epoch": 4.46,
1518
+ "learning_rate": 6.693100770409381e-06,
1519
+ "loss": 0.1138,
1520
+ "step": 2460
1521
+ },
1522
+ {
1523
+ "epoch": 4.47,
1524
+ "learning_rate": 6.663625603972323e-06,
1525
+ "loss": 0.0846,
1526
+ "step": 2470
1527
+ },
1528
+ {
1529
+ "epoch": 4.49,
1530
+ "learning_rate": 6.634085282724252e-06,
1531
+ "loss": 0.1088,
1532
+ "step": 2480
1533
+ },
1534
+ {
1535
+ "epoch": 4.51,
1536
+ "learning_rate": 6.604480963592567e-06,
1537
+ "loss": 0.094,
1538
+ "step": 2490
1539
+ },
1540
+ {
1541
+ "epoch": 4.53,
1542
+ "learning_rate": 6.574813806011107e-06,
1543
+ "loss": 0.1107,
1544
+ "step": 2500
1545
+ },
1546
+ {
1547
+ "epoch": 4.55,
1548
+ "learning_rate": 6.545084971874738e-06,
1549
+ "loss": 0.1171,
1550
+ "step": 2510
1551
+ },
1552
+ {
1553
+ "epoch": 4.56,
1554
+ "learning_rate": 6.515295625493844e-06,
1555
+ "loss": 0.0954,
1556
+ "step": 2520
1557
+ },
1558
+ {
1559
+ "epoch": 4.58,
1560
+ "learning_rate": 6.485446933548735e-06,
1561
+ "loss": 0.0925,
1562
+ "step": 2530
1563
+ },
1564
+ {
1565
+ "epoch": 4.6,
1566
+ "learning_rate": 6.45554006504395e-06,
1567
+ "loss": 0.0928,
1568
+ "step": 2540
1569
+ },
1570
+ {
1571
+ "epoch": 4.62,
1572
+ "learning_rate": 6.425576191262476e-06,
1573
+ "loss": 0.1076,
1574
+ "step": 2550
1575
+ },
1576
+ {
1577
+ "epoch": 4.64,
1578
+ "learning_rate": 6.395556485719873e-06,
1579
+ "loss": 0.0999,
1580
+ "step": 2560
1581
+ },
1582
+ {
1583
+ "epoch": 4.65,
1584
+ "learning_rate": 6.3654821241183155e-06,
1585
+ "loss": 0.1252,
1586
+ "step": 2570
1587
+ },
1588
+ {
1589
+ "epoch": 4.67,
1590
+ "learning_rate": 6.33535428430055e-06,
1591
+ "loss": 0.1073,
1592
+ "step": 2580
1593
+ },
1594
+ {
1595
+ "epoch": 4.69,
1596
+ "learning_rate": 6.305174146203757e-06,
1597
+ "loss": 0.0933,
1598
+ "step": 2590
1599
+ },
1600
+ {
1601
+ "epoch": 4.71,
1602
+ "learning_rate": 6.27494289181335e-06,
1603
+ "loss": 0.0947,
1604
+ "step": 2600
1605
+ },
1606
+ {
1607
+ "epoch": 4.73,
1608
+ "learning_rate": 6.244661705116678e-06,
1609
+ "loss": 0.1053,
1610
+ "step": 2610
1611
+ },
1612
+ {
1613
+ "epoch": 4.75,
1614
+ "learning_rate": 6.214331772056652e-06,
1615
+ "loss": 0.0978,
1616
+ "step": 2620
1617
+ },
1618
+ {
1619
+ "epoch": 4.76,
1620
+ "learning_rate": 6.183954280485307e-06,
1621
+ "loss": 0.0874,
1622
+ "step": 2630
1623
+ },
1624
+ {
1625
+ "epoch": 4.78,
1626
+ "learning_rate": 6.153530420117275e-06,
1627
+ "loss": 0.1032,
1628
+ "step": 2640
1629
+ },
1630
+ {
1631
+ "epoch": 4.8,
1632
+ "learning_rate": 6.123061382483189e-06,
1633
+ "loss": 0.1009,
1634
+ "step": 2650
1635
+ },
1636
+ {
1637
+ "epoch": 4.82,
1638
+ "learning_rate": 6.092548360883024e-06,
1639
+ "loss": 0.1125,
1640
+ "step": 2660
1641
+ },
1642
+ {
1643
+ "epoch": 4.84,
1644
+ "learning_rate": 6.061992550339357e-06,
1645
+ "loss": 0.0917,
1646
+ "step": 2670
1647
+ },
1648
+ {
1649
+ "epoch": 4.85,
1650
+ "learning_rate": 6.031395147550566e-06,
1651
+ "loss": 0.0977,
1652
+ "step": 2680
1653
+ },
1654
+ {
1655
+ "epoch": 4.87,
1656
+ "learning_rate": 6.00075735084396e-06,
1657
+ "loss": 0.1141,
1658
+ "step": 2690
1659
+ },
1660
+ {
1661
+ "epoch": 4.89,
1662
+ "learning_rate": 5.970080360128851e-06,
1663
+ "loss": 0.0985,
1664
+ "step": 2700
1665
+ },
1666
+ {
1667
+ "epoch": 4.91,
1668
+ "learning_rate": 5.939365376849564e-06,
1669
+ "loss": 0.1162,
1670
+ "step": 2710
1671
+ },
1672
+ {
1673
+ "epoch": 4.93,
1674
+ "learning_rate": 5.908613603938367e-06,
1675
+ "loss": 0.0979,
1676
+ "step": 2720
1677
+ },
1678
+ {
1679
+ "epoch": 4.94,
1680
+ "learning_rate": 5.8778262457683776e-06,
1681
+ "loss": 0.0816,
1682
+ "step": 2730
1683
+ },
1684
+ {
1685
+ "epoch": 4.96,
1686
+ "learning_rate": 5.847004508106386e-06,
1687
+ "loss": 0.109,
1688
+ "step": 2740
1689
+ },
1690
+ {
1691
+ "epoch": 4.98,
1692
+ "learning_rate": 5.816149598065632e-06,
1693
+ "loss": 0.105,
1694
+ "step": 2750
1695
+ },
1696
+ {
1697
+ "epoch": 5.0,
1698
+ "learning_rate": 5.78526272405853e-06,
1699
+ "loss": 0.1236,
1700
+ "step": 2760
1701
+ },
1702
+ {
1703
+ "epoch": 5.0,
1704
+ "eval_loss": 0.46689143776893616,
1705
+ "eval_runtime": 244.9078,
1706
+ "eval_samples_per_second": 8.023,
1707
+ "eval_steps_per_second": 4.014,
1708
+ "eval_wer": 0.1603413221612963,
1709
+ "step": 2760
1710
+ },
1711
+ {
1712
+ "epoch": 5.02,
1713
+ "learning_rate": 5.754345095749342e-06,
1714
+ "loss": 0.1158,
1715
+ "step": 2770
1716
+ },
1717
+ {
1718
+ "epoch": 5.04,
1719
+ "learning_rate": 5.723397924006802e-06,
1720
+ "loss": 0.1168,
1721
+ "step": 2780
1722
+ },
1723
+ {
1724
+ "epoch": 5.05,
1725
+ "learning_rate": 5.692422420856694e-06,
1726
+ "loss": 0.1023,
1727
+ "step": 2790
1728
+ },
1729
+ {
1730
+ "epoch": 5.07,
1731
+ "learning_rate": 5.66141979943438e-06,
1732
+ "loss": 0.1093,
1733
+ "step": 2800
1734
+ },
1735
+ {
1736
+ "epoch": 5.09,
1737
+ "learning_rate": 5.6303912739373e-06,
1738
+ "loss": 0.1154,
1739
+ "step": 2810
1740
+ },
1741
+ {
1742
+ "epoch": 5.11,
1743
+ "learning_rate": 5.5993380595773995e-06,
1744
+ "loss": 0.1025,
1745
+ "step": 2820
1746
+ },
1747
+ {
1748
+ "epoch": 5.13,
1749
+ "learning_rate": 5.5682613725335534e-06,
1750
+ "loss": 0.1118,
1751
+ "step": 2830
1752
+ },
1753
+ {
1754
+ "epoch": 5.14,
1755
+ "learning_rate": 5.537162429903928e-06,
1756
+ "loss": 0.106,
1757
+ "step": 2840
1758
+ },
1759
+ {
1760
+ "epoch": 5.16,
1761
+ "learning_rate": 5.5060424496583145e-06,
1762
+ "loss": 0.1161,
1763
+ "step": 2850
1764
+ },
1765
+ {
1766
+ "epoch": 5.18,
1767
+ "learning_rate": 5.474902650590428e-06,
1768
+ "loss": 0.1126,
1769
+ "step": 2860
1770
+ },
1771
+ {
1772
+ "epoch": 5.2,
1773
+ "learning_rate": 5.443744252270178e-06,
1774
+ "loss": 0.1083,
1775
+ "step": 2870
1776
+ },
1777
+ {
1778
+ "epoch": 5.22,
1779
+ "learning_rate": 5.412568474995896e-06,
1780
+ "loss": 0.1042,
1781
+ "step": 2880
1782
+ },
1783
+ {
1784
+ "epoch": 5.24,
1785
+ "learning_rate": 5.381376539746553e-06,
1786
+ "loss": 0.1085,
1787
+ "step": 2890
1788
+ },
1789
+ {
1790
+ "epoch": 5.25,
1791
+ "learning_rate": 5.350169668133937e-06,
1792
+ "loss": 0.0952,
1793
+ "step": 2900
1794
+ },
1795
+ {
1796
+ "epoch": 5.27,
1797
+ "learning_rate": 5.318949082354808e-06,
1798
+ "loss": 0.1085,
1799
+ "step": 2910
1800
+ },
1801
+ {
1802
+ "epoch": 5.29,
1803
+ "learning_rate": 5.287716005143028e-06,
1804
+ "loss": 0.106,
1805
+ "step": 2920
1806
+ },
1807
+ {
1808
+ "epoch": 5.31,
1809
+ "learning_rate": 5.256471659721685e-06,
1810
+ "loss": 0.1054,
1811
+ "step": 2930
1812
+ },
1813
+ {
1814
+ "epoch": 5.33,
1815
+ "learning_rate": 5.225217269755176e-06,
1816
+ "loss": 0.0925,
1817
+ "step": 2940
1818
+ },
1819
+ {
1820
+ "epoch": 5.34,
1821
+ "learning_rate": 5.193954059301285e-06,
1822
+ "loss": 0.1117,
1823
+ "step": 2950
1824
+ },
1825
+ {
1826
+ "epoch": 5.36,
1827
+ "learning_rate": 5.162683252763246e-06,
1828
+ "loss": 0.1271,
1829
+ "step": 2960
1830
+ },
1831
+ {
1832
+ "epoch": 5.38,
1833
+ "learning_rate": 5.131406074841787e-06,
1834
+ "loss": 0.1072,
1835
+ "step": 2970
1836
+ },
1837
+ {
1838
+ "epoch": 5.4,
1839
+ "learning_rate": 5.100123750487168e-06,
1840
+ "loss": 0.0969,
1841
+ "step": 2980
1842
+ },
1843
+ {
1844
+ "epoch": 5.42,
1845
+ "learning_rate": 5.068837504851208e-06,
1846
+ "loss": 0.1127,
1847
+ "step": 2990
1848
+ },
1849
+ {
1850
+ "epoch": 5.43,
1851
+ "learning_rate": 5.0375485632392975e-06,
1852
+ "loss": 0.1013,
1853
+ "step": 3000
1854
+ },
1855
+ {
1856
+ "epoch": 5.45,
1857
+ "learning_rate": 5.006258151062413e-06,
1858
+ "loss": 0.1036,
1859
+ "step": 3010
1860
+ },
1861
+ {
1862
+ "epoch": 5.47,
1863
+ "learning_rate": 4.974967493789126e-06,
1864
+ "loss": 0.0939,
1865
+ "step": 3020
1866
+ },
1867
+ {
1868
+ "epoch": 5.49,
1869
+ "learning_rate": 4.943677816897609e-06,
1870
+ "loss": 0.0949,
1871
+ "step": 3030
1872
+ },
1873
+ {
1874
+ "epoch": 5.51,
1875
+ "learning_rate": 4.912390345827634e-06,
1876
+ "loss": 0.1036,
1877
+ "step": 3040
1878
+ },
1879
+ {
1880
+ "epoch": 5.52,
1881
+ "learning_rate": 4.881106305932585e-06,
1882
+ "loss": 0.1076,
1883
+ "step": 3050
1884
+ },
1885
+ {
1886
+ "epoch": 5.54,
1887
+ "learning_rate": 4.849826922431469e-06,
1888
+ "loss": 0.1034,
1889
+ "step": 3060
1890
+ },
1891
+ {
1892
+ "epoch": 5.56,
1893
+ "learning_rate": 4.818553420360927e-06,
1894
+ "loss": 0.1004,
1895
+ "step": 3070
1896
+ },
1897
+ {
1898
+ "epoch": 5.58,
1899
+ "learning_rate": 4.787287024527254e-06,
1900
+ "loss": 0.1096,
1901
+ "step": 3080
1902
+ },
1903
+ {
1904
+ "epoch": 5.6,
1905
+ "learning_rate": 4.756028959458438e-06,
1906
+ "loss": 0.09,
1907
+ "step": 3090
1908
+ },
1909
+ {
1910
+ "epoch": 5.62,
1911
+ "learning_rate": 4.724780449356199e-06,
1912
+ "loss": 0.1059,
1913
+ "step": 3100
1914
+ },
1915
+ {
1916
+ "epoch": 5.63,
1917
+ "learning_rate": 4.693542718048037e-06,
1918
+ "loss": 0.1041,
1919
+ "step": 3110
1920
+ },
1921
+ {
1922
+ "epoch": 5.65,
1923
+ "learning_rate": 4.662316988939314e-06,
1924
+ "loss": 0.1053,
1925
+ "step": 3120
1926
+ },
1927
+ {
1928
+ "epoch": 5.67,
1929
+ "learning_rate": 4.631104484965331e-06,
1930
+ "loss": 0.1033,
1931
+ "step": 3130
1932
+ },
1933
+ {
1934
+ "epoch": 5.69,
1935
+ "learning_rate": 4.599906428543435e-06,
1936
+ "loss": 0.0967,
1937
+ "step": 3140
1938
+ },
1939
+ {
1940
+ "epoch": 5.71,
1941
+ "learning_rate": 4.568724041525143e-06,
1942
+ "loss": 0.1109,
1943
+ "step": 3150
1944
+ },
1945
+ {
1946
+ "epoch": 5.72,
1947
+ "learning_rate": 4.537558545148292e-06,
1948
+ "loss": 0.1208,
1949
+ "step": 3160
1950
+ },
1951
+ {
1952
+ "epoch": 5.74,
1953
+ "learning_rate": 4.50641115998921e-06,
1954
+ "loss": 0.1004,
1955
+ "step": 3170
1956
+ },
1957
+ {
1958
+ "epoch": 5.76,
1959
+ "learning_rate": 4.475283105914905e-06,
1960
+ "loss": 0.1081,
1961
+ "step": 3180
1962
+ },
1963
+ {
1964
+ "epoch": 5.78,
1965
+ "learning_rate": 4.444175602035301e-06,
1966
+ "loss": 0.0996,
1967
+ "step": 3190
1968
+ },
1969
+ {
1970
+ "epoch": 5.8,
1971
+ "learning_rate": 4.4130898666554845e-06,
1972
+ "loss": 0.0956,
1973
+ "step": 3200
1974
+ },
1975
+ {
1976
+ "epoch": 5.81,
1977
+ "learning_rate": 4.382027117227993e-06,
1978
+ "loss": 0.1069,
1979
+ "step": 3210
1980
+ },
1981
+ {
1982
+ "epoch": 5.83,
1983
+ "learning_rate": 4.350988570305136e-06,
1984
+ "loss": 0.0872,
1985
+ "step": 3220
1986
+ },
1987
+ {
1988
+ "epoch": 5.85,
1989
+ "learning_rate": 4.3199754414913455e-06,
1990
+ "loss": 0.1086,
1991
+ "step": 3230
1992
+ },
1993
+ {
1994
+ "epoch": 5.87,
1995
+ "learning_rate": 4.288988945395572e-06,
1996
+ "loss": 0.1122,
1997
+ "step": 3240
1998
+ },
1999
+ {
2000
+ "epoch": 5.89,
2001
+ "learning_rate": 4.258030295583709e-06,
2002
+ "loss": 0.1044,
2003
+ "step": 3250
2004
+ },
2005
+ {
2006
+ "epoch": 5.9,
2007
+ "learning_rate": 4.227100704531075e-06,
2008
+ "loss": 0.1083,
2009
+ "step": 3260
2010
+ },
2011
+ {
2012
+ "epoch": 5.92,
2013
+ "learning_rate": 4.1962013835749145e-06,
2014
+ "loss": 0.1024,
2015
+ "step": 3270
2016
+ },
2017
+ {
2018
+ "epoch": 5.94,
2019
+ "learning_rate": 4.1653335428669675e-06,
2020
+ "loss": 0.1036,
2021
+ "step": 3280
2022
+ },
2023
+ {
2024
+ "epoch": 5.96,
2025
+ "learning_rate": 4.134498391326069e-06,
2026
+ "loss": 0.1133,
2027
+ "step": 3290
2028
+ },
2029
+ {
2030
+ "epoch": 5.98,
2031
+ "learning_rate": 4.1036971365908055e-06,
2032
+ "loss": 0.095,
2033
+ "step": 3300
2034
+ },
2035
+ {
2036
+ "epoch": 6.0,
2037
+ "learning_rate": 4.072930984972215e-06,
2038
+ "loss": 0.1118,
2039
+ "step": 3310
2040
+ },
2041
+ {
2042
+ "epoch": 6.0,
2043
+ "eval_loss": 0.46403050422668457,
2044
+ "eval_runtime": 245.9183,
2045
+ "eval_samples_per_second": 7.99,
2046
+ "eval_steps_per_second": 3.997,
2047
+ "eval_wer": 0.15673710800320376,
2048
+ "step": 3312
2049
+ },
2050
+ {
2051
+ "epoch": 6.01,
2052
+ "learning_rate": 4.042201141406547e-06,
2053
+ "loss": 0.1158,
2054
+ "step": 3320
2055
+ },
2056
+ {
2057
+ "epoch": 6.03,
2058
+ "learning_rate": 4.011508809408068e-06,
2059
+ "loss": 0.1031,
2060
+ "step": 3330
2061
+ },
2062
+ {
2063
+ "epoch": 6.05,
2064
+ "learning_rate": 3.980855191021931e-06,
2065
+ "loss": 0.0957,
2066
+ "step": 3340
2067
+ },
2068
+ {
2069
+ "epoch": 6.07,
2070
+ "learning_rate": 3.9502414867770916e-06,
2071
+ "loss": 0.1012,
2072
+ "step": 3350
2073
+ },
2074
+ {
2075
+ "epoch": 6.09,
2076
+ "learning_rate": 3.919668895639298e-06,
2077
+ "loss": 0.1092,
2078
+ "step": 3360
2079
+ },
2080
+ {
2081
+ "epoch": 6.1,
2082
+ "learning_rate": 3.8891386149641304e-06,
2083
+ "loss": 0.1065,
2084
+ "step": 3370
2085
+ },
2086
+ {
2087
+ "epoch": 6.12,
2088
+ "learning_rate": 3.858651840450103e-06,
2089
+ "loss": 0.102,
2090
+ "step": 3380
2091
+ },
2092
+ {
2093
+ "epoch": 6.14,
2094
+ "learning_rate": 3.828209766091844e-06,
2095
+ "loss": 0.1022,
2096
+ "step": 3390
2097
+ },
2098
+ {
2099
+ "epoch": 6.16,
2100
+ "learning_rate": 3.797813584133327e-06,
2101
+ "loss": 0.0941,
2102
+ "step": 3400
2103
+ },
2104
+ {
2105
+ "epoch": 6.18,
2106
+ "learning_rate": 3.7674644850211815e-06,
2107
+ "loss": 0.1002,
2108
+ "step": 3410
2109
+ },
2110
+ {
2111
+ "epoch": 6.2,
2112
+ "learning_rate": 3.7401915340728467e-06,
2113
+ "loss": 0.1012,
2114
+ "step": 3420
2115
+ },
2116
+ {
2117
+ "epoch": 6.21,
2118
+ "learning_rate": 3.7099351654075777e-06,
2119
+ "loss": 0.0918,
2120
+ "step": 3430
2121
+ },
2122
+ {
2123
+ "epoch": 6.23,
2124
+ "learning_rate": 3.6797293212895947e-06,
2125
+ "loss": 0.0987,
2126
+ "step": 3440
2127
+ },
2128
+ {
2129
+ "epoch": 6.25,
2130
+ "learning_rate": 3.649575184711069e-06,
2131
+ "loss": 0.0862,
2132
+ "step": 3450
2133
+ },
2134
+ {
2135
+ "epoch": 6.27,
2136
+ "learning_rate": 3.6194739366390806e-06,
2137
+ "loss": 0.0944,
2138
+ "step": 3460
2139
+ },
2140
+ {
2141
+ "epoch": 6.29,
2142
+ "learning_rate": 3.5894267559693628e-06,
2143
+ "loss": 0.1108,
2144
+ "step": 3470
2145
+ },
2146
+ {
2147
+ "epoch": 6.3,
2148
+ "learning_rate": 3.559434819480137e-06,
2149
+ "loss": 0.0942,
2150
+ "step": 3480
2151
+ },
2152
+ {
2153
+ "epoch": 6.32,
2154
+ "learning_rate": 3.5294993017860226e-06,
2155
+ "loss": 0.0873,
2156
+ "step": 3490
2157
+ },
2158
+ {
2159
+ "epoch": 6.34,
2160
+ "learning_rate": 3.499621375292031e-06,
2161
+ "loss": 0.0988,
2162
+ "step": 3500
2163
+ },
2164
+ {
2165
+ "epoch": 6.36,
2166
+ "learning_rate": 3.4698022101476548e-06,
2167
+ "loss": 0.0908,
2168
+ "step": 3510
2169
+ },
2170
+ {
2171
+ "epoch": 6.38,
2172
+ "learning_rate": 3.440042974201034e-06,
2173
+ "loss": 0.0977,
2174
+ "step": 3520
2175
+ },
2176
+ {
2177
+ "epoch": 6.39,
2178
+ "learning_rate": 3.410344832953224e-06,
2179
+ "loss": 0.0942,
2180
+ "step": 3530
2181
+ },
2182
+ {
2183
+ "epoch": 6.41,
2184
+ "learning_rate": 3.380708949512543e-06,
2185
+ "loss": 0.1131,
2186
+ "step": 3540
2187
+ },
2188
+ {
2189
+ "epoch": 6.43,
2190
+ "learning_rate": 3.3511364845490237e-06,
2191
+ "loss": 0.0919,
2192
+ "step": 3550
2193
+ },
2194
+ {
2195
+ "epoch": 6.45,
2196
+ "learning_rate": 3.321628596248958e-06,
2197
+ "loss": 0.0998,
2198
+ "step": 3560
2199
+ },
2200
+ {
2201
+ "epoch": 6.47,
2202
+ "learning_rate": 3.29218644026953e-06,
2203
+ "loss": 0.1176,
2204
+ "step": 3570
2205
+ },
2206
+ {
2207
+ "epoch": 6.48,
2208
+ "learning_rate": 3.2628111696935655e-06,
2209
+ "loss": 0.1156,
2210
+ "step": 3580
2211
+ },
2212
+ {
2213
+ "epoch": 6.5,
2214
+ "learning_rate": 3.2335039349843646e-06,
2215
+ "loss": 0.0825,
2216
+ "step": 3590
2217
+ },
2218
+ {
2219
+ "epoch": 6.52,
2220
+ "learning_rate": 3.2042658839406453e-06,
2221
+ "loss": 0.0967,
2222
+ "step": 3600
2223
+ },
2224
+ {
2225
+ "epoch": 6.54,
2226
+ "learning_rate": 3.1750981616515965e-06,
2227
+ "loss": 0.1033,
2228
+ "step": 3610
2229
+ },
2230
+ {
2231
+ "epoch": 6.56,
2232
+ "learning_rate": 3.1460019104520246e-06,
2233
+ "loss": 0.1165,
2234
+ "step": 3620
2235
+ },
2236
+ {
2237
+ "epoch": 6.58,
2238
+ "learning_rate": 3.1169782698776195e-06,
2239
+ "loss": 0.0958,
2240
+ "step": 3630
2241
+ },
2242
+ {
2243
+ "epoch": 6.59,
2244
+ "learning_rate": 3.0880283766203203e-06,
2245
+ "loss": 0.1033,
2246
+ "step": 3640
2247
+ },
2248
+ {
2249
+ "epoch": 6.61,
2250
+ "learning_rate": 3.0591533644838024e-06,
2251
+ "loss": 0.0975,
2252
+ "step": 3650
2253
+ },
2254
+ {
2255
+ "epoch": 6.63,
2256
+ "learning_rate": 3.0303543643390722e-06,
2257
+ "loss": 0.0931,
2258
+ "step": 3660
2259
+ },
2260
+ {
2261
+ "epoch": 6.65,
2262
+ "learning_rate": 3.001632504080173e-06,
2263
+ "loss": 0.1036,
2264
+ "step": 3670
2265
+ },
2266
+ {
2267
+ "epoch": 6.67,
2268
+ "learning_rate": 2.9729889085800174e-06,
2269
+ "loss": 0.1015,
2270
+ "step": 3680
2271
+ },
2272
+ {
2273
+ "epoch": 6.68,
2274
+ "learning_rate": 2.944424699646329e-06,
2275
+ "loss": 0.0972,
2276
+ "step": 3690
2277
+ },
2278
+ {
2279
+ "epoch": 6.7,
2280
+ "learning_rate": 2.9159409959777086e-06,
2281
+ "loss": 0.088,
2282
+ "step": 3700
2283
+ },
2284
+ {
2285
+ "epoch": 6.72,
2286
+ "learning_rate": 2.8875389131198174e-06,
2287
+ "loss": 0.1096,
2288
+ "step": 3710
2289
+ },
2290
+ {
2291
+ "epoch": 6.74,
2292
+ "learning_rate": 2.859219563421699e-06,
2293
+ "loss": 0.0976,
2294
+ "step": 3720
2295
+ },
2296
+ {
2297
+ "epoch": 6.76,
2298
+ "learning_rate": 2.830984055992201e-06,
2299
+ "loss": 0.112,
2300
+ "step": 3730
2301
+ },
2302
+ {
2303
+ "epoch": 6.77,
2304
+ "learning_rate": 2.8028334966565395e-06,
2305
+ "loss": 0.094,
2306
+ "step": 3740
2307
+ },
2308
+ {
2309
+ "epoch": 6.79,
2310
+ "learning_rate": 2.7747689879130024e-06,
2311
+ "loss": 0.1054,
2312
+ "step": 3750
2313
+ },
2314
+ {
2315
+ "epoch": 6.81,
2316
+ "learning_rate": 2.7467916288897567e-06,
2317
+ "loss": 0.1094,
2318
+ "step": 3760
2319
+ },
2320
+ {
2321
+ "epoch": 6.83,
2322
+ "learning_rate": 2.7189025153018095e-06,
2323
+ "loss": 0.1053,
2324
+ "step": 3770
2325
+ },
2326
+ {
2327
+ "epoch": 6.85,
2328
+ "learning_rate": 2.6911027394080923e-06,
2329
+ "loss": 0.0901,
2330
+ "step": 3780
2331
+ },
2332
+ {
2333
+ "epoch": 6.86,
2334
+ "learning_rate": 2.6633933899686858e-06,
2335
+ "loss": 0.0987,
2336
+ "step": 3790
2337
+ },
2338
+ {
2339
+ "epoch": 6.88,
2340
+ "learning_rate": 2.6357755522021767e-06,
2341
+ "loss": 0.1008,
2342
+ "step": 3800
2343
+ },
2344
+ {
2345
+ "epoch": 6.9,
2346
+ "learning_rate": 2.608250307743153e-06,
2347
+ "loss": 0.079,
2348
+ "step": 3810
2349
+ },
2350
+ {
2351
+ "epoch": 6.92,
2352
+ "learning_rate": 2.5808187345998513e-06,
2353
+ "loss": 0.0898,
2354
+ "step": 3820
2355
+ },
2356
+ {
2357
+ "epoch": 6.94,
2358
+ "learning_rate": 2.5534819071119315e-06,
2359
+ "loss": 0.0951,
2360
+ "step": 3830
2361
+ },
2362
+ {
2363
+ "epoch": 6.96,
2364
+ "learning_rate": 2.5262408959084e-06,
2365
+ "loss": 0.0952,
2366
+ "step": 3840
2367
+ },
2368
+ {
2369
+ "epoch": 6.97,
2370
+ "learning_rate": 2.499096767865682e-06,
2371
+ "loss": 0.0866,
2372
+ "step": 3850
2373
+ },
2374
+ {
2375
+ "epoch": 6.99,
2376
+ "learning_rate": 2.472050586065836e-06,
2377
+ "loss": 0.0916,
2378
+ "step": 3860
2379
+ },
2380
+ {
2381
+ "epoch": 7.0,
2382
+ "eval_loss": 0.46777281165122986,
2383
+ "eval_runtime": 245.4593,
2384
+ "eval_samples_per_second": 8.005,
2385
+ "eval_steps_per_second": 4.005,
2386
+ "eval_wer": 0.15547409278541063,
2387
+ "step": 3864
2388
+ },
2389
+ {
2390
+ "epoch": 7.01,
2391
+ "learning_rate": 2.4451034097549237e-06,
2392
+ "loss": 0.1281,
2393
+ "step": 3870
2394
+ },
2395
+ {
2396
+ "epoch": 7.03,
2397
+ "learning_rate": 2.418256294301513e-06,
2398
+ "loss": 0.0963,
2399
+ "step": 3880
2400
+ },
2401
+ {
2402
+ "epoch": 7.05,
2403
+ "learning_rate": 2.3915102911553616e-06,
2404
+ "loss": 0.1001,
2405
+ "step": 3890
2406
+ },
2407
+ {
2408
+ "epoch": 7.07,
2409
+ "learning_rate": 2.364866447806227e-06,
2410
+ "loss": 0.0893,
2411
+ "step": 3900
2412
+ },
2413
+ {
2414
+ "epoch": 7.08,
2415
+ "learning_rate": 2.338325807742845e-06,
2416
+ "loss": 0.0929,
2417
+ "step": 3910
2418
+ },
2419
+ {
2420
+ "epoch": 7.1,
2421
+ "learning_rate": 2.3118894104120614e-06,
2422
+ "loss": 0.0893,
2423
+ "step": 3920
2424
+ },
2425
+ {
2426
+ "epoch": 7.12,
2427
+ "learning_rate": 2.285558291178126e-06,
2428
+ "loss": 0.0957,
2429
+ "step": 3930
2430
+ },
2431
+ {
2432
+ "epoch": 7.14,
2433
+ "learning_rate": 2.2593334812821356e-06,
2434
+ "loss": 0.095,
2435
+ "step": 3940
2436
+ },
2437
+ {
2438
+ "epoch": 7.16,
2439
+ "learning_rate": 2.2332160078016546e-06,
2440
+ "loss": 0.0996,
2441
+ "step": 3950
2442
+ },
2443
+ {
2444
+ "epoch": 7.17,
2445
+ "learning_rate": 2.207206893610487e-06,
2446
+ "loss": 0.1002,
2447
+ "step": 3960
2448
+ },
2449
+ {
2450
+ "epoch": 7.19,
2451
+ "learning_rate": 2.181307157338615e-06,
2452
+ "loss": 0.1148,
2453
+ "step": 3970
2454
+ },
2455
+ {
2456
+ "epoch": 7.21,
2457
+ "learning_rate": 2.155517813332307e-06,
2458
+ "loss": 0.0909,
2459
+ "step": 3980
2460
+ },
2461
+ {
2462
+ "epoch": 7.23,
2463
+ "learning_rate": 2.1298398716143897e-06,
2464
+ "loss": 0.0808,
2465
+ "step": 3990
2466
+ },
2467
+ {
2468
+ "epoch": 7.25,
2469
+ "learning_rate": 2.104274337844692e-06,
2470
+ "loss": 0.1056,
2471
+ "step": 4000
2472
+ },
2473
+ {
2474
+ "epoch": 7.26,
2475
+ "learning_rate": 2.0788222132806566e-06,
2476
+ "loss": 0.0892,
2477
+ "step": 4010
2478
+ },
2479
+ {
2480
+ "epoch": 7.28,
2481
+ "learning_rate": 2.0534844947381322e-06,
2482
+ "loss": 0.1216,
2483
+ "step": 4020
2484
+ },
2485
+ {
2486
+ "epoch": 7.3,
2487
+ "learning_rate": 2.0282621745523285e-06,
2488
+ "loss": 0.0946,
2489
+ "step": 4030
2490
+ },
2491
+ {
2492
+ "epoch": 7.32,
2493
+ "learning_rate": 2.003156240538954e-06,
2494
+ "loss": 0.089,
2495
+ "step": 4040
2496
+ },
2497
+ {
2498
+ "epoch": 7.34,
2499
+ "learning_rate": 1.9781676759555295e-06,
2500
+ "loss": 0.0973,
2501
+ "step": 4050
2502
+ },
2503
+ {
2504
+ "epoch": 7.35,
2505
+ "learning_rate": 1.9532974594628783e-06,
2506
+ "loss": 0.0887,
2507
+ "step": 4060
2508
+ },
2509
+ {
2510
+ "epoch": 7.37,
2511
+ "learning_rate": 1.928546565086799e-06,
2512
+ "loss": 0.114,
2513
+ "step": 4070
2514
+ },
2515
+ {
2516
+ "epoch": 7.39,
2517
+ "learning_rate": 1.903915962179913e-06,
2518
+ "loss": 0.0899,
2519
+ "step": 4080
2520
+ },
2521
+ {
2522
+ "epoch": 7.41,
2523
+ "learning_rate": 1.8794066153837092e-06,
2524
+ "loss": 0.1057,
2525
+ "step": 4090
2526
+ },
2527
+ {
2528
+ "epoch": 7.43,
2529
+ "learning_rate": 1.85501948459076e-06,
2530
+ "loss": 0.1055,
2531
+ "step": 4100
2532
+ },
2533
+ {
2534
+ "epoch": 7.45,
2535
+ "learning_rate": 1.8307555249071252e-06,
2536
+ "loss": 0.0903,
2537
+ "step": 4110
2538
+ },
2539
+ {
2540
+ "epoch": 7.46,
2541
+ "learning_rate": 1.806615686614951e-06,
2542
+ "loss": 0.1048,
2543
+ "step": 4120
2544
+ },
2545
+ {
2546
+ "epoch": 7.48,
2547
+ "learning_rate": 1.7826009151352486e-06,
2548
+ "loss": 0.0992,
2549
+ "step": 4130
2550
+ },
2551
+ {
2552
+ "epoch": 7.5,
2553
+ "learning_rate": 1.7587121509908723e-06,
2554
+ "loss": 0.1014,
2555
+ "step": 4140
2556
+ },
2557
+ {
2558
+ "epoch": 7.52,
2559
+ "learning_rate": 1.7349503297696745e-06,
2560
+ "loss": 0.091,
2561
+ "step": 4150
2562
+ },
2563
+ {
2564
+ "epoch": 7.54,
2565
+ "learning_rate": 1.7113163820878775e-06,
2566
+ "loss": 0.1005,
2567
+ "step": 4160
2568
+ },
2569
+ {
2570
+ "epoch": 7.55,
2571
+ "learning_rate": 1.6878112335536173e-06,
2572
+ "loss": 0.1074,
2573
+ "step": 4170
2574
+ },
2575
+ {
2576
+ "epoch": 7.57,
2577
+ "learning_rate": 1.6644358047306957e-06,
2578
+ "loss": 0.0981,
2579
+ "step": 4180
2580
+ },
2581
+ {
2582
+ "epoch": 7.59,
2583
+ "learning_rate": 1.6411910111025265e-06,
2584
+ "loss": 0.0943,
2585
+ "step": 4190
2586
+ },
2587
+ {
2588
+ "epoch": 7.61,
2589
+ "learning_rate": 1.6180777630362832e-06,
2590
+ "loss": 0.1027,
2591
+ "step": 4200
2592
+ },
2593
+ {
2594
+ "epoch": 7.63,
2595
+ "learning_rate": 1.5950969657472375e-06,
2596
+ "loss": 0.0869,
2597
+ "step": 4210
2598
+ },
2599
+ {
2600
+ "epoch": 7.64,
2601
+ "learning_rate": 1.5722495192633198e-06,
2602
+ "loss": 0.117,
2603
+ "step": 4220
2604
+ },
2605
+ {
2606
+ "epoch": 7.66,
2607
+ "learning_rate": 1.5495363183898594e-06,
2608
+ "loss": 0.1003,
2609
+ "step": 4230
2610
+ },
2611
+ {
2612
+ "epoch": 7.68,
2613
+ "learning_rate": 1.5269582526745448e-06,
2614
+ "loss": 0.0911,
2615
+ "step": 4240
2616
+ },
2617
+ {
2618
+ "epoch": 7.7,
2619
+ "learning_rate": 1.5045162063725838e-06,
2620
+ "loss": 0.0874,
2621
+ "step": 4250
2622
+ },
2623
+ {
2624
+ "epoch": 7.72,
2625
+ "learning_rate": 1.4822110584120735e-06,
2626
+ "loss": 0.1026,
2627
+ "step": 4260
2628
+ },
2629
+ {
2630
+ "epoch": 7.73,
2631
+ "learning_rate": 1.4600436823595787e-06,
2632
+ "loss": 0.0936,
2633
+ "step": 4270
2634
+ },
2635
+ {
2636
+ "epoch": 7.75,
2637
+ "learning_rate": 1.43801494638591e-06,
2638
+ "loss": 0.0835,
2639
+ "step": 4280
2640
+ },
2641
+ {
2642
+ "epoch": 7.77,
2643
+ "learning_rate": 1.4161257132321377e-06,
2644
+ "loss": 0.0989,
2645
+ "step": 4290
2646
+ },
2647
+ {
2648
+ "epoch": 7.79,
2649
+ "learning_rate": 1.3943768401757917e-06,
2650
+ "loss": 0.0931,
2651
+ "step": 4300
2652
+ },
2653
+ {
2654
+ "epoch": 7.81,
2655
+ "learning_rate": 1.3727691789972913e-06,
2656
+ "loss": 0.0815,
2657
+ "step": 4310
2658
+ },
2659
+ {
2660
+ "epoch": 7.83,
2661
+ "learning_rate": 1.3513035759465841e-06,
2662
+ "loss": 0.1001,
2663
+ "step": 4320
2664
+ },
2665
+ {
2666
+ "epoch": 7.84,
2667
+ "learning_rate": 1.3299808717100037e-06,
2668
+ "loss": 0.0907,
2669
+ "step": 4330
2670
+ },
2671
+ {
2672
+ "epoch": 7.86,
2673
+ "learning_rate": 1.3088019013773456e-06,
2674
+ "loss": 0.0899,
2675
+ "step": 4340
2676
+ },
2677
+ {
2678
+ "epoch": 7.88,
2679
+ "learning_rate": 1.2877674944091605e-06,
2680
+ "loss": 0.0953,
2681
+ "step": 4350
2682
+ },
2683
+ {
2684
+ "epoch": 7.9,
2685
+ "learning_rate": 1.266878474604269e-06,
2686
+ "loss": 0.081,
2687
+ "step": 4360
2688
+ },
2689
+ {
2690
+ "epoch": 7.92,
2691
+ "learning_rate": 1.2461356600674984e-06,
2692
+ "loss": 0.1002,
2693
+ "step": 4370
2694
+ },
2695
+ {
2696
+ "epoch": 7.93,
2697
+ "learning_rate": 1.2255398631776427e-06,
2698
+ "loss": 0.088,
2699
+ "step": 4380
2700
+ },
2701
+ {
2702
+ "epoch": 7.95,
2703
+ "learning_rate": 1.2050918905556463e-06,
2704
+ "loss": 0.1001,
2705
+ "step": 4390
2706
+ },
2707
+ {
2708
+ "epoch": 7.97,
2709
+ "learning_rate": 1.1847925430330093e-06,
2710
+ "loss": 0.1004,
2711
+ "step": 4400
2712
+ },
2713
+ {
2714
+ "epoch": 7.99,
2715
+ "learning_rate": 1.16464261562043e-06,
2716
+ "loss": 0.1,
2717
+ "step": 4410
2718
+ },
2719
+ {
2720
+ "epoch": 8.0,
2721
+ "eval_loss": 0.47049203515052795,
2722
+ "eval_runtime": 243.653,
2723
+ "eval_samples_per_second": 8.065,
2724
+ "eval_steps_per_second": 4.034,
2725
+ "eval_wer": 0.15495040354876471,
2726
+ "step": 4416
2727
+ },
2728
+ {
2729
+ "epoch": 8.01,
2730
+ "learning_rate": 1.1446428974766643e-06,
2731
+ "loss": 0.1086,
2732
+ "step": 4420
2733
+ },
2734
+ {
2735
+ "epoch": 8.03,
2736
+ "learning_rate": 1.1247941718776212e-06,
2737
+ "loss": 0.0893,
2738
+ "step": 4430
2739
+ },
2740
+ {
2741
+ "epoch": 8.04,
2742
+ "learning_rate": 1.1050972161856843e-06,
2743
+ "loss": 0.1016,
2744
+ "step": 4440
2745
+ },
2746
+ {
2747
+ "epoch": 8.06,
2748
+ "learning_rate": 1.0855528018192679e-06,
2749
+ "loss": 0.1021,
2750
+ "step": 4450
2751
+ },
2752
+ {
2753
+ "epoch": 8.08,
2754
+ "learning_rate": 1.066161694222606e-06,
2755
+ "loss": 0.1029,
2756
+ "step": 4460
2757
+ },
2758
+ {
2759
+ "epoch": 8.1,
2760
+ "learning_rate": 1.0469246528357707e-06,
2761
+ "loss": 0.1003,
2762
+ "step": 4470
2763
+ },
2764
+ {
2765
+ "epoch": 8.12,
2766
+ "learning_rate": 1.0278424310649327e-06,
2767
+ "loss": 0.0996,
2768
+ "step": 4480
2769
+ },
2770
+ {
2771
+ "epoch": 8.13,
2772
+ "learning_rate": 1.008915776252854e-06,
2773
+ "loss": 0.1053,
2774
+ "step": 4490
2775
+ },
2776
+ {
2777
+ "epoch": 8.15,
2778
+ "learning_rate": 9.901454296496182e-07,
2779
+ "loss": 0.0977,
2780
+ "step": 4500
2781
+ },
2782
+ {
2783
+ "epoch": 8.17,
2784
+ "learning_rate": 9.715321263835996e-07,
2785
+ "loss": 0.1032,
2786
+ "step": 4510
2787
+ },
2788
+ {
2789
+ "epoch": 8.19,
2790
+ "learning_rate": 9.530765954326726e-07,
2791
+ "loss": 0.1114,
2792
+ "step": 4520
2793
+ },
2794
+ {
2795
+ "epoch": 8.21,
2796
+ "learning_rate": 9.347795595956632e-07,
2797
+ "loss": 0.0806,
2798
+ "step": 4530
2799
+ },
2800
+ {
2801
+ "epoch": 8.22,
2802
+ "learning_rate": 9.16641735464035e-07,
2803
+ "loss": 0.0995,
2804
+ "step": 4540
2805
+ },
2806
+ {
2807
+ "epoch": 8.24,
2808
+ "learning_rate": 8.986638333938352e-07,
2809
+ "loss": 0.1056,
2810
+ "step": 4550
2811
+ },
2812
+ {
2813
+ "epoch": 8.26,
2814
+ "learning_rate": 8.808465574778635e-07,
2815
+ "loss": 0.0918,
2816
+ "step": 4560
2817
+ },
2818
+ {
2819
+ "epoch": 8.28,
2820
+ "learning_rate": 8.631906055181044e-07,
2821
+ "loss": 0.0987,
2822
+ "step": 4570
2823
+ },
2824
+ {
2825
+ "epoch": 8.3,
2826
+ "learning_rate": 8.45696668998392e-07,
2827
+ "loss": 0.0999,
2828
+ "step": 4580
2829
+ },
2830
+ {
2831
+ "epoch": 8.31,
2832
+ "learning_rate": 8.283654330573326e-07,
2833
+ "loss": 0.0911,
2834
+ "step": 4590
2835
+ },
2836
+ {
2837
+ "epoch": 8.33,
2838
+ "learning_rate": 8.111975764614732e-07,
2839
+ "loss": 0.1001,
2840
+ "step": 4600
2841
+ },
2842
+ {
2843
+ "epoch": 8.35,
2844
+ "learning_rate": 7.94193771578709e-07,
2845
+ "loss": 0.1019,
2846
+ "step": 4610
2847
+ },
2848
+ {
2849
+ "epoch": 8.37,
2850
+ "learning_rate": 7.773546843519619e-07,
2851
+ "loss": 0.0912,
2852
+ "step": 4620
2853
+ },
2854
+ {
2855
+ "epoch": 8.39,
2856
+ "learning_rate": 7.606809742730942e-07,
2857
+ "loss": 0.095,
2858
+ "step": 4630
2859
+ },
2860
+ {
2861
+ "epoch": 8.41,
2862
+ "learning_rate": 7.441732943570784e-07,
2863
+ "loss": 0.0907,
2864
+ "step": 4640
2865
+ },
2866
+ {
2867
+ "epoch": 8.42,
2868
+ "learning_rate": 7.278322911164243e-07,
2869
+ "loss": 0.0883,
2870
+ "step": 4650
2871
+ },
2872
+ {
2873
+ "epoch": 8.44,
2874
+ "learning_rate": 7.116586045358609e-07,
2875
+ "loss": 0.1234,
2876
+ "step": 4660
2877
+ },
2878
+ {
2879
+ "epoch": 8.46,
2880
+ "learning_rate": 6.956528680472641e-07,
2881
+ "loss": 0.1065,
2882
+ "step": 4670
2883
+ },
2884
+ {
2885
+ "epoch": 8.48,
2886
+ "learning_rate": 6.79815708504859e-07,
2887
+ "loss": 0.098,
2888
+ "step": 4680
2889
+ },
2890
+ {
2891
+ "epoch": 8.5,
2892
+ "learning_rate": 6.641477461606632e-07,
2893
+ "loss": 0.1049,
2894
+ "step": 4690
2895
+ },
2896
+ {
2897
+ "epoch": 8.51,
2898
+ "learning_rate": 6.48649594640196e-07,
2899
+ "loss": 0.0924,
2900
+ "step": 4700
2901
+ },
2902
+ {
2903
+ "epoch": 8.53,
2904
+ "learning_rate": 6.333218609184466e-07,
2905
+ "loss": 0.1016,
2906
+ "step": 4710
2907
+ },
2908
+ {
2909
+ "epoch": 8.55,
2910
+ "learning_rate": 6.181651452961023e-07,
2911
+ "loss": 0.1077,
2912
+ "step": 4720
2913
+ },
2914
+ {
2915
+ "epoch": 8.57,
2916
+ "learning_rate": 6.031800413760408e-07,
2917
+ "loss": 0.0862,
2918
+ "step": 4730
2919
+ },
2920
+ {
2921
+ "epoch": 8.59,
2922
+ "learning_rate": 5.883671360400739e-07,
2923
+ "loss": 0.0892,
2924
+ "step": 4740
2925
+ },
2926
+ {
2927
+ "epoch": 8.6,
2928
+ "learning_rate": 5.737270094259723e-07,
2929
+ "loss": 0.081,
2930
+ "step": 4750
2931
+ },
2932
+ {
2933
+ "epoch": 8.62,
2934
+ "learning_rate": 5.592602349047405e-07,
2935
+ "loss": 0.0864,
2936
+ "step": 4760
2937
+ },
2938
+ {
2939
+ "epoch": 8.64,
2940
+ "learning_rate": 5.449673790581611e-07,
2941
+ "loss": 0.1028,
2942
+ "step": 4770
2943
+ },
2944
+ {
2945
+ "epoch": 8.66,
2946
+ "learning_rate": 5.308490016566042e-07,
2947
+ "loss": 0.0988,
2948
+ "step": 4780
2949
+ },
2950
+ {
2951
+ "epoch": 8.68,
2952
+ "learning_rate": 5.169056556371077e-07,
2953
+ "loss": 0.114,
2954
+ "step": 4790
2955
+ },
2956
+ {
2957
+ "epoch": 8.69,
2958
+ "learning_rate": 5.031378870817182e-07,
2959
+ "loss": 0.107,
2960
+ "step": 4800
2961
+ },
2962
+ {
2963
+ "epoch": 8.71,
2964
+ "learning_rate": 4.895462351961039e-07,
2965
+ "loss": 0.0935,
2966
+ "step": 4810
2967
+ },
2968
+ {
2969
+ "epoch": 8.73,
2970
+ "learning_rate": 4.761312322884415e-07,
2971
+ "loss": 0.0963,
2972
+ "step": 4820
2973
+ },
2974
+ {
2975
+ "epoch": 8.75,
2976
+ "learning_rate": 4.628934037485644e-07,
2977
+ "loss": 0.0836,
2978
+ "step": 4830
2979
+ },
2980
+ {
2981
+ "epoch": 8.77,
2982
+ "learning_rate": 4.498332680273881e-07,
2983
+ "loss": 0.1062,
2984
+ "step": 4840
2985
+ },
2986
+ {
2987
+ "epoch": 8.79,
2988
+ "learning_rate": 4.369513366166039e-07,
2989
+ "loss": 0.0952,
2990
+ "step": 4850
2991
+ },
2992
+ {
2993
+ "epoch": 8.8,
2994
+ "learning_rate": 4.242481140286486e-07,
2995
+ "loss": 0.0869,
2996
+ "step": 4860
2997
+ },
2998
+ {
2999
+ "epoch": 8.82,
3000
+ "learning_rate": 4.1172409777694433e-07,
3001
+ "loss": 0.0906,
3002
+ "step": 4870
3003
+ },
3004
+ {
3005
+ "epoch": 8.84,
3006
+ "learning_rate": 3.9937977835641294e-07,
3007
+ "loss": 0.087,
3008
+ "step": 4880
3009
+ },
3010
+ {
3011
+ "epoch": 8.86,
3012
+ "learning_rate": 3.872156392242682e-07,
3013
+ "loss": 0.0754,
3014
+ "step": 4890
3015
+ },
3016
+ {
3017
+ "epoch": 8.88,
3018
+ "learning_rate": 3.7523215678108015e-07,
3019
+ "loss": 0.0993,
3020
+ "step": 4900
3021
+ },
3022
+ {
3023
+ "epoch": 8.89,
3024
+ "learning_rate": 3.634298003521186e-07,
3025
+ "loss": 0.0977,
3026
+ "step": 4910
3027
+ },
3028
+ {
3029
+ "epoch": 8.91,
3030
+ "learning_rate": 3.518090321689693e-07,
3031
+ "loss": 0.1168,
3032
+ "step": 4920
3033
+ },
3034
+ {
3035
+ "epoch": 8.93,
3036
+ "learning_rate": 3.4037030735143474e-07,
3037
+ "loss": 0.0901,
3038
+ "step": 4930
3039
+ },
3040
+ {
3041
+ "epoch": 8.95,
3042
+ "learning_rate": 3.291140738897064e-07,
3043
+ "loss": 0.1056,
3044
+ "step": 4940
3045
+ },
3046
+ {
3047
+ "epoch": 8.97,
3048
+ "learning_rate": 3.1804077262682085e-07,
3049
+ "loss": 0.0966,
3050
+ "step": 4950
3051
+ },
3052
+ {
3053
+ "epoch": 8.98,
3054
+ "learning_rate": 3.0715083724139503e-07,
3055
+ "loss": 0.1301,
3056
+ "step": 4960
3057
+ },
3058
+ {
3059
+ "epoch": 9.0,
3060
+ "eval_loss": 0.47403913736343384,
3061
+ "eval_runtime": 244.7056,
3062
+ "eval_samples_per_second": 8.03,
3063
+ "eval_steps_per_second": 4.017,
3064
+ "eval_wer": 0.1551352350440515,
3065
+ "step": 4968
3066
+ },
3067
+ {
3068
+ "epoch": 9.0,
3069
+ "learning_rate": 2.964446942306426e-07,
3070
+ "loss": 0.108,
3071
+ "step": 4970
3072
+ },
3073
+ {
3074
+ "epoch": 9.02,
3075
+ "learning_rate": 2.8592276289366563e-07,
3076
+ "loss": 0.0953,
3077
+ "step": 4980
3078
+ },
3079
+ {
3080
+ "epoch": 9.04,
3081
+ "learning_rate": 2.755854553150394e-07,
3082
+ "loss": 0.101,
3083
+ "step": 4990
3084
+ },
3085
+ {
3086
+ "epoch": 9.06,
3087
+ "learning_rate": 2.654331763486673e-07,
3088
+ "loss": 0.0919,
3089
+ "step": 5000
3090
+ },
3091
+ {
3092
+ "epoch": 9.08,
3093
+ "learning_rate": 2.5546632360192893e-07,
3094
+ "loss": 0.094,
3095
+ "step": 5010
3096
+ },
3097
+ {
3098
+ "epoch": 9.09,
3099
+ "learning_rate": 2.456852874201082e-07,
3100
+ "loss": 0.1065,
3101
+ "step": 5020
3102
+ },
3103
+ {
3104
+ "epoch": 9.11,
3105
+ "learning_rate": 2.3609045087110317e-07,
3106
+ "loss": 0.0769,
3107
+ "step": 5030
3108
+ },
3109
+ {
3110
+ "epoch": 9.13,
3111
+ "learning_rate": 2.26682189730425e-07,
3112
+ "loss": 0.097,
3113
+ "step": 5040
3114
+ },
3115
+ {
3116
+ "epoch": 9.15,
3117
+ "learning_rate": 2.1746087246648172e-07,
3118
+ "loss": 0.0975,
3119
+ "step": 5050
3120
+ },
3121
+ {
3122
+ "epoch": 9.17,
3123
+ "learning_rate": 2.0842686022614412e-07,
3124
+ "loss": 0.0946,
3125
+ "step": 5060
3126
+ },
3127
+ {
3128
+ "epoch": 9.18,
3129
+ "learning_rate": 1.9958050682060692e-07,
3130
+ "loss": 0.0989,
3131
+ "step": 5070
3132
+ },
3133
+ {
3134
+ "epoch": 9.2,
3135
+ "learning_rate": 1.9092215871152786e-07,
3136
+ "loss": 0.0874,
3137
+ "step": 5080
3138
+ },
3139
+ {
3140
+ "epoch": 9.22,
3141
+ "learning_rate": 1.8245215499745838e-07,
3142
+ "loss": 0.1016,
3143
+ "step": 5090
3144
+ },
3145
+ {
3146
+ "epoch": 9.24,
3147
+ "learning_rate": 1.7417082740056713e-07,
3148
+ "loss": 0.1044,
3149
+ "step": 5100
3150
+ },
3151
+ {
3152
+ "epoch": 9.26,
3153
+ "learning_rate": 1.660785002536447e-07,
3154
+ "loss": 0.1017,
3155
+ "step": 5110
3156
+ },
3157
+ {
3158
+ "epoch": 9.28,
3159
+ "learning_rate": 1.5817549048740166e-07,
3160
+ "loss": 0.0955,
3161
+ "step": 5120
3162
+ },
3163
+ {
3164
+ "epoch": 9.29,
3165
+ "learning_rate": 1.504621076180579e-07,
3166
+ "loss": 0.0875,
3167
+ "step": 5130
3168
+ },
3169
+ {
3170
+ "epoch": 9.31,
3171
+ "learning_rate": 1.4293865373521953e-07,
3172
+ "loss": 0.1041,
3173
+ "step": 5140
3174
+ },
3175
+ {
3176
+ "epoch": 9.33,
3177
+ "learning_rate": 1.3560542349004723e-07,
3178
+ "loss": 0.0988,
3179
+ "step": 5150
3180
+ },
3181
+ {
3182
+ "epoch": 9.35,
3183
+ "learning_rate": 1.2846270408371832e-07,
3184
+ "loss": 0.0909,
3185
+ "step": 5160
3186
+ },
3187
+ {
3188
+ "epoch": 9.37,
3189
+ "learning_rate": 1.215107752561756e-07,
3190
+ "loss": 0.1083,
3191
+ "step": 5170
3192
+ },
3193
+ {
3194
+ "epoch": 9.38,
3195
+ "learning_rate": 1.1474990927517527e-07,
3196
+ "loss": 0.1009,
3197
+ "step": 5180
3198
+ },
3199
+ {
3200
+ "epoch": 9.4,
3201
+ "learning_rate": 1.0818037092562128e-07,
3202
+ "loss": 0.1099,
3203
+ "step": 5190
3204
+ },
3205
+ {
3206
+ "epoch": 9.42,
3207
+ "learning_rate": 1.018024174991944e-07,
3208
+ "loss": 0.0862,
3209
+ "step": 5200
3210
+ },
3211
+ {
3212
+ "epoch": 9.44,
3213
+ "learning_rate": 9.561629878427847e-08,
3214
+ "loss": 0.0782,
3215
+ "step": 5210
3216
+ },
3217
+ {
3218
+ "epoch": 9.46,
3219
+ "learning_rate": 8.962225705617556e-08,
3220
+ "loss": 0.1061,
3221
+ "step": 5220
3222
+ },
3223
+ {
3224
+ "epoch": 9.47,
3225
+ "learning_rate": 8.38205270676179e-08,
3226
+ "loss": 0.0864,
3227
+ "step": 5230
3228
+ },
3229
+ {
3230
+ "epoch": 9.49,
3231
+ "learning_rate": 7.82113360395742e-08,
3232
+ "loss": 0.1151,
3233
+ "step": 5240
3234
+ },
3235
+ {
3236
+ "epoch": 9.51,
3237
+ "learning_rate": 7.279490365235065e-08,
3238
+ "loss": 0.0906,
3239
+ "step": 5250
3240
+ },
3241
+ {
3242
+ "epoch": 9.53,
3243
+ "learning_rate": 6.757144203698607e-08,
3244
+ "loss": 0.0912,
3245
+ "step": 5260
3246
+ },
3247
+ {
3248
+ "epoch": 9.55,
3249
+ "learning_rate": 6.25411557669453e-08,
3250
+ "loss": 0.105,
3251
+ "step": 5270
3252
+ },
3253
+ {
3254
+ "epoch": 9.56,
3255
+ "learning_rate": 5.770424185010726e-08,
3256
+ "loss": 0.1117,
3257
+ "step": 5280
3258
+ },
3259
+ {
3260
+ "epoch": 9.58,
3261
+ "learning_rate": 5.306088972104939e-08,
3262
+ "loss": 0.0891,
3263
+ "step": 5290
3264
+ },
3265
+ {
3266
+ "epoch": 9.6,
3267
+ "learning_rate": 4.861128123362646e-08,
3268
+ "loss": 0.0901,
3269
+ "step": 5300
3270
+ },
3271
+ {
3272
+ "epoch": 9.62,
3273
+ "learning_rate": 4.43555906538512e-08,
3274
+ "loss": 0.0931,
3275
+ "step": 5310
3276
+ },
3277
+ {
3278
+ "epoch": 9.64,
3279
+ "learning_rate": 4.0293984653068105e-08,
3280
+ "loss": 0.1138,
3281
+ "step": 5320
3282
+ },
3283
+ {
3284
+ "epoch": 9.65,
3285
+ "learning_rate": 3.64266223014248e-08,
3286
+ "loss": 0.1282,
3287
+ "step": 5330
3288
+ },
3289
+ {
3290
+ "epoch": 9.67,
3291
+ "learning_rate": 3.275365506164363e-08,
3292
+ "loss": 0.0919,
3293
+ "step": 5340
3294
+ },
3295
+ {
3296
+ "epoch": 9.69,
3297
+ "learning_rate": 2.9275226783089806e-08,
3298
+ "loss": 0.0988,
3299
+ "step": 5350
3300
+ },
3301
+ {
3302
+ "epoch": 9.71,
3303
+ "learning_rate": 2.5991473696136416e-08,
3304
+ "loss": 0.11,
3305
+ "step": 5360
3306
+ },
3307
+ {
3308
+ "epoch": 9.73,
3309
+ "learning_rate": 2.2902524406830385e-08,
3310
+ "loss": 0.0971,
3311
+ "step": 5370
3312
+ },
3313
+ {
3314
+ "epoch": 9.75,
3315
+ "learning_rate": 2.0008499891853715e-08,
3316
+ "loss": 0.1008,
3317
+ "step": 5380
3318
+ },
3319
+ {
3320
+ "epoch": 9.76,
3321
+ "learning_rate": 1.7309513493788955e-08,
3322
+ "loss": 0.0955,
3323
+ "step": 5390
3324
+ },
3325
+ {
3326
+ "epoch": 9.78,
3327
+ "learning_rate": 1.4805670916676063e-08,
3328
+ "loss": 0.0855,
3329
+ "step": 5400
3330
+ },
3331
+ {
3332
+ "epoch": 9.8,
3333
+ "learning_rate": 1.249707022187574e-08,
3334
+ "loss": 0.1014,
3335
+ "step": 5410
3336
+ },
3337
+ {
3338
+ "epoch": 9.82,
3339
+ "learning_rate": 1.0586336291851795e-08,
3340
+ "loss": 0.1153,
3341
+ "step": 5420
3342
+ },
3343
+ {
3344
+ "epoch": 9.84,
3345
+ "learning_rate": 8.64893794362054e-09,
3346
+ "loss": 0.0956,
3347
+ "step": 5430
3348
+ },
3349
+ {
3350
+ "epoch": 9.85,
3351
+ "learning_rate": 6.907022602130054e-09,
3352
+ "loss": 0.0805,
3353
+ "step": 5440
3354
+ },
3355
+ {
3356
+ "epoch": 9.87,
3357
+ "learning_rate": 5.36065848835865e-09,
3358
+ "loss": 0.0883,
3359
+ "step": 5450
3360
+ },
3361
+ {
3362
+ "epoch": 9.89,
3363
+ "learning_rate": 4.009906164646471e-09,
3364
+ "loss": 0.0965,
3365
+ "step": 5460
3366
+ },
3367
+ {
3368
+ "epoch": 9.91,
3369
+ "learning_rate": 2.8548185323279407e-09,
3370
+ "loss": 0.1014,
3371
+ "step": 5470
3372
+ },
3373
+ {
3374
+ "epoch": 9.93,
3375
+ "learning_rate": 1.895440829656203e-09,
3376
+ "loss": 0.0836,
3377
+ "step": 5480
3378
+ },
3379
+ {
3380
+ "epoch": 9.94,
3381
+ "learning_rate": 1.13181063003176e-09,
3382
+ "loss": 0.0809,
3383
+ "step": 5490
3384
+ },
3385
+ {
3386
+ "epoch": 9.96,
3387
+ "learning_rate": 5.639578405342017e-10,
3388
+ "loss": 0.0959,
3389
+ "step": 5500
3390
+ },
3391
+ {
3392
+ "epoch": 9.98,
3393
+ "learning_rate": 1.9190470074703648e-10,
3394
+ "loss": 0.1045,
3395
+ "step": 5510
3396
+ },
3397
+ {
3398
+ "epoch": 10.0,
3399
+ "learning_rate": 1.56657818878303e-11,
3400
+ "loss": 0.0885,
3401
+ "step": 5520
3402
+ },
3403
+ {
3404
+ "epoch": 10.0,
3405
+ "eval_loss": 0.47021371126174927,
3406
+ "eval_runtime": 245.8434,
3407
+ "eval_samples_per_second": 7.993,
3408
+ "eval_steps_per_second": 3.998,
3409
+ "eval_wer": 0.15458074055819113,
3410
+ "step": 5520
3411
+ },
3412
+ {
3413
+ "epoch": 10.0,
3414
+ "step": 5520,
3415
+ "total_flos": 5.359318129634805e+19,
3416
+ "train_loss": 0.11096900397020837,
3417
+ "train_runtime": 38440.8489,
3418
+ "train_samples_per_second": 4.601,
3419
+ "train_steps_per_second": 0.144
3420
+ }
3421
+ ],
3422
+ "max_steps": 5520,
3423
+ "num_train_epochs": 10,
3424
+ "total_flos": 5.359318129634805e+19,
3425
+ "trial_name": null,
3426
+ "trial_params": null
3427
+ }