Commit
β’
cf7ab1e
1
Parent(s):
9f100c5
End of training
Browse files- all_results.json +10 -10
- checkpoint-1000/trainer_state.json +0 -94
- checkpoint-1500/trainer_state.json +0 -133
- {checkpoint-1000 β checkpoint-16000}/config.json +0 -0
- {checkpoint-1000 β checkpoint-16000}/optimizer.pt +1 -1
- {checkpoint-1000 β checkpoint-16000}/preprocessor_config.json +0 -0
- {checkpoint-1500 β checkpoint-16000}/pytorch_model.bin +1 -1
- {checkpoint-500 β checkpoint-16000}/rng_state.pth +2 -2
- {checkpoint-500 β checkpoint-16000}/scaler.pt +1 -1
- {checkpoint-1000 β checkpoint-16000}/scheduler.pt +1 -1
- checkpoint-16000/trainer_state.json +1264 -0
- {checkpoint-1000 β checkpoint-16000}/training_args.bin +0 -0
- {checkpoint-1500 β checkpoint-16500}/config.json +0 -0
- {checkpoint-1500 β checkpoint-16500}/optimizer.pt +1 -1
- {checkpoint-1500 β checkpoint-16500}/preprocessor_config.json +0 -0
- {checkpoint-1000 β checkpoint-16500}/pytorch_model.bin +1 -1
- {checkpoint-1000 β checkpoint-16500}/rng_state.pth +2 -2
- {checkpoint-1500 β checkpoint-16500}/scaler.pt +1 -1
- {checkpoint-500 β checkpoint-16500}/scheduler.pt +1 -1
- checkpoint-16500/trainer_state.json +1303 -0
- {checkpoint-1500 β checkpoint-16500}/training_args.bin +0 -0
- {checkpoint-500 β checkpoint-17000}/config.json +0 -0
- {checkpoint-500 β checkpoint-17000}/optimizer.pt +1 -1
- {checkpoint-500 β checkpoint-17000}/preprocessor_config.json +0 -0
- {checkpoint-500 β checkpoint-17000}/pytorch_model.bin +1 -1
- {checkpoint-1500 β checkpoint-17000}/rng_state.pth +2 -2
- {checkpoint-1000 β checkpoint-17000}/scaler.pt +1 -1
- {checkpoint-1500 β checkpoint-17000}/scheduler.pt +1 -1
- checkpoint-17000/trainer_state.json +1342 -0
- {checkpoint-500 β checkpoint-17000}/training_args.bin +0 -0
- checkpoint-500/trainer_state.json +0 -55
- eval_results.json +6 -6
- pytorch_model.bin +1 -1
- train_results.json +5 -5
- trainer_state.json +1269 -60
all_results.json
CHANGED
@@ -1,14 +1,14 @@
|
|
1 |
{
|
2 |
-
"epoch":
|
3 |
-
"eval_loss":
|
4 |
-
"eval_runtime":
|
5 |
"eval_samples": 4620,
|
6 |
-
"eval_samples_per_second":
|
7 |
-
"eval_steps_per_second": 4.
|
8 |
-
"eval_wer": 0.
|
9 |
-
"train_loss":
|
10 |
-
"train_runtime":
|
11 |
"train_samples": 11030,
|
12 |
-
"train_samples_per_second": 18.
|
13 |
-
"train_steps_per_second": 0.
|
14 |
}
|
|
|
1 |
{
|
2 |
+
"epoch": 50.0,
|
3 |
+
"eval_loss": 0.31790396571159363,
|
4 |
+
"eval_runtime": 136.0793,
|
5 |
"eval_samples": 4620,
|
6 |
+
"eval_samples_per_second": 33.951,
|
7 |
+
"eval_steps_per_second": 4.248,
|
8 |
+
"eval_wer": 0.2734810010402007,
|
9 |
+
"train_loss": 1.4555730460410894,
|
10 |
+
"train_runtime": 29339.4334,
|
11 |
"train_samples": 11030,
|
12 |
+
"train_samples_per_second": 18.797,
|
13 |
+
"train_steps_per_second": 0.586
|
14 |
}
|
checkpoint-1000/trainer_state.json
DELETED
@@ -1,94 +0,0 @@
|
|
1 |
-
{
|
2 |
-
"best_metric": null,
|
3 |
-
"best_model_checkpoint": null,
|
4 |
-
"epoch": 2.9050036258158087,
|
5 |
-
"global_step": 1000,
|
6 |
-
"is_hyper_param_search": false,
|
7 |
-
"is_local_process_zero": true,
|
8 |
-
"is_world_process_zero": true,
|
9 |
-
"log_history": [
|
10 |
-
{
|
11 |
-
"epoch": 0.29,
|
12 |
-
"learning_rate": 3.7125e-06,
|
13 |
-
"loss": 12.2713,
|
14 |
-
"step": 100
|
15 |
-
},
|
16 |
-
{
|
17 |
-
"epoch": 0.58,
|
18 |
-
"learning_rate": 7.4625e-06,
|
19 |
-
"loss": 6.2026,
|
20 |
-
"step": 200
|
21 |
-
},
|
22 |
-
{
|
23 |
-
"epoch": 0.87,
|
24 |
-
"learning_rate": 1.1212499999999998e-05,
|
25 |
-
"loss": 4.1374,
|
26 |
-
"step": 300
|
27 |
-
},
|
28 |
-
{
|
29 |
-
"epoch": 1.16,
|
30 |
-
"learning_rate": 1.49625e-05,
|
31 |
-
"loss": 3.6755,
|
32 |
-
"step": 400
|
33 |
-
},
|
34 |
-
{
|
35 |
-
"epoch": 1.45,
|
36 |
-
"learning_rate": 1.8712499999999997e-05,
|
37 |
-
"loss": 3.3332,
|
38 |
-
"step": 500
|
39 |
-
},
|
40 |
-
{
|
41 |
-
"epoch": 1.45,
|
42 |
-
"eval_loss": 3.292029857635498,
|
43 |
-
"eval_runtime": 137.2974,
|
44 |
-
"eval_samples_per_second": 33.65,
|
45 |
-
"eval_steps_per_second": 4.21,
|
46 |
-
"eval_wer": 1.0,
|
47 |
-
"step": 500
|
48 |
-
},
|
49 |
-
{
|
50 |
-
"epoch": 1.74,
|
51 |
-
"learning_rate": 2.2462499999999997e-05,
|
52 |
-
"loss": 3.1418,
|
53 |
-
"step": 600
|
54 |
-
},
|
55 |
-
{
|
56 |
-
"epoch": 2.03,
|
57 |
-
"learning_rate": 2.6212499999999997e-05,
|
58 |
-
"loss": 3.0879,
|
59 |
-
"step": 700
|
60 |
-
},
|
61 |
-
{
|
62 |
-
"epoch": 2.32,
|
63 |
-
"learning_rate": 2.99625e-05,
|
64 |
-
"loss": 3.0216,
|
65 |
-
"step": 800
|
66 |
-
},
|
67 |
-
{
|
68 |
-
"epoch": 2.61,
|
69 |
-
"learning_rate": 3.37125e-05,
|
70 |
-
"loss": 2.9595,
|
71 |
-
"step": 900
|
72 |
-
},
|
73 |
-
{
|
74 |
-
"epoch": 2.91,
|
75 |
-
"learning_rate": 3.7462499999999996e-05,
|
76 |
-
"loss": 2.9269,
|
77 |
-
"step": 1000
|
78 |
-
},
|
79 |
-
{
|
80 |
-
"epoch": 2.91,
|
81 |
-
"eval_loss": 2.941540241241455,
|
82 |
-
"eval_runtime": 178.4388,
|
83 |
-
"eval_samples_per_second": 25.891,
|
84 |
-
"eval_steps_per_second": 3.239,
|
85 |
-
"eval_wer": 0.9966346448020559,
|
86 |
-
"step": 1000
|
87 |
-
}
|
88 |
-
],
|
89 |
-
"max_steps": 17200,
|
90 |
-
"num_train_epochs": 50,
|
91 |
-
"total_flos": 2.9636799545011507e+18,
|
92 |
-
"trial_name": null,
|
93 |
-
"trial_params": null
|
94 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-1500/trainer_state.json
DELETED
@@ -1,133 +0,0 @@
|
|
1 |
-
{
|
2 |
-
"best_metric": null,
|
3 |
-
"best_model_checkpoint": null,
|
4 |
-
"epoch": 4.359680928208847,
|
5 |
-
"global_step": 1500,
|
6 |
-
"is_hyper_param_search": false,
|
7 |
-
"is_local_process_zero": true,
|
8 |
-
"is_world_process_zero": true,
|
9 |
-
"log_history": [
|
10 |
-
{
|
11 |
-
"epoch": 0.29,
|
12 |
-
"learning_rate": 3.7125e-06,
|
13 |
-
"loss": 12.2713,
|
14 |
-
"step": 100
|
15 |
-
},
|
16 |
-
{
|
17 |
-
"epoch": 0.58,
|
18 |
-
"learning_rate": 7.4625e-06,
|
19 |
-
"loss": 6.2026,
|
20 |
-
"step": 200
|
21 |
-
},
|
22 |
-
{
|
23 |
-
"epoch": 0.87,
|
24 |
-
"learning_rate": 1.1212499999999998e-05,
|
25 |
-
"loss": 4.1374,
|
26 |
-
"step": 300
|
27 |
-
},
|
28 |
-
{
|
29 |
-
"epoch": 1.16,
|
30 |
-
"learning_rate": 1.49625e-05,
|
31 |
-
"loss": 3.6755,
|
32 |
-
"step": 400
|
33 |
-
},
|
34 |
-
{
|
35 |
-
"epoch": 1.45,
|
36 |
-
"learning_rate": 1.8712499999999997e-05,
|
37 |
-
"loss": 3.3332,
|
38 |
-
"step": 500
|
39 |
-
},
|
40 |
-
{
|
41 |
-
"epoch": 1.45,
|
42 |
-
"eval_loss": 3.292029857635498,
|
43 |
-
"eval_runtime": 137.2974,
|
44 |
-
"eval_samples_per_second": 33.65,
|
45 |
-
"eval_steps_per_second": 4.21,
|
46 |
-
"eval_wer": 1.0,
|
47 |
-
"step": 500
|
48 |
-
},
|
49 |
-
{
|
50 |
-
"epoch": 1.74,
|
51 |
-
"learning_rate": 2.2462499999999997e-05,
|
52 |
-
"loss": 3.1418,
|
53 |
-
"step": 600
|
54 |
-
},
|
55 |
-
{
|
56 |
-
"epoch": 2.03,
|
57 |
-
"learning_rate": 2.6212499999999997e-05,
|
58 |
-
"loss": 3.0879,
|
59 |
-
"step": 700
|
60 |
-
},
|
61 |
-
{
|
62 |
-
"epoch": 2.32,
|
63 |
-
"learning_rate": 2.99625e-05,
|
64 |
-
"loss": 3.0216,
|
65 |
-
"step": 800
|
66 |
-
},
|
67 |
-
{
|
68 |
-
"epoch": 2.61,
|
69 |
-
"learning_rate": 3.37125e-05,
|
70 |
-
"loss": 2.9595,
|
71 |
-
"step": 900
|
72 |
-
},
|
73 |
-
{
|
74 |
-
"epoch": 2.91,
|
75 |
-
"learning_rate": 3.7462499999999996e-05,
|
76 |
-
"loss": 2.9269,
|
77 |
-
"step": 1000
|
78 |
-
},
|
79 |
-
{
|
80 |
-
"epoch": 2.91,
|
81 |
-
"eval_loss": 2.941540241241455,
|
82 |
-
"eval_runtime": 178.4388,
|
83 |
-
"eval_samples_per_second": 25.891,
|
84 |
-
"eval_steps_per_second": 3.239,
|
85 |
-
"eval_wer": 0.9966346448020559,
|
86 |
-
"step": 1000
|
87 |
-
},
|
88 |
-
{
|
89 |
-
"epoch": 3.2,
|
90 |
-
"learning_rate": 4.12125e-05,
|
91 |
-
"loss": 2.914,
|
92 |
-
"step": 1100
|
93 |
-
},
|
94 |
-
{
|
95 |
-
"epoch": 3.49,
|
96 |
-
"learning_rate": 4.4962499999999995e-05,
|
97 |
-
"loss": 2.8432,
|
98 |
-
"step": 1200
|
99 |
-
},
|
100 |
-
{
|
101 |
-
"epoch": 3.78,
|
102 |
-
"learning_rate": 4.871249999999999e-05,
|
103 |
-
"loss": 2.6828,
|
104 |
-
"step": 1300
|
105 |
-
},
|
106 |
-
{
|
107 |
-
"epoch": 4.07,
|
108 |
-
"learning_rate": 5.2462499999999994e-05,
|
109 |
-
"loss": 2.355,
|
110 |
-
"step": 1400
|
111 |
-
},
|
112 |
-
{
|
113 |
-
"epoch": 4.36,
|
114 |
-
"learning_rate": 5.62125e-05,
|
115 |
-
"loss": 2.0719,
|
116 |
-
"step": 1500
|
117 |
-
},
|
118 |
-
{
|
119 |
-
"epoch": 4.36,
|
120 |
-
"eval_loss": 1.164096713066101,
|
121 |
-
"eval_runtime": 134.2981,
|
122 |
-
"eval_samples_per_second": 34.401,
|
123 |
-
"eval_steps_per_second": 4.304,
|
124 |
-
"eval_wer": 0.8507923881784251,
|
125 |
-
"step": 1500
|
126 |
-
}
|
127 |
-
],
|
128 |
-
"max_steps": 17200,
|
129 |
-
"num_train_epochs": 50,
|
130 |
-
"total_flos": 4.445118706030802e+18,
|
131 |
-
"trial_name": null,
|
132 |
-
"trial_params": null
|
133 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
{checkpoint-1000 β checkpoint-16000}/config.json
RENAMED
File without changes
|
{checkpoint-1000 β checkpoint-16000}/optimizer.pt
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 2490362385
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:23505cec8634d6a983156ca9aaadbd850d5f159d59400b96c32cdb115a378ee4
|
3 |
size 2490362385
|
{checkpoint-1000 β checkpoint-16000}/preprocessor_config.json
RENAMED
File without changes
|
{checkpoint-1500 β checkpoint-16000}/pytorch_model.bin
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 1262075377
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4470667bcf97fe5d54dee8281ff4760245200a1adf08a6e672825247d4a08b79
|
3 |
size 1262075377
|
{checkpoint-500 β checkpoint-16000}/rng_state.pth
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:61b1412288f80f87d3b78af99d5eeabfcc510fdb01676bb0cd4bcc6470be1c10
|
3 |
+
size 14567
|
{checkpoint-500 β checkpoint-16000}/scaler.pt
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 559
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8bcb42b3bf1898edd5501a8a54e2a8aeeb1c707564ea3cf64a8e4a8749b091df
|
3 |
size 559
|
{checkpoint-1000 β checkpoint-16000}/scheduler.pt
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 623
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:812531ae10018fac0bc1fdeb95a26197edbd60dd570f21800a631a5c4b646bc9
|
3 |
size 623
|
checkpoint-16000/trainer_state.json
ADDED
@@ -0,0 +1,1264 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 46.51051486584481,
|
5 |
+
"global_step": 16000,
|
6 |
+
"is_hyper_param_search": false,
|
7 |
+
"is_local_process_zero": true,
|
8 |
+
"is_world_process_zero": true,
|
9 |
+
"log_history": [
|
10 |
+
{
|
11 |
+
"epoch": 0.29,
|
12 |
+
"learning_rate": 3.7125e-06,
|
13 |
+
"loss": 12.2713,
|
14 |
+
"step": 100
|
15 |
+
},
|
16 |
+
{
|
17 |
+
"epoch": 0.58,
|
18 |
+
"learning_rate": 7.4625e-06,
|
19 |
+
"loss": 6.2026,
|
20 |
+
"step": 200
|
21 |
+
},
|
22 |
+
{
|
23 |
+
"epoch": 0.87,
|
24 |
+
"learning_rate": 1.1212499999999998e-05,
|
25 |
+
"loss": 4.1374,
|
26 |
+
"step": 300
|
27 |
+
},
|
28 |
+
{
|
29 |
+
"epoch": 1.16,
|
30 |
+
"learning_rate": 1.49625e-05,
|
31 |
+
"loss": 3.6755,
|
32 |
+
"step": 400
|
33 |
+
},
|
34 |
+
{
|
35 |
+
"epoch": 1.45,
|
36 |
+
"learning_rate": 1.8712499999999997e-05,
|
37 |
+
"loss": 3.3332,
|
38 |
+
"step": 500
|
39 |
+
},
|
40 |
+
{
|
41 |
+
"epoch": 1.45,
|
42 |
+
"eval_loss": 3.292029857635498,
|
43 |
+
"eval_runtime": 137.2974,
|
44 |
+
"eval_samples_per_second": 33.65,
|
45 |
+
"eval_steps_per_second": 4.21,
|
46 |
+
"eval_wer": 1.0,
|
47 |
+
"step": 500
|
48 |
+
},
|
49 |
+
{
|
50 |
+
"epoch": 1.74,
|
51 |
+
"learning_rate": 2.2462499999999997e-05,
|
52 |
+
"loss": 3.1418,
|
53 |
+
"step": 600
|
54 |
+
},
|
55 |
+
{
|
56 |
+
"epoch": 2.03,
|
57 |
+
"learning_rate": 2.6212499999999997e-05,
|
58 |
+
"loss": 3.0879,
|
59 |
+
"step": 700
|
60 |
+
},
|
61 |
+
{
|
62 |
+
"epoch": 2.32,
|
63 |
+
"learning_rate": 2.99625e-05,
|
64 |
+
"loss": 3.0216,
|
65 |
+
"step": 800
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 2.61,
|
69 |
+
"learning_rate": 3.37125e-05,
|
70 |
+
"loss": 2.9595,
|
71 |
+
"step": 900
|
72 |
+
},
|
73 |
+
{
|
74 |
+
"epoch": 2.91,
|
75 |
+
"learning_rate": 3.7462499999999996e-05,
|
76 |
+
"loss": 2.9269,
|
77 |
+
"step": 1000
|
78 |
+
},
|
79 |
+
{
|
80 |
+
"epoch": 2.91,
|
81 |
+
"eval_loss": 2.941540241241455,
|
82 |
+
"eval_runtime": 178.4388,
|
83 |
+
"eval_samples_per_second": 25.891,
|
84 |
+
"eval_steps_per_second": 3.239,
|
85 |
+
"eval_wer": 0.9966346448020559,
|
86 |
+
"step": 1000
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 3.2,
|
90 |
+
"learning_rate": 4.12125e-05,
|
91 |
+
"loss": 2.914,
|
92 |
+
"step": 1100
|
93 |
+
},
|
94 |
+
{
|
95 |
+
"epoch": 3.49,
|
96 |
+
"learning_rate": 4.4962499999999995e-05,
|
97 |
+
"loss": 2.8432,
|
98 |
+
"step": 1200
|
99 |
+
},
|
100 |
+
{
|
101 |
+
"epoch": 3.78,
|
102 |
+
"learning_rate": 4.871249999999999e-05,
|
103 |
+
"loss": 2.6828,
|
104 |
+
"step": 1300
|
105 |
+
},
|
106 |
+
{
|
107 |
+
"epoch": 4.07,
|
108 |
+
"learning_rate": 5.2462499999999994e-05,
|
109 |
+
"loss": 2.355,
|
110 |
+
"step": 1400
|
111 |
+
},
|
112 |
+
{
|
113 |
+
"epoch": 4.36,
|
114 |
+
"learning_rate": 5.62125e-05,
|
115 |
+
"loss": 2.0719,
|
116 |
+
"step": 1500
|
117 |
+
},
|
118 |
+
{
|
119 |
+
"epoch": 4.36,
|
120 |
+
"eval_loss": 1.164096713066101,
|
121 |
+
"eval_runtime": 134.2981,
|
122 |
+
"eval_samples_per_second": 34.401,
|
123 |
+
"eval_steps_per_second": 4.304,
|
124 |
+
"eval_wer": 0.8507923881784251,
|
125 |
+
"step": 1500
|
126 |
+
},
|
127 |
+
{
|
128 |
+
"epoch": 4.65,
|
129 |
+
"learning_rate": 5.9962499999999994e-05,
|
130 |
+
"loss": 1.944,
|
131 |
+
"step": 1600
|
132 |
+
},
|
133 |
+
{
|
134 |
+
"epoch": 4.94,
|
135 |
+
"learning_rate": 6.37125e-05,
|
136 |
+
"loss": 1.8772,
|
137 |
+
"step": 1700
|
138 |
+
},
|
139 |
+
{
|
140 |
+
"epoch": 5.23,
|
141 |
+
"learning_rate": 6.746249999999999e-05,
|
142 |
+
"loss": 1.807,
|
143 |
+
"step": 1800
|
144 |
+
},
|
145 |
+
{
|
146 |
+
"epoch": 5.52,
|
147 |
+
"learning_rate": 7.121249999999999e-05,
|
148 |
+
"loss": 1.7612,
|
149 |
+
"step": 1900
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 5.81,
|
153 |
+
"learning_rate": 7.492499999999999e-05,
|
154 |
+
"loss": 1.7404,
|
155 |
+
"step": 2000
|
156 |
+
},
|
157 |
+
{
|
158 |
+
"epoch": 5.81,
|
159 |
+
"eval_loss": 0.7280541658401489,
|
160 |
+
"eval_runtime": 174.7873,
|
161 |
+
"eval_samples_per_second": 26.432,
|
162 |
+
"eval_steps_per_second": 3.307,
|
163 |
+
"eval_wer": 0.6846356238144772,
|
164 |
+
"step": 2000
|
165 |
+
},
|
166 |
+
{
|
167 |
+
"epoch": 6.1,
|
168 |
+
"learning_rate": 7.451644736842105e-05,
|
169 |
+
"loss": 1.6883,
|
170 |
+
"step": 2100
|
171 |
+
},
|
172 |
+
{
|
173 |
+
"epoch": 6.39,
|
174 |
+
"learning_rate": 7.402302631578947e-05,
|
175 |
+
"loss": 1.6463,
|
176 |
+
"step": 2200
|
177 |
+
},
|
178 |
+
{
|
179 |
+
"epoch": 6.68,
|
180 |
+
"learning_rate": 7.352960526315788e-05,
|
181 |
+
"loss": 1.6216,
|
182 |
+
"step": 2300
|
183 |
+
},
|
184 |
+
{
|
185 |
+
"epoch": 6.97,
|
186 |
+
"learning_rate": 7.30361842105263e-05,
|
187 |
+
"loss": 1.586,
|
188 |
+
"step": 2400
|
189 |
+
},
|
190 |
+
{
|
191 |
+
"epoch": 7.27,
|
192 |
+
"learning_rate": 7.254276315789473e-05,
|
193 |
+
"loss": 1.5921,
|
194 |
+
"step": 2500
|
195 |
+
},
|
196 |
+
{
|
197 |
+
"epoch": 7.27,
|
198 |
+
"eval_loss": 0.5885757803916931,
|
199 |
+
"eval_runtime": 137.296,
|
200 |
+
"eval_samples_per_second": 33.65,
|
201 |
+
"eval_steps_per_second": 4.21,
|
202 |
+
"eval_wer": 0.5146545921801383,
|
203 |
+
"step": 2500
|
204 |
+
},
|
205 |
+
{
|
206 |
+
"epoch": 7.56,
|
207 |
+
"learning_rate": 7.204934210526316e-05,
|
208 |
+
"loss": 1.5694,
|
209 |
+
"step": 2600
|
210 |
+
},
|
211 |
+
{
|
212 |
+
"epoch": 7.85,
|
213 |
+
"learning_rate": 7.155592105263157e-05,
|
214 |
+
"loss": 1.5279,
|
215 |
+
"step": 2700
|
216 |
+
},
|
217 |
+
{
|
218 |
+
"epoch": 8.14,
|
219 |
+
"learning_rate": 7.10625e-05,
|
220 |
+
"loss": 1.5435,
|
221 |
+
"step": 2800
|
222 |
+
},
|
223 |
+
{
|
224 |
+
"epoch": 8.43,
|
225 |
+
"learning_rate": 7.056907894736841e-05,
|
226 |
+
"loss": 1.5163,
|
227 |
+
"step": 2900
|
228 |
+
},
|
229 |
+
{
|
230 |
+
"epoch": 8.72,
|
231 |
+
"learning_rate": 7.007565789473684e-05,
|
232 |
+
"loss": 1.4941,
|
233 |
+
"step": 3000
|
234 |
+
},
|
235 |
+
{
|
236 |
+
"epoch": 8.72,
|
237 |
+
"eval_loss": 0.5183178186416626,
|
238 |
+
"eval_runtime": 134.3566,
|
239 |
+
"eval_samples_per_second": 34.386,
|
240 |
+
"eval_steps_per_second": 4.302,
|
241 |
+
"eval_wer": 0.5063023924616044,
|
242 |
+
"step": 3000
|
243 |
+
},
|
244 |
+
{
|
245 |
+
"epoch": 9.01,
|
246 |
+
"learning_rate": 6.958223684210525e-05,
|
247 |
+
"loss": 1.5061,
|
248 |
+
"step": 3100
|
249 |
+
},
|
250 |
+
{
|
251 |
+
"epoch": 9.3,
|
252 |
+
"learning_rate": 6.908881578947368e-05,
|
253 |
+
"loss": 1.4551,
|
254 |
+
"step": 3200
|
255 |
+
},
|
256 |
+
{
|
257 |
+
"epoch": 9.59,
|
258 |
+
"learning_rate": 6.859539473684209e-05,
|
259 |
+
"loss": 1.4674,
|
260 |
+
"step": 3300
|
261 |
+
},
|
262 |
+
{
|
263 |
+
"epoch": 9.88,
|
264 |
+
"learning_rate": 6.810197368421052e-05,
|
265 |
+
"loss": 1.4691,
|
266 |
+
"step": 3400
|
267 |
+
},
|
268 |
+
{
|
269 |
+
"epoch": 10.17,
|
270 |
+
"learning_rate": 6.760855263157895e-05,
|
271 |
+
"loss": 1.4486,
|
272 |
+
"step": 3500
|
273 |
+
},
|
274 |
+
{
|
275 |
+
"epoch": 10.17,
|
276 |
+
"eval_loss": 0.47492966055870056,
|
277 |
+
"eval_runtime": 135.1792,
|
278 |
+
"eval_samples_per_second": 34.177,
|
279 |
+
"eval_steps_per_second": 4.276,
|
280 |
+
"eval_wer": 0.46760080768524753,
|
281 |
+
"step": 3500
|
282 |
+
},
|
283 |
+
{
|
284 |
+
"epoch": 10.46,
|
285 |
+
"learning_rate": 6.711513157894736e-05,
|
286 |
+
"loss": 1.4274,
|
287 |
+
"step": 3600
|
288 |
+
},
|
289 |
+
{
|
290 |
+
"epoch": 10.75,
|
291 |
+
"learning_rate": 6.66266447368421e-05,
|
292 |
+
"loss": 1.4277,
|
293 |
+
"step": 3700
|
294 |
+
},
|
295 |
+
{
|
296 |
+
"epoch": 11.05,
|
297 |
+
"learning_rate": 6.613322368421052e-05,
|
298 |
+
"loss": 1.4445,
|
299 |
+
"step": 3800
|
300 |
+
},
|
301 |
+
{
|
302 |
+
"epoch": 11.34,
|
303 |
+
"learning_rate": 6.563980263157894e-05,
|
304 |
+
"loss": 1.4101,
|
305 |
+
"step": 3900
|
306 |
+
},
|
307 |
+
{
|
308 |
+
"epoch": 11.63,
|
309 |
+
"learning_rate": 6.514638157894736e-05,
|
310 |
+
"loss": 1.3899,
|
311 |
+
"step": 4000
|
312 |
+
},
|
313 |
+
{
|
314 |
+
"epoch": 11.63,
|
315 |
+
"eval_loss": 0.4564875364303589,
|
316 |
+
"eval_runtime": 134.5257,
|
317 |
+
"eval_samples_per_second": 34.343,
|
318 |
+
"eval_steps_per_second": 4.297,
|
319 |
+
"eval_wer": 0.4432172795692345,
|
320 |
+
"step": 4000
|
321 |
+
},
|
322 |
+
{
|
323 |
+
"epoch": 11.92,
|
324 |
+
"learning_rate": 6.465296052631578e-05,
|
325 |
+
"loss": 1.404,
|
326 |
+
"step": 4100
|
327 |
+
},
|
328 |
+
{
|
329 |
+
"epoch": 12.21,
|
330 |
+
"learning_rate": 6.415953947368421e-05,
|
331 |
+
"loss": 1.3911,
|
332 |
+
"step": 4200
|
333 |
+
},
|
334 |
+
{
|
335 |
+
"epoch": 12.5,
|
336 |
+
"learning_rate": 6.366611842105262e-05,
|
337 |
+
"loss": 1.3873,
|
338 |
+
"step": 4300
|
339 |
+
},
|
340 |
+
{
|
341 |
+
"epoch": 12.79,
|
342 |
+
"learning_rate": 6.317269736842105e-05,
|
343 |
+
"loss": 1.3712,
|
344 |
+
"step": 4400
|
345 |
+
},
|
346 |
+
{
|
347 |
+
"epoch": 13.08,
|
348 |
+
"learning_rate": 6.267927631578946e-05,
|
349 |
+
"loss": 1.3881,
|
350 |
+
"step": 4500
|
351 |
+
},
|
352 |
+
{
|
353 |
+
"epoch": 13.08,
|
354 |
+
"eval_loss": 0.43156012892723083,
|
355 |
+
"eval_runtime": 134.756,
|
356 |
+
"eval_samples_per_second": 34.284,
|
357 |
+
"eval_steps_per_second": 4.289,
|
358 |
+
"eval_wer": 0.42278039527626504,
|
359 |
+
"step": 4500
|
360 |
+
},
|
361 |
+
{
|
362 |
+
"epoch": 13.37,
|
363 |
+
"learning_rate": 6.218585526315789e-05,
|
364 |
+
"loss": 1.3538,
|
365 |
+
"step": 4600
|
366 |
+
},
|
367 |
+
{
|
368 |
+
"epoch": 13.66,
|
369 |
+
"learning_rate": 6.16924342105263e-05,
|
370 |
+
"loss": 1.355,
|
371 |
+
"step": 4700
|
372 |
+
},
|
373 |
+
{
|
374 |
+
"epoch": 13.95,
|
375 |
+
"learning_rate": 6.119901315789473e-05,
|
376 |
+
"loss": 1.341,
|
377 |
+
"step": 4800
|
378 |
+
},
|
379 |
+
{
|
380 |
+
"epoch": 14.24,
|
381 |
+
"learning_rate": 6.070559210526316e-05,
|
382 |
+
"loss": 1.3474,
|
383 |
+
"step": 4900
|
384 |
+
},
|
385 |
+
{
|
386 |
+
"epoch": 14.53,
|
387 |
+
"learning_rate": 6.021217105263157e-05,
|
388 |
+
"loss": 1.3572,
|
389 |
+
"step": 5000
|
390 |
+
},
|
391 |
+
{
|
392 |
+
"epoch": 14.53,
|
393 |
+
"eval_loss": 0.4194825291633606,
|
394 |
+
"eval_runtime": 134.9338,
|
395 |
+
"eval_samples_per_second": 34.239,
|
396 |
+
"eval_steps_per_second": 4.284,
|
397 |
+
"eval_wer": 0.3834363335984825,
|
398 |
+
"step": 5000
|
399 |
+
},
|
400 |
+
{
|
401 |
+
"epoch": 14.82,
|
402 |
+
"learning_rate": 5.971875e-05,
|
403 |
+
"loss": 1.3408,
|
404 |
+
"step": 5100
|
405 |
+
},
|
406 |
+
{
|
407 |
+
"epoch": 15.12,
|
408 |
+
"learning_rate": 5.922532894736842e-05,
|
409 |
+
"loss": 1.3357,
|
410 |
+
"step": 5200
|
411 |
+
},
|
412 |
+
{
|
413 |
+
"epoch": 15.41,
|
414 |
+
"learning_rate": 5.873190789473683e-05,
|
415 |
+
"loss": 1.3288,
|
416 |
+
"step": 5300
|
417 |
+
},
|
418 |
+
{
|
419 |
+
"epoch": 15.7,
|
420 |
+
"learning_rate": 5.823848684210526e-05,
|
421 |
+
"loss": 1.3258,
|
422 |
+
"step": 5400
|
423 |
+
},
|
424 |
+
{
|
425 |
+
"epoch": 15.99,
|
426 |
+
"learning_rate": 5.7745065789473685e-05,
|
427 |
+
"loss": 1.3261,
|
428 |
+
"step": 5500
|
429 |
+
},
|
430 |
+
{
|
431 |
+
"epoch": 15.99,
|
432 |
+
"eval_loss": 0.3974045217037201,
|
433 |
+
"eval_runtime": 134.8844,
|
434 |
+
"eval_samples_per_second": 34.252,
|
435 |
+
"eval_steps_per_second": 4.285,
|
436 |
+
"eval_wer": 0.3606742948051153,
|
437 |
+
"step": 5500
|
438 |
+
},
|
439 |
+
{
|
440 |
+
"epoch": 16.28,
|
441 |
+
"learning_rate": 5.72516447368421e-05,
|
442 |
+
"loss": 1.3273,
|
443 |
+
"step": 5600
|
444 |
+
},
|
445 |
+
{
|
446 |
+
"epoch": 16.57,
|
447 |
+
"learning_rate": 5.675822368421052e-05,
|
448 |
+
"loss": 1.3131,
|
449 |
+
"step": 5700
|
450 |
+
},
|
451 |
+
{
|
452 |
+
"epoch": 16.86,
|
453 |
+
"learning_rate": 5.6264802631578946e-05,
|
454 |
+
"loss": 1.2739,
|
455 |
+
"step": 5800
|
456 |
+
},
|
457 |
+
{
|
458 |
+
"epoch": 17.15,
|
459 |
+
"learning_rate": 5.577138157894736e-05,
|
460 |
+
"loss": 1.3141,
|
461 |
+
"step": 5900
|
462 |
+
},
|
463 |
+
{
|
464 |
+
"epoch": 17.44,
|
465 |
+
"learning_rate": 5.5277960526315786e-05,
|
466 |
+
"loss": 1.2809,
|
467 |
+
"step": 6000
|
468 |
+
},
|
469 |
+
{
|
470 |
+
"epoch": 17.44,
|
471 |
+
"eval_loss": 0.3844749331474304,
|
472 |
+
"eval_runtime": 136.3592,
|
473 |
+
"eval_samples_per_second": 33.881,
|
474 |
+
"eval_steps_per_second": 4.239,
|
475 |
+
"eval_wer": 0.3466621795264027,
|
476 |
+
"step": 6000
|
477 |
+
},
|
478 |
+
{
|
479 |
+
"epoch": 17.73,
|
480 |
+
"learning_rate": 5.4784539473684206e-05,
|
481 |
+
"loss": 1.2847,
|
482 |
+
"step": 6100
|
483 |
+
},
|
484 |
+
{
|
485 |
+
"epoch": 18.02,
|
486 |
+
"learning_rate": 5.4291118421052626e-05,
|
487 |
+
"loss": 1.303,
|
488 |
+
"step": 6200
|
489 |
+
},
|
490 |
+
{
|
491 |
+
"epoch": 18.31,
|
492 |
+
"learning_rate": 5.3797697368421046e-05,
|
493 |
+
"loss": 1.2733,
|
494 |
+
"step": 6300
|
495 |
+
},
|
496 |
+
{
|
497 |
+
"epoch": 18.6,
|
498 |
+
"learning_rate": 5.330427631578947e-05,
|
499 |
+
"loss": 1.2707,
|
500 |
+
"step": 6400
|
501 |
+
},
|
502 |
+
{
|
503 |
+
"epoch": 18.89,
|
504 |
+
"learning_rate": 5.2810855263157886e-05,
|
505 |
+
"loss": 1.2713,
|
506 |
+
"step": 6500
|
507 |
+
},
|
508 |
+
{
|
509 |
+
"epoch": 18.89,
|
510 |
+
"eval_loss": 0.3831607401371002,
|
511 |
+
"eval_runtime": 138.2456,
|
512 |
+
"eval_samples_per_second": 33.419,
|
513 |
+
"eval_steps_per_second": 4.181,
|
514 |
+
"eval_wer": 0.3449795019274307,
|
515 |
+
"step": 6500
|
516 |
+
},
|
517 |
+
{
|
518 |
+
"epoch": 19.19,
|
519 |
+
"learning_rate": 5.231743421052631e-05,
|
520 |
+
"loss": 1.2562,
|
521 |
+
"step": 6600
|
522 |
+
},
|
523 |
+
{
|
524 |
+
"epoch": 19.48,
|
525 |
+
"learning_rate": 5.182894736842105e-05,
|
526 |
+
"loss": 1.2721,
|
527 |
+
"step": 6700
|
528 |
+
},
|
529 |
+
{
|
530 |
+
"epoch": 19.77,
|
531 |
+
"learning_rate": 5.133552631578947e-05,
|
532 |
+
"loss": 1.2536,
|
533 |
+
"step": 6800
|
534 |
+
},
|
535 |
+
{
|
536 |
+
"epoch": 20.06,
|
537 |
+
"learning_rate": 5.084210526315789e-05,
|
538 |
+
"loss": 1.2503,
|
539 |
+
"step": 6900
|
540 |
+
},
|
541 |
+
{
|
542 |
+
"epoch": 20.35,
|
543 |
+
"learning_rate": 5.0353618421052625e-05,
|
544 |
+
"loss": 1.257,
|
545 |
+
"step": 7000
|
546 |
+
},
|
547 |
+
{
|
548 |
+
"epoch": 20.35,
|
549 |
+
"eval_loss": 0.3778809607028961,
|
550 |
+
"eval_runtime": 138.2403,
|
551 |
+
"eval_samples_per_second": 33.42,
|
552 |
+
"eval_steps_per_second": 4.181,
|
553 |
+
"eval_wer": 0.33726977911032247,
|
554 |
+
"step": 7000
|
555 |
+
},
|
556 |
+
{
|
557 |
+
"epoch": 20.64,
|
558 |
+
"learning_rate": 4.9860197368421045e-05,
|
559 |
+
"loss": 1.2557,
|
560 |
+
"step": 7100
|
561 |
+
},
|
562 |
+
{
|
563 |
+
"epoch": 20.93,
|
564 |
+
"learning_rate": 4.936677631578947e-05,
|
565 |
+
"loss": 1.2397,
|
566 |
+
"step": 7200
|
567 |
+
},
|
568 |
+
{
|
569 |
+
"epoch": 21.22,
|
570 |
+
"learning_rate": 4.8873355263157886e-05,
|
571 |
+
"loss": 1.2527,
|
572 |
+
"step": 7300
|
573 |
+
},
|
574 |
+
{
|
575 |
+
"epoch": 21.51,
|
576 |
+
"learning_rate": 4.837993421052631e-05,
|
577 |
+
"loss": 1.2273,
|
578 |
+
"step": 7400
|
579 |
+
},
|
580 |
+
{
|
581 |
+
"epoch": 21.8,
|
582 |
+
"learning_rate": 4.788651315789473e-05,
|
583 |
+
"loss": 1.2298,
|
584 |
+
"step": 7500
|
585 |
+
},
|
586 |
+
{
|
587 |
+
"epoch": 21.8,
|
588 |
+
"eval_loss": 0.3743567168712616,
|
589 |
+
"eval_runtime": 136.1375,
|
590 |
+
"eval_samples_per_second": 33.936,
|
591 |
+
"eval_steps_per_second": 4.246,
|
592 |
+
"eval_wer": 0.33913602153827327,
|
593 |
+
"step": 7500
|
594 |
+
},
|
595 |
+
{
|
596 |
+
"epoch": 22.09,
|
597 |
+
"learning_rate": 4.739309210526315e-05,
|
598 |
+
"loss": 1.2444,
|
599 |
+
"step": 7600
|
600 |
+
},
|
601 |
+
{
|
602 |
+
"epoch": 22.38,
|
603 |
+
"learning_rate": 4.689967105263157e-05,
|
604 |
+
"loss": 1.2153,
|
605 |
+
"step": 7700
|
606 |
+
},
|
607 |
+
{
|
608 |
+
"epoch": 22.67,
|
609 |
+
"learning_rate": 4.640625e-05,
|
610 |
+
"loss": 1.2327,
|
611 |
+
"step": 7800
|
612 |
+
},
|
613 |
+
{
|
614 |
+
"epoch": 22.96,
|
615 |
+
"learning_rate": 4.591282894736841e-05,
|
616 |
+
"loss": 1.2404,
|
617 |
+
"step": 7900
|
618 |
+
},
|
619 |
+
{
|
620 |
+
"epoch": 23.26,
|
621 |
+
"learning_rate": 4.541940789473684e-05,
|
622 |
+
"loss": 1.2173,
|
623 |
+
"step": 8000
|
624 |
+
},
|
625 |
+
{
|
626 |
+
"epoch": 23.26,
|
627 |
+
"eval_loss": 0.3745496869087219,
|
628 |
+
"eval_runtime": 139.3565,
|
629 |
+
"eval_samples_per_second": 33.152,
|
630 |
+
"eval_steps_per_second": 4.148,
|
631 |
+
"eval_wer": 0.32619470109527016,
|
632 |
+
"step": 8000
|
633 |
+
},
|
634 |
+
{
|
635 |
+
"epoch": 23.55,
|
636 |
+
"learning_rate": 4.492598684210526e-05,
|
637 |
+
"loss": 1.222,
|
638 |
+
"step": 8100
|
639 |
+
},
|
640 |
+
{
|
641 |
+
"epoch": 23.84,
|
642 |
+
"learning_rate": 4.443256578947369e-05,
|
643 |
+
"loss": 1.1868,
|
644 |
+
"step": 8200
|
645 |
+
},
|
646 |
+
{
|
647 |
+
"epoch": 24.13,
|
648 |
+
"learning_rate": 4.39391447368421e-05,
|
649 |
+
"loss": 1.2272,
|
650 |
+
"step": 8300
|
651 |
+
},
|
652 |
+
{
|
653 |
+
"epoch": 24.42,
|
654 |
+
"learning_rate": 4.344572368421052e-05,
|
655 |
+
"loss": 1.2212,
|
656 |
+
"step": 8400
|
657 |
+
},
|
658 |
+
{
|
659 |
+
"epoch": 24.71,
|
660 |
+
"learning_rate": 4.295230263157895e-05,
|
661 |
+
"loss": 1.1966,
|
662 |
+
"step": 8500
|
663 |
+
},
|
664 |
+
{
|
665 |
+
"epoch": 24.71,
|
666 |
+
"eval_loss": 0.36800575256347656,
|
667 |
+
"eval_runtime": 137.1685,
|
668 |
+
"eval_samples_per_second": 33.681,
|
669 |
+
"eval_steps_per_second": 4.214,
|
670 |
+
"eval_wer": 0.32411429970017747,
|
671 |
+
"step": 8500
|
672 |
+
},
|
673 |
+
{
|
674 |
+
"epoch": 25.0,
|
675 |
+
"learning_rate": 4.245888157894736e-05,
|
676 |
+
"loss": 1.2128,
|
677 |
+
"step": 8600
|
678 |
+
},
|
679 |
+
{
|
680 |
+
"epoch": 25.29,
|
681 |
+
"learning_rate": 4.196546052631579e-05,
|
682 |
+
"loss": 1.2118,
|
683 |
+
"step": 8700
|
684 |
+
},
|
685 |
+
{
|
686 |
+
"epoch": 25.58,
|
687 |
+
"learning_rate": 4.147203947368421e-05,
|
688 |
+
"loss": 1.2009,
|
689 |
+
"step": 8800
|
690 |
+
},
|
691 |
+
{
|
692 |
+
"epoch": 25.87,
|
693 |
+
"learning_rate": 4.097861842105263e-05,
|
694 |
+
"loss": 1.1684,
|
695 |
+
"step": 8900
|
696 |
+
},
|
697 |
+
{
|
698 |
+
"epoch": 26.16,
|
699 |
+
"learning_rate": 4.048519736842105e-05,
|
700 |
+
"loss": 1.1925,
|
701 |
+
"step": 9000
|
702 |
+
},
|
703 |
+
{
|
704 |
+
"epoch": 26.16,
|
705 |
+
"eval_loss": 0.3604615330696106,
|
706 |
+
"eval_runtime": 153.4534,
|
707 |
+
"eval_samples_per_second": 30.107,
|
708 |
+
"eval_steps_per_second": 3.767,
|
709 |
+
"eval_wer": 0.31713883619898425,
|
710 |
+
"step": 9000
|
711 |
+
},
|
712 |
+
{
|
713 |
+
"epoch": 26.45,
|
714 |
+
"learning_rate": 3.9991776315789475e-05,
|
715 |
+
"loss": 1.1759,
|
716 |
+
"step": 9100
|
717 |
+
},
|
718 |
+
{
|
719 |
+
"epoch": 26.74,
|
720 |
+
"learning_rate": 3.949835526315789e-05,
|
721 |
+
"loss": 1.1754,
|
722 |
+
"step": 9200
|
723 |
+
},
|
724 |
+
{
|
725 |
+
"epoch": 27.03,
|
726 |
+
"learning_rate": 3.9004934210526315e-05,
|
727 |
+
"loss": 1.198,
|
728 |
+
"step": 9300
|
729 |
+
},
|
730 |
+
{
|
731 |
+
"epoch": 27.32,
|
732 |
+
"learning_rate": 3.8511513157894735e-05,
|
733 |
+
"loss": 1.1602,
|
734 |
+
"step": 9400
|
735 |
+
},
|
736 |
+
{
|
737 |
+
"epoch": 27.61,
|
738 |
+
"learning_rate": 3.801809210526315e-05,
|
739 |
+
"loss": 1.1692,
|
740 |
+
"step": 9500
|
741 |
+
},
|
742 |
+
{
|
743 |
+
"epoch": 27.61,
|
744 |
+
"eval_loss": 0.3512294292449951,
|
745 |
+
"eval_runtime": 166.9619,
|
746 |
+
"eval_samples_per_second": 27.671,
|
747 |
+
"eval_steps_per_second": 3.462,
|
748 |
+
"eval_wer": 0.31472189928409716,
|
749 |
+
"step": 9500
|
750 |
+
},
|
751 |
+
{
|
752 |
+
"epoch": 27.91,
|
753 |
+
"learning_rate": 3.7524671052631575e-05,
|
754 |
+
"loss": 1.1495,
|
755 |
+
"step": 9600
|
756 |
+
},
|
757 |
+
{
|
758 |
+
"epoch": 28.2,
|
759 |
+
"learning_rate": 3.7031249999999995e-05,
|
760 |
+
"loss": 1.1722,
|
761 |
+
"step": 9700
|
762 |
+
},
|
763 |
+
{
|
764 |
+
"epoch": 28.49,
|
765 |
+
"learning_rate": 3.6537828947368416e-05,
|
766 |
+
"loss": 1.1443,
|
767 |
+
"step": 9800
|
768 |
+
},
|
769 |
+
{
|
770 |
+
"epoch": 28.78,
|
771 |
+
"learning_rate": 3.6044407894736836e-05,
|
772 |
+
"loss": 1.1535,
|
773 |
+
"step": 9900
|
774 |
+
},
|
775 |
+
{
|
776 |
+
"epoch": 29.07,
|
777 |
+
"learning_rate": 3.555098684210526e-05,
|
778 |
+
"loss": 1.1704,
|
779 |
+
"step": 10000
|
780 |
+
},
|
781 |
+
{
|
782 |
+
"epoch": 29.07,
|
783 |
+
"eval_loss": 0.3532196581363678,
|
784 |
+
"eval_runtime": 134.5452,
|
785 |
+
"eval_samples_per_second": 34.338,
|
786 |
+
"eval_steps_per_second": 4.296,
|
787 |
+
"eval_wer": 0.30976564890167047,
|
788 |
+
"step": 10000
|
789 |
+
},
|
790 |
+
{
|
791 |
+
"epoch": 29.36,
|
792 |
+
"learning_rate": 3.505756578947368e-05,
|
793 |
+
"loss": 1.1519,
|
794 |
+
"step": 10100
|
795 |
+
},
|
796 |
+
{
|
797 |
+
"epoch": 29.65,
|
798 |
+
"learning_rate": 3.45641447368421e-05,
|
799 |
+
"loss": 1.1507,
|
800 |
+
"step": 10200
|
801 |
+
},
|
802 |
+
{
|
803 |
+
"epoch": 29.94,
|
804 |
+
"learning_rate": 3.407072368421052e-05,
|
805 |
+
"loss": 1.1517,
|
806 |
+
"step": 10300
|
807 |
+
},
|
808 |
+
{
|
809 |
+
"epoch": 30.23,
|
810 |
+
"learning_rate": 3.357730263157894e-05,
|
811 |
+
"loss": 1.1447,
|
812 |
+
"step": 10400
|
813 |
+
},
|
814 |
+
{
|
815 |
+
"epoch": 30.52,
|
816 |
+
"learning_rate": 3.308388157894737e-05,
|
817 |
+
"loss": 1.1595,
|
818 |
+
"step": 10500
|
819 |
+
},
|
820 |
+
{
|
821 |
+
"epoch": 30.52,
|
822 |
+
"eval_loss": 0.34246256947517395,
|
823 |
+
"eval_runtime": 159.7393,
|
824 |
+
"eval_samples_per_second": 28.922,
|
825 |
+
"eval_steps_per_second": 3.618,
|
826 |
+
"eval_wer": 0.3038915743743499,
|
827 |
+
"step": 10500
|
828 |
+
},
|
829 |
+
{
|
830 |
+
"epoch": 30.81,
|
831 |
+
"learning_rate": 3.259046052631579e-05,
|
832 |
+
"loss": 1.1451,
|
833 |
+
"step": 10600
|
834 |
+
},
|
835 |
+
{
|
836 |
+
"epoch": 31.1,
|
837 |
+
"learning_rate": 3.2097039473684203e-05,
|
838 |
+
"loss": 1.1304,
|
839 |
+
"step": 10700
|
840 |
+
},
|
841 |
+
{
|
842 |
+
"epoch": 31.39,
|
843 |
+
"learning_rate": 3.160361842105263e-05,
|
844 |
+
"loss": 1.1316,
|
845 |
+
"step": 10800
|
846 |
+
},
|
847 |
+
{
|
848 |
+
"epoch": 31.68,
|
849 |
+
"learning_rate": 3.111019736842105e-05,
|
850 |
+
"loss": 1.1301,
|
851 |
+
"step": 10900
|
852 |
+
},
|
853 |
+
{
|
854 |
+
"epoch": 31.97,
|
855 |
+
"learning_rate": 3.061677631578947e-05,
|
856 |
+
"loss": 1.1433,
|
857 |
+
"step": 11000
|
858 |
+
},
|
859 |
+
{
|
860 |
+
"epoch": 31.97,
|
861 |
+
"eval_loss": 0.35683709383010864,
|
862 |
+
"eval_runtime": 158.7503,
|
863 |
+
"eval_samples_per_second": 29.102,
|
864 |
+
"eval_steps_per_second": 3.641,
|
865 |
+
"eval_wer": 0.30257602643333537,
|
866 |
+
"step": 11000
|
867 |
+
},
|
868 |
+
{
|
869 |
+
"epoch": 32.27,
|
870 |
+
"learning_rate": 3.0123355263157894e-05,
|
871 |
+
"loss": 1.1337,
|
872 |
+
"step": 11100
|
873 |
+
},
|
874 |
+
{
|
875 |
+
"epoch": 32.56,
|
876 |
+
"learning_rate": 2.9629934210526314e-05,
|
877 |
+
"loss": 1.1227,
|
878 |
+
"step": 11200
|
879 |
+
},
|
880 |
+
{
|
881 |
+
"epoch": 32.85,
|
882 |
+
"learning_rate": 2.913651315789473e-05,
|
883 |
+
"loss": 1.1281,
|
884 |
+
"step": 11300
|
885 |
+
},
|
886 |
+
{
|
887 |
+
"epoch": 33.14,
|
888 |
+
"learning_rate": 2.8643092105263154e-05,
|
889 |
+
"loss": 1.1203,
|
890 |
+
"step": 11400
|
891 |
+
},
|
892 |
+
{
|
893 |
+
"epoch": 33.43,
|
894 |
+
"learning_rate": 2.8149671052631574e-05,
|
895 |
+
"loss": 1.1295,
|
896 |
+
"step": 11500
|
897 |
+
},
|
898 |
+
{
|
899 |
+
"epoch": 33.43,
|
900 |
+
"eval_loss": 0.34605443477630615,
|
901 |
+
"eval_runtime": 160.1749,
|
902 |
+
"eval_samples_per_second": 28.843,
|
903 |
+
"eval_steps_per_second": 3.609,
|
904 |
+
"eval_wer": 0.2992106712353913,
|
905 |
+
"step": 11500
|
906 |
+
},
|
907 |
+
{
|
908 |
+
"epoch": 33.72,
|
909 |
+
"learning_rate": 2.7656249999999998e-05,
|
910 |
+
"loss": 1.0915,
|
911 |
+
"step": 11600
|
912 |
+
},
|
913 |
+
{
|
914 |
+
"epoch": 34.01,
|
915 |
+
"learning_rate": 2.7162828947368418e-05,
|
916 |
+
"loss": 1.1312,
|
917 |
+
"step": 11700
|
918 |
+
},
|
919 |
+
{
|
920 |
+
"epoch": 34.3,
|
921 |
+
"learning_rate": 2.6669407894736838e-05,
|
922 |
+
"loss": 1.1146,
|
923 |
+
"step": 11800
|
924 |
+
},
|
925 |
+
{
|
926 |
+
"epoch": 34.59,
|
927 |
+
"learning_rate": 2.6175986842105262e-05,
|
928 |
+
"loss": 1.1184,
|
929 |
+
"step": 11900
|
930 |
+
},
|
931 |
+
{
|
932 |
+
"epoch": 34.88,
|
933 |
+
"learning_rate": 2.56875e-05,
|
934 |
+
"loss": 1.1131,
|
935 |
+
"step": 12000
|
936 |
+
},
|
937 |
+
{
|
938 |
+
"epoch": 34.88,
|
939 |
+
"eval_loss": 0.3348712623119354,
|
940 |
+
"eval_runtime": 157.6811,
|
941 |
+
"eval_samples_per_second": 29.3,
|
942 |
+
"eval_steps_per_second": 3.666,
|
943 |
+
"eval_wer": 0.29422382671480146,
|
944 |
+
"step": 12000
|
945 |
+
},
|
946 |
+
{
|
947 |
+
"epoch": 35.17,
|
948 |
+
"learning_rate": 2.5194078947368418e-05,
|
949 |
+
"loss": 1.1025,
|
950 |
+
"step": 12100
|
951 |
+
},
|
952 |
+
{
|
953 |
+
"epoch": 35.46,
|
954 |
+
"learning_rate": 2.4700657894736838e-05,
|
955 |
+
"loss": 1.1069,
|
956 |
+
"step": 12200
|
957 |
+
},
|
958 |
+
{
|
959 |
+
"epoch": 35.75,
|
960 |
+
"learning_rate": 2.420723684210526e-05,
|
961 |
+
"loss": 1.076,
|
962 |
+
"step": 12300
|
963 |
+
},
|
964 |
+
{
|
965 |
+
"epoch": 36.05,
|
966 |
+
"learning_rate": 2.371381578947368e-05,
|
967 |
+
"loss": 1.111,
|
968 |
+
"step": 12400
|
969 |
+
},
|
970 |
+
{
|
971 |
+
"epoch": 36.34,
|
972 |
+
"learning_rate": 2.32203947368421e-05,
|
973 |
+
"loss": 1.1015,
|
974 |
+
"step": 12500
|
975 |
+
},
|
976 |
+
{
|
977 |
+
"epoch": 36.34,
|
978 |
+
"eval_loss": 0.3378337025642395,
|
979 |
+
"eval_runtime": 160.9202,
|
980 |
+
"eval_samples_per_second": 28.71,
|
981 |
+
"eval_steps_per_second": 3.592,
|
982 |
+
"eval_wer": 0.29612066328091535,
|
983 |
+
"step": 12500
|
984 |
+
},
|
985 |
+
{
|
986 |
+
"epoch": 36.63,
|
987 |
+
"learning_rate": 2.2726973684210525e-05,
|
988 |
+
"loss": 1.1047,
|
989 |
+
"step": 12600
|
990 |
+
},
|
991 |
+
{
|
992 |
+
"epoch": 36.92,
|
993 |
+
"learning_rate": 2.2233552631578945e-05,
|
994 |
+
"loss": 1.102,
|
995 |
+
"step": 12700
|
996 |
+
},
|
997 |
+
{
|
998 |
+
"epoch": 37.21,
|
999 |
+
"learning_rate": 2.1740131578947365e-05,
|
1000 |
+
"loss": 1.095,
|
1001 |
+
"step": 12800
|
1002 |
+
},
|
1003 |
+
{
|
1004 |
+
"epoch": 37.5,
|
1005 |
+
"learning_rate": 2.124671052631579e-05,
|
1006 |
+
"loss": 1.0709,
|
1007 |
+
"step": 12900
|
1008 |
+
},
|
1009 |
+
{
|
1010 |
+
"epoch": 37.79,
|
1011 |
+
"learning_rate": 2.075328947368421e-05,
|
1012 |
+
"loss": 1.0835,
|
1013 |
+
"step": 13000
|
1014 |
+
},
|
1015 |
+
{
|
1016 |
+
"epoch": 37.79,
|
1017 |
+
"eval_loss": 0.3281959593296051,
|
1018 |
+
"eval_runtime": 159.3279,
|
1019 |
+
"eval_samples_per_second": 28.997,
|
1020 |
+
"eval_steps_per_second": 3.628,
|
1021 |
+
"eval_wer": 0.2865141038976932,
|
1022 |
+
"step": 13000
|
1023 |
+
},
|
1024 |
+
{
|
1025 |
+
"epoch": 38.08,
|
1026 |
+
"learning_rate": 2.0259868421052632e-05,
|
1027 |
+
"loss": 1.0846,
|
1028 |
+
"step": 13100
|
1029 |
+
},
|
1030 |
+
{
|
1031 |
+
"epoch": 38.37,
|
1032 |
+
"learning_rate": 1.9766447368421053e-05,
|
1033 |
+
"loss": 1.0933,
|
1034 |
+
"step": 13200
|
1035 |
+
},
|
1036 |
+
{
|
1037 |
+
"epoch": 38.66,
|
1038 |
+
"learning_rate": 1.927302631578947e-05,
|
1039 |
+
"loss": 1.0661,
|
1040 |
+
"step": 13300
|
1041 |
+
},
|
1042 |
+
{
|
1043 |
+
"epoch": 38.95,
|
1044 |
+
"learning_rate": 1.8779605263157893e-05,
|
1045 |
+
"loss": 1.0902,
|
1046 |
+
"step": 13400
|
1047 |
+
},
|
1048 |
+
{
|
1049 |
+
"epoch": 39.24,
|
1050 |
+
"learning_rate": 1.8286184210526313e-05,
|
1051 |
+
"loss": 1.083,
|
1052 |
+
"step": 13500
|
1053 |
+
},
|
1054 |
+
{
|
1055 |
+
"epoch": 39.24,
|
1056 |
+
"eval_loss": 0.3181643486022949,
|
1057 |
+
"eval_runtime": 158.3527,
|
1058 |
+
"eval_samples_per_second": 29.175,
|
1059 |
+
"eval_steps_per_second": 3.65,
|
1060 |
+
"eval_wer": 0.2825674600746497,
|
1061 |
+
"step": 13500
|
1062 |
+
},
|
1063 |
+
{
|
1064 |
+
"epoch": 39.53,
|
1065 |
+
"learning_rate": 1.7792763157894736e-05,
|
1066 |
+
"loss": 1.0548,
|
1067 |
+
"step": 13600
|
1068 |
+
},
|
1069 |
+
{
|
1070 |
+
"epoch": 39.82,
|
1071 |
+
"learning_rate": 1.7299342105263156e-05,
|
1072 |
+
"loss": 1.0735,
|
1073 |
+
"step": 13700
|
1074 |
+
},
|
1075 |
+
{
|
1076 |
+
"epoch": 40.12,
|
1077 |
+
"learning_rate": 1.6805921052631577e-05,
|
1078 |
+
"loss": 1.085,
|
1079 |
+
"step": 13800
|
1080 |
+
},
|
1081 |
+
{
|
1082 |
+
"epoch": 40.41,
|
1083 |
+
"learning_rate": 1.63125e-05,
|
1084 |
+
"loss": 1.0531,
|
1085 |
+
"step": 13900
|
1086 |
+
},
|
1087 |
+
{
|
1088 |
+
"epoch": 40.7,
|
1089 |
+
"learning_rate": 1.581907894736842e-05,
|
1090 |
+
"loss": 1.0819,
|
1091 |
+
"step": 14000
|
1092 |
+
},
|
1093 |
+
{
|
1094 |
+
"epoch": 40.7,
|
1095 |
+
"eval_loss": 0.32643991708755493,
|
1096 |
+
"eval_runtime": 159.1099,
|
1097 |
+
"eval_samples_per_second": 29.037,
|
1098 |
+
"eval_steps_per_second": 3.633,
|
1099 |
+
"eval_wer": 0.28504558526586304,
|
1100 |
+
"step": 14000
|
1101 |
+
},
|
1102 |
+
{
|
1103 |
+
"epoch": 40.99,
|
1104 |
+
"learning_rate": 1.532565789473684e-05,
|
1105 |
+
"loss": 1.0679,
|
1106 |
+
"step": 14100
|
1107 |
+
},
|
1108 |
+
{
|
1109 |
+
"epoch": 41.28,
|
1110 |
+
"learning_rate": 1.4832236842105262e-05,
|
1111 |
+
"loss": 1.0773,
|
1112 |
+
"step": 14200
|
1113 |
+
},
|
1114 |
+
{
|
1115 |
+
"epoch": 41.57,
|
1116 |
+
"learning_rate": 1.4338815789473682e-05,
|
1117 |
+
"loss": 1.0587,
|
1118 |
+
"step": 14300
|
1119 |
+
},
|
1120 |
+
{
|
1121 |
+
"epoch": 41.86,
|
1122 |
+
"learning_rate": 1.3845394736842104e-05,
|
1123 |
+
"loss": 1.0287,
|
1124 |
+
"step": 14400
|
1125 |
+
},
|
1126 |
+
{
|
1127 |
+
"epoch": 42.15,
|
1128 |
+
"learning_rate": 1.3351973684210524e-05,
|
1129 |
+
"loss": 1.072,
|
1130 |
+
"step": 14500
|
1131 |
+
},
|
1132 |
+
{
|
1133 |
+
"epoch": 42.15,
|
1134 |
+
"eval_loss": 0.32792535424232483,
|
1135 |
+
"eval_runtime": 155.2521,
|
1136 |
+
"eval_samples_per_second": 29.758,
|
1137 |
+
"eval_steps_per_second": 3.723,
|
1138 |
+
"eval_wer": 0.2817108242060821,
|
1139 |
+
"step": 14500
|
1140 |
+
},
|
1141 |
+
{
|
1142 |
+
"epoch": 42.44,
|
1143 |
+
"learning_rate": 1.2863486842105262e-05,
|
1144 |
+
"loss": 1.0508,
|
1145 |
+
"step": 14600
|
1146 |
+
},
|
1147 |
+
{
|
1148 |
+
"epoch": 42.73,
|
1149 |
+
"learning_rate": 1.2370065789473684e-05,
|
1150 |
+
"loss": 1.0408,
|
1151 |
+
"step": 14700
|
1152 |
+
},
|
1153 |
+
{
|
1154 |
+
"epoch": 43.02,
|
1155 |
+
"learning_rate": 1.1876644736842105e-05,
|
1156 |
+
"loss": 1.0725,
|
1157 |
+
"step": 14800
|
1158 |
+
},
|
1159 |
+
{
|
1160 |
+
"epoch": 43.31,
|
1161 |
+
"learning_rate": 1.1383223684210525e-05,
|
1162 |
+
"loss": 1.0381,
|
1163 |
+
"step": 14900
|
1164 |
+
},
|
1165 |
+
{
|
1166 |
+
"epoch": 43.6,
|
1167 |
+
"learning_rate": 1.0889802631578946e-05,
|
1168 |
+
"loss": 1.0456,
|
1169 |
+
"step": 15000
|
1170 |
+
},
|
1171 |
+
{
|
1172 |
+
"epoch": 43.6,
|
1173 |
+
"eval_loss": 0.323445200920105,
|
1174 |
+
"eval_runtime": 137.3413,
|
1175 |
+
"eval_samples_per_second": 33.639,
|
1176 |
+
"eval_steps_per_second": 4.208,
|
1177 |
+
"eval_wer": 0.2792632931530319,
|
1178 |
+
"step": 15000
|
1179 |
+
},
|
1180 |
+
{
|
1181 |
+
"epoch": 43.89,
|
1182 |
+
"learning_rate": 1.0396381578947367e-05,
|
1183 |
+
"loss": 1.0655,
|
1184 |
+
"step": 15100
|
1185 |
+
},
|
1186 |
+
{
|
1187 |
+
"epoch": 44.19,
|
1188 |
+
"learning_rate": 9.90296052631579e-06,
|
1189 |
+
"loss": 1.0509,
|
1190 |
+
"step": 15200
|
1191 |
+
},
|
1192 |
+
{
|
1193 |
+
"epoch": 44.48,
|
1194 |
+
"learning_rate": 9.40953947368421e-06,
|
1195 |
+
"loss": 1.0357,
|
1196 |
+
"step": 15300
|
1197 |
+
},
|
1198 |
+
{
|
1199 |
+
"epoch": 44.77,
|
1200 |
+
"learning_rate": 8.916118421052631e-06,
|
1201 |
+
"loss": 1.0467,
|
1202 |
+
"step": 15400
|
1203 |
+
},
|
1204 |
+
{
|
1205 |
+
"epoch": 45.06,
|
1206 |
+
"learning_rate": 8.422697368421051e-06,
|
1207 |
+
"loss": 1.0581,
|
1208 |
+
"step": 15500
|
1209 |
+
},
|
1210 |
+
{
|
1211 |
+
"epoch": 45.06,
|
1212 |
+
"eval_loss": 0.32200726866722107,
|
1213 |
+
"eval_runtime": 138.9222,
|
1214 |
+
"eval_samples_per_second": 33.256,
|
1215 |
+
"eval_steps_per_second": 4.161,
|
1216 |
+
"eval_wer": 0.27791715107385423,
|
1217 |
+
"step": 15500
|
1218 |
+
},
|
1219 |
+
{
|
1220 |
+
"epoch": 45.35,
|
1221 |
+
"learning_rate": 7.929276315789473e-06,
|
1222 |
+
"loss": 1.0512,
|
1223 |
+
"step": 15600
|
1224 |
+
},
|
1225 |
+
{
|
1226 |
+
"epoch": 45.64,
|
1227 |
+
"learning_rate": 7.435855263157894e-06,
|
1228 |
+
"loss": 1.0397,
|
1229 |
+
"step": 15700
|
1230 |
+
},
|
1231 |
+
{
|
1232 |
+
"epoch": 45.93,
|
1233 |
+
"learning_rate": 6.9473684210526315e-06,
|
1234 |
+
"loss": 1.0455,
|
1235 |
+
"step": 15800
|
1236 |
+
},
|
1237 |
+
{
|
1238 |
+
"epoch": 46.22,
|
1239 |
+
"learning_rate": 6.453947368421052e-06,
|
1240 |
+
"loss": 1.0347,
|
1241 |
+
"step": 15900
|
1242 |
+
},
|
1243 |
+
{
|
1244 |
+
"epoch": 46.51,
|
1245 |
+
"learning_rate": 5.9605263157894735e-06,
|
1246 |
+
"loss": 1.0406,
|
1247 |
+
"step": 16000
|
1248 |
+
},
|
1249 |
+
{
|
1250 |
+
"epoch": 46.51,
|
1251 |
+
"eval_loss": 0.32076749205589294,
|
1252 |
+
"eval_runtime": 134.5334,
|
1253 |
+
"eval_samples_per_second": 34.341,
|
1254 |
+
"eval_steps_per_second": 4.296,
|
1255 |
+
"eval_wer": 0.2762344734748822,
|
1256 |
+
"step": 16000
|
1257 |
+
}
|
1258 |
+
],
|
1259 |
+
"max_steps": 17200,
|
1260 |
+
"num_train_epochs": 50,
|
1261 |
+
"total_flos": 4.722636215807744e+19,
|
1262 |
+
"trial_name": null,
|
1263 |
+
"trial_params": null
|
1264 |
+
}
|
{checkpoint-1000 β checkpoint-16000}/training_args.bin
RENAMED
File without changes
|
{checkpoint-1500 β checkpoint-16500}/config.json
RENAMED
File without changes
|
{checkpoint-1500 β checkpoint-16500}/optimizer.pt
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 2490362385
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8f228b737be7b0ab54382497d712270d476c95b36ecae1d727785dc7bff0978e
|
3 |
size 2490362385
|
{checkpoint-1500 β checkpoint-16500}/preprocessor_config.json
RENAMED
File without changes
|
{checkpoint-1000 β checkpoint-16500}/pytorch_model.bin
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 1262075377
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3c6a62856623bc4550b5f336bd7209b1bd2e3e670b1e7950a79cb005b3fd7534
|
3 |
size 1262075377
|
{checkpoint-1000 β checkpoint-16500}/rng_state.pth
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a27ba1fd455e94e8b7a6027e7ed9cf82cd26113cc0bd18ebed10a0b5d449392f
|
3 |
+
size 14567
|
{checkpoint-1500 β checkpoint-16500}/scaler.pt
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 559
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d9dc60172b032edcc9f74ac3e1b49feca378ebd1f3b02a4d7cfc53aef7447534
|
3 |
size 559
|
{checkpoint-500 β checkpoint-16500}/scheduler.pt
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 623
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e93700db66c6e0b98be12ee747665bf9e5cb13cb3f69d7765f6d28a5384f3c0b
|
3 |
size 623
|
checkpoint-16500/trainer_state.json
ADDED
@@ -0,0 +1,1303 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 47.963016678752716,
|
5 |
+
"global_step": 16500,
|
6 |
+
"is_hyper_param_search": false,
|
7 |
+
"is_local_process_zero": true,
|
8 |
+
"is_world_process_zero": true,
|
9 |
+
"log_history": [
|
10 |
+
{
|
11 |
+
"epoch": 0.29,
|
12 |
+
"learning_rate": 3.7125e-06,
|
13 |
+
"loss": 12.2713,
|
14 |
+
"step": 100
|
15 |
+
},
|
16 |
+
{
|
17 |
+
"epoch": 0.58,
|
18 |
+
"learning_rate": 7.4625e-06,
|
19 |
+
"loss": 6.2026,
|
20 |
+
"step": 200
|
21 |
+
},
|
22 |
+
{
|
23 |
+
"epoch": 0.87,
|
24 |
+
"learning_rate": 1.1212499999999998e-05,
|
25 |
+
"loss": 4.1374,
|
26 |
+
"step": 300
|
27 |
+
},
|
28 |
+
{
|
29 |
+
"epoch": 1.16,
|
30 |
+
"learning_rate": 1.49625e-05,
|
31 |
+
"loss": 3.6755,
|
32 |
+
"step": 400
|
33 |
+
},
|
34 |
+
{
|
35 |
+
"epoch": 1.45,
|
36 |
+
"learning_rate": 1.8712499999999997e-05,
|
37 |
+
"loss": 3.3332,
|
38 |
+
"step": 500
|
39 |
+
},
|
40 |
+
{
|
41 |
+
"epoch": 1.45,
|
42 |
+
"eval_loss": 3.292029857635498,
|
43 |
+
"eval_runtime": 137.2974,
|
44 |
+
"eval_samples_per_second": 33.65,
|
45 |
+
"eval_steps_per_second": 4.21,
|
46 |
+
"eval_wer": 1.0,
|
47 |
+
"step": 500
|
48 |
+
},
|
49 |
+
{
|
50 |
+
"epoch": 1.74,
|
51 |
+
"learning_rate": 2.2462499999999997e-05,
|
52 |
+
"loss": 3.1418,
|
53 |
+
"step": 600
|
54 |
+
},
|
55 |
+
{
|
56 |
+
"epoch": 2.03,
|
57 |
+
"learning_rate": 2.6212499999999997e-05,
|
58 |
+
"loss": 3.0879,
|
59 |
+
"step": 700
|
60 |
+
},
|
61 |
+
{
|
62 |
+
"epoch": 2.32,
|
63 |
+
"learning_rate": 2.99625e-05,
|
64 |
+
"loss": 3.0216,
|
65 |
+
"step": 800
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 2.61,
|
69 |
+
"learning_rate": 3.37125e-05,
|
70 |
+
"loss": 2.9595,
|
71 |
+
"step": 900
|
72 |
+
},
|
73 |
+
{
|
74 |
+
"epoch": 2.91,
|
75 |
+
"learning_rate": 3.7462499999999996e-05,
|
76 |
+
"loss": 2.9269,
|
77 |
+
"step": 1000
|
78 |
+
},
|
79 |
+
{
|
80 |
+
"epoch": 2.91,
|
81 |
+
"eval_loss": 2.941540241241455,
|
82 |
+
"eval_runtime": 178.4388,
|
83 |
+
"eval_samples_per_second": 25.891,
|
84 |
+
"eval_steps_per_second": 3.239,
|
85 |
+
"eval_wer": 0.9966346448020559,
|
86 |
+
"step": 1000
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 3.2,
|
90 |
+
"learning_rate": 4.12125e-05,
|
91 |
+
"loss": 2.914,
|
92 |
+
"step": 1100
|
93 |
+
},
|
94 |
+
{
|
95 |
+
"epoch": 3.49,
|
96 |
+
"learning_rate": 4.4962499999999995e-05,
|
97 |
+
"loss": 2.8432,
|
98 |
+
"step": 1200
|
99 |
+
},
|
100 |
+
{
|
101 |
+
"epoch": 3.78,
|
102 |
+
"learning_rate": 4.871249999999999e-05,
|
103 |
+
"loss": 2.6828,
|
104 |
+
"step": 1300
|
105 |
+
},
|
106 |
+
{
|
107 |
+
"epoch": 4.07,
|
108 |
+
"learning_rate": 5.2462499999999994e-05,
|
109 |
+
"loss": 2.355,
|
110 |
+
"step": 1400
|
111 |
+
},
|
112 |
+
{
|
113 |
+
"epoch": 4.36,
|
114 |
+
"learning_rate": 5.62125e-05,
|
115 |
+
"loss": 2.0719,
|
116 |
+
"step": 1500
|
117 |
+
},
|
118 |
+
{
|
119 |
+
"epoch": 4.36,
|
120 |
+
"eval_loss": 1.164096713066101,
|
121 |
+
"eval_runtime": 134.2981,
|
122 |
+
"eval_samples_per_second": 34.401,
|
123 |
+
"eval_steps_per_second": 4.304,
|
124 |
+
"eval_wer": 0.8507923881784251,
|
125 |
+
"step": 1500
|
126 |
+
},
|
127 |
+
{
|
128 |
+
"epoch": 4.65,
|
129 |
+
"learning_rate": 5.9962499999999994e-05,
|
130 |
+
"loss": 1.944,
|
131 |
+
"step": 1600
|
132 |
+
},
|
133 |
+
{
|
134 |
+
"epoch": 4.94,
|
135 |
+
"learning_rate": 6.37125e-05,
|
136 |
+
"loss": 1.8772,
|
137 |
+
"step": 1700
|
138 |
+
},
|
139 |
+
{
|
140 |
+
"epoch": 5.23,
|
141 |
+
"learning_rate": 6.746249999999999e-05,
|
142 |
+
"loss": 1.807,
|
143 |
+
"step": 1800
|
144 |
+
},
|
145 |
+
{
|
146 |
+
"epoch": 5.52,
|
147 |
+
"learning_rate": 7.121249999999999e-05,
|
148 |
+
"loss": 1.7612,
|
149 |
+
"step": 1900
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 5.81,
|
153 |
+
"learning_rate": 7.492499999999999e-05,
|
154 |
+
"loss": 1.7404,
|
155 |
+
"step": 2000
|
156 |
+
},
|
157 |
+
{
|
158 |
+
"epoch": 5.81,
|
159 |
+
"eval_loss": 0.7280541658401489,
|
160 |
+
"eval_runtime": 174.7873,
|
161 |
+
"eval_samples_per_second": 26.432,
|
162 |
+
"eval_steps_per_second": 3.307,
|
163 |
+
"eval_wer": 0.6846356238144772,
|
164 |
+
"step": 2000
|
165 |
+
},
|
166 |
+
{
|
167 |
+
"epoch": 6.1,
|
168 |
+
"learning_rate": 7.451644736842105e-05,
|
169 |
+
"loss": 1.6883,
|
170 |
+
"step": 2100
|
171 |
+
},
|
172 |
+
{
|
173 |
+
"epoch": 6.39,
|
174 |
+
"learning_rate": 7.402302631578947e-05,
|
175 |
+
"loss": 1.6463,
|
176 |
+
"step": 2200
|
177 |
+
},
|
178 |
+
{
|
179 |
+
"epoch": 6.68,
|
180 |
+
"learning_rate": 7.352960526315788e-05,
|
181 |
+
"loss": 1.6216,
|
182 |
+
"step": 2300
|
183 |
+
},
|
184 |
+
{
|
185 |
+
"epoch": 6.97,
|
186 |
+
"learning_rate": 7.30361842105263e-05,
|
187 |
+
"loss": 1.586,
|
188 |
+
"step": 2400
|
189 |
+
},
|
190 |
+
{
|
191 |
+
"epoch": 7.27,
|
192 |
+
"learning_rate": 7.254276315789473e-05,
|
193 |
+
"loss": 1.5921,
|
194 |
+
"step": 2500
|
195 |
+
},
|
196 |
+
{
|
197 |
+
"epoch": 7.27,
|
198 |
+
"eval_loss": 0.5885757803916931,
|
199 |
+
"eval_runtime": 137.296,
|
200 |
+
"eval_samples_per_second": 33.65,
|
201 |
+
"eval_steps_per_second": 4.21,
|
202 |
+
"eval_wer": 0.5146545921801383,
|
203 |
+
"step": 2500
|
204 |
+
},
|
205 |
+
{
|
206 |
+
"epoch": 7.56,
|
207 |
+
"learning_rate": 7.204934210526316e-05,
|
208 |
+
"loss": 1.5694,
|
209 |
+
"step": 2600
|
210 |
+
},
|
211 |
+
{
|
212 |
+
"epoch": 7.85,
|
213 |
+
"learning_rate": 7.155592105263157e-05,
|
214 |
+
"loss": 1.5279,
|
215 |
+
"step": 2700
|
216 |
+
},
|
217 |
+
{
|
218 |
+
"epoch": 8.14,
|
219 |
+
"learning_rate": 7.10625e-05,
|
220 |
+
"loss": 1.5435,
|
221 |
+
"step": 2800
|
222 |
+
},
|
223 |
+
{
|
224 |
+
"epoch": 8.43,
|
225 |
+
"learning_rate": 7.056907894736841e-05,
|
226 |
+
"loss": 1.5163,
|
227 |
+
"step": 2900
|
228 |
+
},
|
229 |
+
{
|
230 |
+
"epoch": 8.72,
|
231 |
+
"learning_rate": 7.007565789473684e-05,
|
232 |
+
"loss": 1.4941,
|
233 |
+
"step": 3000
|
234 |
+
},
|
235 |
+
{
|
236 |
+
"epoch": 8.72,
|
237 |
+
"eval_loss": 0.5183178186416626,
|
238 |
+
"eval_runtime": 134.3566,
|
239 |
+
"eval_samples_per_second": 34.386,
|
240 |
+
"eval_steps_per_second": 4.302,
|
241 |
+
"eval_wer": 0.5063023924616044,
|
242 |
+
"step": 3000
|
243 |
+
},
|
244 |
+
{
|
245 |
+
"epoch": 9.01,
|
246 |
+
"learning_rate": 6.958223684210525e-05,
|
247 |
+
"loss": 1.5061,
|
248 |
+
"step": 3100
|
249 |
+
},
|
250 |
+
{
|
251 |
+
"epoch": 9.3,
|
252 |
+
"learning_rate": 6.908881578947368e-05,
|
253 |
+
"loss": 1.4551,
|
254 |
+
"step": 3200
|
255 |
+
},
|
256 |
+
{
|
257 |
+
"epoch": 9.59,
|
258 |
+
"learning_rate": 6.859539473684209e-05,
|
259 |
+
"loss": 1.4674,
|
260 |
+
"step": 3300
|
261 |
+
},
|
262 |
+
{
|
263 |
+
"epoch": 9.88,
|
264 |
+
"learning_rate": 6.810197368421052e-05,
|
265 |
+
"loss": 1.4691,
|
266 |
+
"step": 3400
|
267 |
+
},
|
268 |
+
{
|
269 |
+
"epoch": 10.17,
|
270 |
+
"learning_rate": 6.760855263157895e-05,
|
271 |
+
"loss": 1.4486,
|
272 |
+
"step": 3500
|
273 |
+
},
|
274 |
+
{
|
275 |
+
"epoch": 10.17,
|
276 |
+
"eval_loss": 0.47492966055870056,
|
277 |
+
"eval_runtime": 135.1792,
|
278 |
+
"eval_samples_per_second": 34.177,
|
279 |
+
"eval_steps_per_second": 4.276,
|
280 |
+
"eval_wer": 0.46760080768524753,
|
281 |
+
"step": 3500
|
282 |
+
},
|
283 |
+
{
|
284 |
+
"epoch": 10.46,
|
285 |
+
"learning_rate": 6.711513157894736e-05,
|
286 |
+
"loss": 1.4274,
|
287 |
+
"step": 3600
|
288 |
+
},
|
289 |
+
{
|
290 |
+
"epoch": 10.75,
|
291 |
+
"learning_rate": 6.66266447368421e-05,
|
292 |
+
"loss": 1.4277,
|
293 |
+
"step": 3700
|
294 |
+
},
|
295 |
+
{
|
296 |
+
"epoch": 11.05,
|
297 |
+
"learning_rate": 6.613322368421052e-05,
|
298 |
+
"loss": 1.4445,
|
299 |
+
"step": 3800
|
300 |
+
},
|
301 |
+
{
|
302 |
+
"epoch": 11.34,
|
303 |
+
"learning_rate": 6.563980263157894e-05,
|
304 |
+
"loss": 1.4101,
|
305 |
+
"step": 3900
|
306 |
+
},
|
307 |
+
{
|
308 |
+
"epoch": 11.63,
|
309 |
+
"learning_rate": 6.514638157894736e-05,
|
310 |
+
"loss": 1.3899,
|
311 |
+
"step": 4000
|
312 |
+
},
|
313 |
+
{
|
314 |
+
"epoch": 11.63,
|
315 |
+
"eval_loss": 0.4564875364303589,
|
316 |
+
"eval_runtime": 134.5257,
|
317 |
+
"eval_samples_per_second": 34.343,
|
318 |
+
"eval_steps_per_second": 4.297,
|
319 |
+
"eval_wer": 0.4432172795692345,
|
320 |
+
"step": 4000
|
321 |
+
},
|
322 |
+
{
|
323 |
+
"epoch": 11.92,
|
324 |
+
"learning_rate": 6.465296052631578e-05,
|
325 |
+
"loss": 1.404,
|
326 |
+
"step": 4100
|
327 |
+
},
|
328 |
+
{
|
329 |
+
"epoch": 12.21,
|
330 |
+
"learning_rate": 6.415953947368421e-05,
|
331 |
+
"loss": 1.3911,
|
332 |
+
"step": 4200
|
333 |
+
},
|
334 |
+
{
|
335 |
+
"epoch": 12.5,
|
336 |
+
"learning_rate": 6.366611842105262e-05,
|
337 |
+
"loss": 1.3873,
|
338 |
+
"step": 4300
|
339 |
+
},
|
340 |
+
{
|
341 |
+
"epoch": 12.79,
|
342 |
+
"learning_rate": 6.317269736842105e-05,
|
343 |
+
"loss": 1.3712,
|
344 |
+
"step": 4400
|
345 |
+
},
|
346 |
+
{
|
347 |
+
"epoch": 13.08,
|
348 |
+
"learning_rate": 6.267927631578946e-05,
|
349 |
+
"loss": 1.3881,
|
350 |
+
"step": 4500
|
351 |
+
},
|
352 |
+
{
|
353 |
+
"epoch": 13.08,
|
354 |
+
"eval_loss": 0.43156012892723083,
|
355 |
+
"eval_runtime": 134.756,
|
356 |
+
"eval_samples_per_second": 34.284,
|
357 |
+
"eval_steps_per_second": 4.289,
|
358 |
+
"eval_wer": 0.42278039527626504,
|
359 |
+
"step": 4500
|
360 |
+
},
|
361 |
+
{
|
362 |
+
"epoch": 13.37,
|
363 |
+
"learning_rate": 6.218585526315789e-05,
|
364 |
+
"loss": 1.3538,
|
365 |
+
"step": 4600
|
366 |
+
},
|
367 |
+
{
|
368 |
+
"epoch": 13.66,
|
369 |
+
"learning_rate": 6.16924342105263e-05,
|
370 |
+
"loss": 1.355,
|
371 |
+
"step": 4700
|
372 |
+
},
|
373 |
+
{
|
374 |
+
"epoch": 13.95,
|
375 |
+
"learning_rate": 6.119901315789473e-05,
|
376 |
+
"loss": 1.341,
|
377 |
+
"step": 4800
|
378 |
+
},
|
379 |
+
{
|
380 |
+
"epoch": 14.24,
|
381 |
+
"learning_rate": 6.070559210526316e-05,
|
382 |
+
"loss": 1.3474,
|
383 |
+
"step": 4900
|
384 |
+
},
|
385 |
+
{
|
386 |
+
"epoch": 14.53,
|
387 |
+
"learning_rate": 6.021217105263157e-05,
|
388 |
+
"loss": 1.3572,
|
389 |
+
"step": 5000
|
390 |
+
},
|
391 |
+
{
|
392 |
+
"epoch": 14.53,
|
393 |
+
"eval_loss": 0.4194825291633606,
|
394 |
+
"eval_runtime": 134.9338,
|
395 |
+
"eval_samples_per_second": 34.239,
|
396 |
+
"eval_steps_per_second": 4.284,
|
397 |
+
"eval_wer": 0.3834363335984825,
|
398 |
+
"step": 5000
|
399 |
+
},
|
400 |
+
{
|
401 |
+
"epoch": 14.82,
|
402 |
+
"learning_rate": 5.971875e-05,
|
403 |
+
"loss": 1.3408,
|
404 |
+
"step": 5100
|
405 |
+
},
|
406 |
+
{
|
407 |
+
"epoch": 15.12,
|
408 |
+
"learning_rate": 5.922532894736842e-05,
|
409 |
+
"loss": 1.3357,
|
410 |
+
"step": 5200
|
411 |
+
},
|
412 |
+
{
|
413 |
+
"epoch": 15.41,
|
414 |
+
"learning_rate": 5.873190789473683e-05,
|
415 |
+
"loss": 1.3288,
|
416 |
+
"step": 5300
|
417 |
+
},
|
418 |
+
{
|
419 |
+
"epoch": 15.7,
|
420 |
+
"learning_rate": 5.823848684210526e-05,
|
421 |
+
"loss": 1.3258,
|
422 |
+
"step": 5400
|
423 |
+
},
|
424 |
+
{
|
425 |
+
"epoch": 15.99,
|
426 |
+
"learning_rate": 5.7745065789473685e-05,
|
427 |
+
"loss": 1.3261,
|
428 |
+
"step": 5500
|
429 |
+
},
|
430 |
+
{
|
431 |
+
"epoch": 15.99,
|
432 |
+
"eval_loss": 0.3974045217037201,
|
433 |
+
"eval_runtime": 134.8844,
|
434 |
+
"eval_samples_per_second": 34.252,
|
435 |
+
"eval_steps_per_second": 4.285,
|
436 |
+
"eval_wer": 0.3606742948051153,
|
437 |
+
"step": 5500
|
438 |
+
},
|
439 |
+
{
|
440 |
+
"epoch": 16.28,
|
441 |
+
"learning_rate": 5.72516447368421e-05,
|
442 |
+
"loss": 1.3273,
|
443 |
+
"step": 5600
|
444 |
+
},
|
445 |
+
{
|
446 |
+
"epoch": 16.57,
|
447 |
+
"learning_rate": 5.675822368421052e-05,
|
448 |
+
"loss": 1.3131,
|
449 |
+
"step": 5700
|
450 |
+
},
|
451 |
+
{
|
452 |
+
"epoch": 16.86,
|
453 |
+
"learning_rate": 5.6264802631578946e-05,
|
454 |
+
"loss": 1.2739,
|
455 |
+
"step": 5800
|
456 |
+
},
|
457 |
+
{
|
458 |
+
"epoch": 17.15,
|
459 |
+
"learning_rate": 5.577138157894736e-05,
|
460 |
+
"loss": 1.3141,
|
461 |
+
"step": 5900
|
462 |
+
},
|
463 |
+
{
|
464 |
+
"epoch": 17.44,
|
465 |
+
"learning_rate": 5.5277960526315786e-05,
|
466 |
+
"loss": 1.2809,
|
467 |
+
"step": 6000
|
468 |
+
},
|
469 |
+
{
|
470 |
+
"epoch": 17.44,
|
471 |
+
"eval_loss": 0.3844749331474304,
|
472 |
+
"eval_runtime": 136.3592,
|
473 |
+
"eval_samples_per_second": 33.881,
|
474 |
+
"eval_steps_per_second": 4.239,
|
475 |
+
"eval_wer": 0.3466621795264027,
|
476 |
+
"step": 6000
|
477 |
+
},
|
478 |
+
{
|
479 |
+
"epoch": 17.73,
|
480 |
+
"learning_rate": 5.4784539473684206e-05,
|
481 |
+
"loss": 1.2847,
|
482 |
+
"step": 6100
|
483 |
+
},
|
484 |
+
{
|
485 |
+
"epoch": 18.02,
|
486 |
+
"learning_rate": 5.4291118421052626e-05,
|
487 |
+
"loss": 1.303,
|
488 |
+
"step": 6200
|
489 |
+
},
|
490 |
+
{
|
491 |
+
"epoch": 18.31,
|
492 |
+
"learning_rate": 5.3797697368421046e-05,
|
493 |
+
"loss": 1.2733,
|
494 |
+
"step": 6300
|
495 |
+
},
|
496 |
+
{
|
497 |
+
"epoch": 18.6,
|
498 |
+
"learning_rate": 5.330427631578947e-05,
|
499 |
+
"loss": 1.2707,
|
500 |
+
"step": 6400
|
501 |
+
},
|
502 |
+
{
|
503 |
+
"epoch": 18.89,
|
504 |
+
"learning_rate": 5.2810855263157886e-05,
|
505 |
+
"loss": 1.2713,
|
506 |
+
"step": 6500
|
507 |
+
},
|
508 |
+
{
|
509 |
+
"epoch": 18.89,
|
510 |
+
"eval_loss": 0.3831607401371002,
|
511 |
+
"eval_runtime": 138.2456,
|
512 |
+
"eval_samples_per_second": 33.419,
|
513 |
+
"eval_steps_per_second": 4.181,
|
514 |
+
"eval_wer": 0.3449795019274307,
|
515 |
+
"step": 6500
|
516 |
+
},
|
517 |
+
{
|
518 |
+
"epoch": 19.19,
|
519 |
+
"learning_rate": 5.231743421052631e-05,
|
520 |
+
"loss": 1.2562,
|
521 |
+
"step": 6600
|
522 |
+
},
|
523 |
+
{
|
524 |
+
"epoch": 19.48,
|
525 |
+
"learning_rate": 5.182894736842105e-05,
|
526 |
+
"loss": 1.2721,
|
527 |
+
"step": 6700
|
528 |
+
},
|
529 |
+
{
|
530 |
+
"epoch": 19.77,
|
531 |
+
"learning_rate": 5.133552631578947e-05,
|
532 |
+
"loss": 1.2536,
|
533 |
+
"step": 6800
|
534 |
+
},
|
535 |
+
{
|
536 |
+
"epoch": 20.06,
|
537 |
+
"learning_rate": 5.084210526315789e-05,
|
538 |
+
"loss": 1.2503,
|
539 |
+
"step": 6900
|
540 |
+
},
|
541 |
+
{
|
542 |
+
"epoch": 20.35,
|
543 |
+
"learning_rate": 5.0353618421052625e-05,
|
544 |
+
"loss": 1.257,
|
545 |
+
"step": 7000
|
546 |
+
},
|
547 |
+
{
|
548 |
+
"epoch": 20.35,
|
549 |
+
"eval_loss": 0.3778809607028961,
|
550 |
+
"eval_runtime": 138.2403,
|
551 |
+
"eval_samples_per_second": 33.42,
|
552 |
+
"eval_steps_per_second": 4.181,
|
553 |
+
"eval_wer": 0.33726977911032247,
|
554 |
+
"step": 7000
|
555 |
+
},
|
556 |
+
{
|
557 |
+
"epoch": 20.64,
|
558 |
+
"learning_rate": 4.9860197368421045e-05,
|
559 |
+
"loss": 1.2557,
|
560 |
+
"step": 7100
|
561 |
+
},
|
562 |
+
{
|
563 |
+
"epoch": 20.93,
|
564 |
+
"learning_rate": 4.936677631578947e-05,
|
565 |
+
"loss": 1.2397,
|
566 |
+
"step": 7200
|
567 |
+
},
|
568 |
+
{
|
569 |
+
"epoch": 21.22,
|
570 |
+
"learning_rate": 4.8873355263157886e-05,
|
571 |
+
"loss": 1.2527,
|
572 |
+
"step": 7300
|
573 |
+
},
|
574 |
+
{
|
575 |
+
"epoch": 21.51,
|
576 |
+
"learning_rate": 4.837993421052631e-05,
|
577 |
+
"loss": 1.2273,
|
578 |
+
"step": 7400
|
579 |
+
},
|
580 |
+
{
|
581 |
+
"epoch": 21.8,
|
582 |
+
"learning_rate": 4.788651315789473e-05,
|
583 |
+
"loss": 1.2298,
|
584 |
+
"step": 7500
|
585 |
+
},
|
586 |
+
{
|
587 |
+
"epoch": 21.8,
|
588 |
+
"eval_loss": 0.3743567168712616,
|
589 |
+
"eval_runtime": 136.1375,
|
590 |
+
"eval_samples_per_second": 33.936,
|
591 |
+
"eval_steps_per_second": 4.246,
|
592 |
+
"eval_wer": 0.33913602153827327,
|
593 |
+
"step": 7500
|
594 |
+
},
|
595 |
+
{
|
596 |
+
"epoch": 22.09,
|
597 |
+
"learning_rate": 4.739309210526315e-05,
|
598 |
+
"loss": 1.2444,
|
599 |
+
"step": 7600
|
600 |
+
},
|
601 |
+
{
|
602 |
+
"epoch": 22.38,
|
603 |
+
"learning_rate": 4.689967105263157e-05,
|
604 |
+
"loss": 1.2153,
|
605 |
+
"step": 7700
|
606 |
+
},
|
607 |
+
{
|
608 |
+
"epoch": 22.67,
|
609 |
+
"learning_rate": 4.640625e-05,
|
610 |
+
"loss": 1.2327,
|
611 |
+
"step": 7800
|
612 |
+
},
|
613 |
+
{
|
614 |
+
"epoch": 22.96,
|
615 |
+
"learning_rate": 4.591282894736841e-05,
|
616 |
+
"loss": 1.2404,
|
617 |
+
"step": 7900
|
618 |
+
},
|
619 |
+
{
|
620 |
+
"epoch": 23.26,
|
621 |
+
"learning_rate": 4.541940789473684e-05,
|
622 |
+
"loss": 1.2173,
|
623 |
+
"step": 8000
|
624 |
+
},
|
625 |
+
{
|
626 |
+
"epoch": 23.26,
|
627 |
+
"eval_loss": 0.3745496869087219,
|
628 |
+
"eval_runtime": 139.3565,
|
629 |
+
"eval_samples_per_second": 33.152,
|
630 |
+
"eval_steps_per_second": 4.148,
|
631 |
+
"eval_wer": 0.32619470109527016,
|
632 |
+
"step": 8000
|
633 |
+
},
|
634 |
+
{
|
635 |
+
"epoch": 23.55,
|
636 |
+
"learning_rate": 4.492598684210526e-05,
|
637 |
+
"loss": 1.222,
|
638 |
+
"step": 8100
|
639 |
+
},
|
640 |
+
{
|
641 |
+
"epoch": 23.84,
|
642 |
+
"learning_rate": 4.443256578947369e-05,
|
643 |
+
"loss": 1.1868,
|
644 |
+
"step": 8200
|
645 |
+
},
|
646 |
+
{
|
647 |
+
"epoch": 24.13,
|
648 |
+
"learning_rate": 4.39391447368421e-05,
|
649 |
+
"loss": 1.2272,
|
650 |
+
"step": 8300
|
651 |
+
},
|
652 |
+
{
|
653 |
+
"epoch": 24.42,
|
654 |
+
"learning_rate": 4.344572368421052e-05,
|
655 |
+
"loss": 1.2212,
|
656 |
+
"step": 8400
|
657 |
+
},
|
658 |
+
{
|
659 |
+
"epoch": 24.71,
|
660 |
+
"learning_rate": 4.295230263157895e-05,
|
661 |
+
"loss": 1.1966,
|
662 |
+
"step": 8500
|
663 |
+
},
|
664 |
+
{
|
665 |
+
"epoch": 24.71,
|
666 |
+
"eval_loss": 0.36800575256347656,
|
667 |
+
"eval_runtime": 137.1685,
|
668 |
+
"eval_samples_per_second": 33.681,
|
669 |
+
"eval_steps_per_second": 4.214,
|
670 |
+
"eval_wer": 0.32411429970017747,
|
671 |
+
"step": 8500
|
672 |
+
},
|
673 |
+
{
|
674 |
+
"epoch": 25.0,
|
675 |
+
"learning_rate": 4.245888157894736e-05,
|
676 |
+
"loss": 1.2128,
|
677 |
+
"step": 8600
|
678 |
+
},
|
679 |
+
{
|
680 |
+
"epoch": 25.29,
|
681 |
+
"learning_rate": 4.196546052631579e-05,
|
682 |
+
"loss": 1.2118,
|
683 |
+
"step": 8700
|
684 |
+
},
|
685 |
+
{
|
686 |
+
"epoch": 25.58,
|
687 |
+
"learning_rate": 4.147203947368421e-05,
|
688 |
+
"loss": 1.2009,
|
689 |
+
"step": 8800
|
690 |
+
},
|
691 |
+
{
|
692 |
+
"epoch": 25.87,
|
693 |
+
"learning_rate": 4.097861842105263e-05,
|
694 |
+
"loss": 1.1684,
|
695 |
+
"step": 8900
|
696 |
+
},
|
697 |
+
{
|
698 |
+
"epoch": 26.16,
|
699 |
+
"learning_rate": 4.048519736842105e-05,
|
700 |
+
"loss": 1.1925,
|
701 |
+
"step": 9000
|
702 |
+
},
|
703 |
+
{
|
704 |
+
"epoch": 26.16,
|
705 |
+
"eval_loss": 0.3604615330696106,
|
706 |
+
"eval_runtime": 153.4534,
|
707 |
+
"eval_samples_per_second": 30.107,
|
708 |
+
"eval_steps_per_second": 3.767,
|
709 |
+
"eval_wer": 0.31713883619898425,
|
710 |
+
"step": 9000
|
711 |
+
},
|
712 |
+
{
|
713 |
+
"epoch": 26.45,
|
714 |
+
"learning_rate": 3.9991776315789475e-05,
|
715 |
+
"loss": 1.1759,
|
716 |
+
"step": 9100
|
717 |
+
},
|
718 |
+
{
|
719 |
+
"epoch": 26.74,
|
720 |
+
"learning_rate": 3.949835526315789e-05,
|
721 |
+
"loss": 1.1754,
|
722 |
+
"step": 9200
|
723 |
+
},
|
724 |
+
{
|
725 |
+
"epoch": 27.03,
|
726 |
+
"learning_rate": 3.9004934210526315e-05,
|
727 |
+
"loss": 1.198,
|
728 |
+
"step": 9300
|
729 |
+
},
|
730 |
+
{
|
731 |
+
"epoch": 27.32,
|
732 |
+
"learning_rate": 3.8511513157894735e-05,
|
733 |
+
"loss": 1.1602,
|
734 |
+
"step": 9400
|
735 |
+
},
|
736 |
+
{
|
737 |
+
"epoch": 27.61,
|
738 |
+
"learning_rate": 3.801809210526315e-05,
|
739 |
+
"loss": 1.1692,
|
740 |
+
"step": 9500
|
741 |
+
},
|
742 |
+
{
|
743 |
+
"epoch": 27.61,
|
744 |
+
"eval_loss": 0.3512294292449951,
|
745 |
+
"eval_runtime": 166.9619,
|
746 |
+
"eval_samples_per_second": 27.671,
|
747 |
+
"eval_steps_per_second": 3.462,
|
748 |
+
"eval_wer": 0.31472189928409716,
|
749 |
+
"step": 9500
|
750 |
+
},
|
751 |
+
{
|
752 |
+
"epoch": 27.91,
|
753 |
+
"learning_rate": 3.7524671052631575e-05,
|
754 |
+
"loss": 1.1495,
|
755 |
+
"step": 9600
|
756 |
+
},
|
757 |
+
{
|
758 |
+
"epoch": 28.2,
|
759 |
+
"learning_rate": 3.7031249999999995e-05,
|
760 |
+
"loss": 1.1722,
|
761 |
+
"step": 9700
|
762 |
+
},
|
763 |
+
{
|
764 |
+
"epoch": 28.49,
|
765 |
+
"learning_rate": 3.6537828947368416e-05,
|
766 |
+
"loss": 1.1443,
|
767 |
+
"step": 9800
|
768 |
+
},
|
769 |
+
{
|
770 |
+
"epoch": 28.78,
|
771 |
+
"learning_rate": 3.6044407894736836e-05,
|
772 |
+
"loss": 1.1535,
|
773 |
+
"step": 9900
|
774 |
+
},
|
775 |
+
{
|
776 |
+
"epoch": 29.07,
|
777 |
+
"learning_rate": 3.555098684210526e-05,
|
778 |
+
"loss": 1.1704,
|
779 |
+
"step": 10000
|
780 |
+
},
|
781 |
+
{
|
782 |
+
"epoch": 29.07,
|
783 |
+
"eval_loss": 0.3532196581363678,
|
784 |
+
"eval_runtime": 134.5452,
|
785 |
+
"eval_samples_per_second": 34.338,
|
786 |
+
"eval_steps_per_second": 4.296,
|
787 |
+
"eval_wer": 0.30976564890167047,
|
788 |
+
"step": 10000
|
789 |
+
},
|
790 |
+
{
|
791 |
+
"epoch": 29.36,
|
792 |
+
"learning_rate": 3.505756578947368e-05,
|
793 |
+
"loss": 1.1519,
|
794 |
+
"step": 10100
|
795 |
+
},
|
796 |
+
{
|
797 |
+
"epoch": 29.65,
|
798 |
+
"learning_rate": 3.45641447368421e-05,
|
799 |
+
"loss": 1.1507,
|
800 |
+
"step": 10200
|
801 |
+
},
|
802 |
+
{
|
803 |
+
"epoch": 29.94,
|
804 |
+
"learning_rate": 3.407072368421052e-05,
|
805 |
+
"loss": 1.1517,
|
806 |
+
"step": 10300
|
807 |
+
},
|
808 |
+
{
|
809 |
+
"epoch": 30.23,
|
810 |
+
"learning_rate": 3.357730263157894e-05,
|
811 |
+
"loss": 1.1447,
|
812 |
+
"step": 10400
|
813 |
+
},
|
814 |
+
{
|
815 |
+
"epoch": 30.52,
|
816 |
+
"learning_rate": 3.308388157894737e-05,
|
817 |
+
"loss": 1.1595,
|
818 |
+
"step": 10500
|
819 |
+
},
|
820 |
+
{
|
821 |
+
"epoch": 30.52,
|
822 |
+
"eval_loss": 0.34246256947517395,
|
823 |
+
"eval_runtime": 159.7393,
|
824 |
+
"eval_samples_per_second": 28.922,
|
825 |
+
"eval_steps_per_second": 3.618,
|
826 |
+
"eval_wer": 0.3038915743743499,
|
827 |
+
"step": 10500
|
828 |
+
},
|
829 |
+
{
|
830 |
+
"epoch": 30.81,
|
831 |
+
"learning_rate": 3.259046052631579e-05,
|
832 |
+
"loss": 1.1451,
|
833 |
+
"step": 10600
|
834 |
+
},
|
835 |
+
{
|
836 |
+
"epoch": 31.1,
|
837 |
+
"learning_rate": 3.2097039473684203e-05,
|
838 |
+
"loss": 1.1304,
|
839 |
+
"step": 10700
|
840 |
+
},
|
841 |
+
{
|
842 |
+
"epoch": 31.39,
|
843 |
+
"learning_rate": 3.160361842105263e-05,
|
844 |
+
"loss": 1.1316,
|
845 |
+
"step": 10800
|
846 |
+
},
|
847 |
+
{
|
848 |
+
"epoch": 31.68,
|
849 |
+
"learning_rate": 3.111019736842105e-05,
|
850 |
+
"loss": 1.1301,
|
851 |
+
"step": 10900
|
852 |
+
},
|
853 |
+
{
|
854 |
+
"epoch": 31.97,
|
855 |
+
"learning_rate": 3.061677631578947e-05,
|
856 |
+
"loss": 1.1433,
|
857 |
+
"step": 11000
|
858 |
+
},
|
859 |
+
{
|
860 |
+
"epoch": 31.97,
|
861 |
+
"eval_loss": 0.35683709383010864,
|
862 |
+
"eval_runtime": 158.7503,
|
863 |
+
"eval_samples_per_second": 29.102,
|
864 |
+
"eval_steps_per_second": 3.641,
|
865 |
+
"eval_wer": 0.30257602643333537,
|
866 |
+
"step": 11000
|
867 |
+
},
|
868 |
+
{
|
869 |
+
"epoch": 32.27,
|
870 |
+
"learning_rate": 3.0123355263157894e-05,
|
871 |
+
"loss": 1.1337,
|
872 |
+
"step": 11100
|
873 |
+
},
|
874 |
+
{
|
875 |
+
"epoch": 32.56,
|
876 |
+
"learning_rate": 2.9629934210526314e-05,
|
877 |
+
"loss": 1.1227,
|
878 |
+
"step": 11200
|
879 |
+
},
|
880 |
+
{
|
881 |
+
"epoch": 32.85,
|
882 |
+
"learning_rate": 2.913651315789473e-05,
|
883 |
+
"loss": 1.1281,
|
884 |
+
"step": 11300
|
885 |
+
},
|
886 |
+
{
|
887 |
+
"epoch": 33.14,
|
888 |
+
"learning_rate": 2.8643092105263154e-05,
|
889 |
+
"loss": 1.1203,
|
890 |
+
"step": 11400
|
891 |
+
},
|
892 |
+
{
|
893 |
+
"epoch": 33.43,
|
894 |
+
"learning_rate": 2.8149671052631574e-05,
|
895 |
+
"loss": 1.1295,
|
896 |
+
"step": 11500
|
897 |
+
},
|
898 |
+
{
|
899 |
+
"epoch": 33.43,
|
900 |
+
"eval_loss": 0.34605443477630615,
|
901 |
+
"eval_runtime": 160.1749,
|
902 |
+
"eval_samples_per_second": 28.843,
|
903 |
+
"eval_steps_per_second": 3.609,
|
904 |
+
"eval_wer": 0.2992106712353913,
|
905 |
+
"step": 11500
|
906 |
+
},
|
907 |
+
{
|
908 |
+
"epoch": 33.72,
|
909 |
+
"learning_rate": 2.7656249999999998e-05,
|
910 |
+
"loss": 1.0915,
|
911 |
+
"step": 11600
|
912 |
+
},
|
913 |
+
{
|
914 |
+
"epoch": 34.01,
|
915 |
+
"learning_rate": 2.7162828947368418e-05,
|
916 |
+
"loss": 1.1312,
|
917 |
+
"step": 11700
|
918 |
+
},
|
919 |
+
{
|
920 |
+
"epoch": 34.3,
|
921 |
+
"learning_rate": 2.6669407894736838e-05,
|
922 |
+
"loss": 1.1146,
|
923 |
+
"step": 11800
|
924 |
+
},
|
925 |
+
{
|
926 |
+
"epoch": 34.59,
|
927 |
+
"learning_rate": 2.6175986842105262e-05,
|
928 |
+
"loss": 1.1184,
|
929 |
+
"step": 11900
|
930 |
+
},
|
931 |
+
{
|
932 |
+
"epoch": 34.88,
|
933 |
+
"learning_rate": 2.56875e-05,
|
934 |
+
"loss": 1.1131,
|
935 |
+
"step": 12000
|
936 |
+
},
|
937 |
+
{
|
938 |
+
"epoch": 34.88,
|
939 |
+
"eval_loss": 0.3348712623119354,
|
940 |
+
"eval_runtime": 157.6811,
|
941 |
+
"eval_samples_per_second": 29.3,
|
942 |
+
"eval_steps_per_second": 3.666,
|
943 |
+
"eval_wer": 0.29422382671480146,
|
944 |
+
"step": 12000
|
945 |
+
},
|
946 |
+
{
|
947 |
+
"epoch": 35.17,
|
948 |
+
"learning_rate": 2.5194078947368418e-05,
|
949 |
+
"loss": 1.1025,
|
950 |
+
"step": 12100
|
951 |
+
},
|
952 |
+
{
|
953 |
+
"epoch": 35.46,
|
954 |
+
"learning_rate": 2.4700657894736838e-05,
|
955 |
+
"loss": 1.1069,
|
956 |
+
"step": 12200
|
957 |
+
},
|
958 |
+
{
|
959 |
+
"epoch": 35.75,
|
960 |
+
"learning_rate": 2.420723684210526e-05,
|
961 |
+
"loss": 1.076,
|
962 |
+
"step": 12300
|
963 |
+
},
|
964 |
+
{
|
965 |
+
"epoch": 36.05,
|
966 |
+
"learning_rate": 2.371381578947368e-05,
|
967 |
+
"loss": 1.111,
|
968 |
+
"step": 12400
|
969 |
+
},
|
970 |
+
{
|
971 |
+
"epoch": 36.34,
|
972 |
+
"learning_rate": 2.32203947368421e-05,
|
973 |
+
"loss": 1.1015,
|
974 |
+
"step": 12500
|
975 |
+
},
|
976 |
+
{
|
977 |
+
"epoch": 36.34,
|
978 |
+
"eval_loss": 0.3378337025642395,
|
979 |
+
"eval_runtime": 160.9202,
|
980 |
+
"eval_samples_per_second": 28.71,
|
981 |
+
"eval_steps_per_second": 3.592,
|
982 |
+
"eval_wer": 0.29612066328091535,
|
983 |
+
"step": 12500
|
984 |
+
},
|
985 |
+
{
|
986 |
+
"epoch": 36.63,
|
987 |
+
"learning_rate": 2.2726973684210525e-05,
|
988 |
+
"loss": 1.1047,
|
989 |
+
"step": 12600
|
990 |
+
},
|
991 |
+
{
|
992 |
+
"epoch": 36.92,
|
993 |
+
"learning_rate": 2.2233552631578945e-05,
|
994 |
+
"loss": 1.102,
|
995 |
+
"step": 12700
|
996 |
+
},
|
997 |
+
{
|
998 |
+
"epoch": 37.21,
|
999 |
+
"learning_rate": 2.1740131578947365e-05,
|
1000 |
+
"loss": 1.095,
|
1001 |
+
"step": 12800
|
1002 |
+
},
|
1003 |
+
{
|
1004 |
+
"epoch": 37.5,
|
1005 |
+
"learning_rate": 2.124671052631579e-05,
|
1006 |
+
"loss": 1.0709,
|
1007 |
+
"step": 12900
|
1008 |
+
},
|
1009 |
+
{
|
1010 |
+
"epoch": 37.79,
|
1011 |
+
"learning_rate": 2.075328947368421e-05,
|
1012 |
+
"loss": 1.0835,
|
1013 |
+
"step": 13000
|
1014 |
+
},
|
1015 |
+
{
|
1016 |
+
"epoch": 37.79,
|
1017 |
+
"eval_loss": 0.3281959593296051,
|
1018 |
+
"eval_runtime": 159.3279,
|
1019 |
+
"eval_samples_per_second": 28.997,
|
1020 |
+
"eval_steps_per_second": 3.628,
|
1021 |
+
"eval_wer": 0.2865141038976932,
|
1022 |
+
"step": 13000
|
1023 |
+
},
|
1024 |
+
{
|
1025 |
+
"epoch": 38.08,
|
1026 |
+
"learning_rate": 2.0259868421052632e-05,
|
1027 |
+
"loss": 1.0846,
|
1028 |
+
"step": 13100
|
1029 |
+
},
|
1030 |
+
{
|
1031 |
+
"epoch": 38.37,
|
1032 |
+
"learning_rate": 1.9766447368421053e-05,
|
1033 |
+
"loss": 1.0933,
|
1034 |
+
"step": 13200
|
1035 |
+
},
|
1036 |
+
{
|
1037 |
+
"epoch": 38.66,
|
1038 |
+
"learning_rate": 1.927302631578947e-05,
|
1039 |
+
"loss": 1.0661,
|
1040 |
+
"step": 13300
|
1041 |
+
},
|
1042 |
+
{
|
1043 |
+
"epoch": 38.95,
|
1044 |
+
"learning_rate": 1.8779605263157893e-05,
|
1045 |
+
"loss": 1.0902,
|
1046 |
+
"step": 13400
|
1047 |
+
},
|
1048 |
+
{
|
1049 |
+
"epoch": 39.24,
|
1050 |
+
"learning_rate": 1.8286184210526313e-05,
|
1051 |
+
"loss": 1.083,
|
1052 |
+
"step": 13500
|
1053 |
+
},
|
1054 |
+
{
|
1055 |
+
"epoch": 39.24,
|
1056 |
+
"eval_loss": 0.3181643486022949,
|
1057 |
+
"eval_runtime": 158.3527,
|
1058 |
+
"eval_samples_per_second": 29.175,
|
1059 |
+
"eval_steps_per_second": 3.65,
|
1060 |
+
"eval_wer": 0.2825674600746497,
|
1061 |
+
"step": 13500
|
1062 |
+
},
|
1063 |
+
{
|
1064 |
+
"epoch": 39.53,
|
1065 |
+
"learning_rate": 1.7792763157894736e-05,
|
1066 |
+
"loss": 1.0548,
|
1067 |
+
"step": 13600
|
1068 |
+
},
|
1069 |
+
{
|
1070 |
+
"epoch": 39.82,
|
1071 |
+
"learning_rate": 1.7299342105263156e-05,
|
1072 |
+
"loss": 1.0735,
|
1073 |
+
"step": 13700
|
1074 |
+
},
|
1075 |
+
{
|
1076 |
+
"epoch": 40.12,
|
1077 |
+
"learning_rate": 1.6805921052631577e-05,
|
1078 |
+
"loss": 1.085,
|
1079 |
+
"step": 13800
|
1080 |
+
},
|
1081 |
+
{
|
1082 |
+
"epoch": 40.41,
|
1083 |
+
"learning_rate": 1.63125e-05,
|
1084 |
+
"loss": 1.0531,
|
1085 |
+
"step": 13900
|
1086 |
+
},
|
1087 |
+
{
|
1088 |
+
"epoch": 40.7,
|
1089 |
+
"learning_rate": 1.581907894736842e-05,
|
1090 |
+
"loss": 1.0819,
|
1091 |
+
"step": 14000
|
1092 |
+
},
|
1093 |
+
{
|
1094 |
+
"epoch": 40.7,
|
1095 |
+
"eval_loss": 0.32643991708755493,
|
1096 |
+
"eval_runtime": 159.1099,
|
1097 |
+
"eval_samples_per_second": 29.037,
|
1098 |
+
"eval_steps_per_second": 3.633,
|
1099 |
+
"eval_wer": 0.28504558526586304,
|
1100 |
+
"step": 14000
|
1101 |
+
},
|
1102 |
+
{
|
1103 |
+
"epoch": 40.99,
|
1104 |
+
"learning_rate": 1.532565789473684e-05,
|
1105 |
+
"loss": 1.0679,
|
1106 |
+
"step": 14100
|
1107 |
+
},
|
1108 |
+
{
|
1109 |
+
"epoch": 41.28,
|
1110 |
+
"learning_rate": 1.4832236842105262e-05,
|
1111 |
+
"loss": 1.0773,
|
1112 |
+
"step": 14200
|
1113 |
+
},
|
1114 |
+
{
|
1115 |
+
"epoch": 41.57,
|
1116 |
+
"learning_rate": 1.4338815789473682e-05,
|
1117 |
+
"loss": 1.0587,
|
1118 |
+
"step": 14300
|
1119 |
+
},
|
1120 |
+
{
|
1121 |
+
"epoch": 41.86,
|
1122 |
+
"learning_rate": 1.3845394736842104e-05,
|
1123 |
+
"loss": 1.0287,
|
1124 |
+
"step": 14400
|
1125 |
+
},
|
1126 |
+
{
|
1127 |
+
"epoch": 42.15,
|
1128 |
+
"learning_rate": 1.3351973684210524e-05,
|
1129 |
+
"loss": 1.072,
|
1130 |
+
"step": 14500
|
1131 |
+
},
|
1132 |
+
{
|
1133 |
+
"epoch": 42.15,
|
1134 |
+
"eval_loss": 0.32792535424232483,
|
1135 |
+
"eval_runtime": 155.2521,
|
1136 |
+
"eval_samples_per_second": 29.758,
|
1137 |
+
"eval_steps_per_second": 3.723,
|
1138 |
+
"eval_wer": 0.2817108242060821,
|
1139 |
+
"step": 14500
|
1140 |
+
},
|
1141 |
+
{
|
1142 |
+
"epoch": 42.44,
|
1143 |
+
"learning_rate": 1.2863486842105262e-05,
|
1144 |
+
"loss": 1.0508,
|
1145 |
+
"step": 14600
|
1146 |
+
},
|
1147 |
+
{
|
1148 |
+
"epoch": 42.73,
|
1149 |
+
"learning_rate": 1.2370065789473684e-05,
|
1150 |
+
"loss": 1.0408,
|
1151 |
+
"step": 14700
|
1152 |
+
},
|
1153 |
+
{
|
1154 |
+
"epoch": 43.02,
|
1155 |
+
"learning_rate": 1.1876644736842105e-05,
|
1156 |
+
"loss": 1.0725,
|
1157 |
+
"step": 14800
|
1158 |
+
},
|
1159 |
+
{
|
1160 |
+
"epoch": 43.31,
|
1161 |
+
"learning_rate": 1.1383223684210525e-05,
|
1162 |
+
"loss": 1.0381,
|
1163 |
+
"step": 14900
|
1164 |
+
},
|
1165 |
+
{
|
1166 |
+
"epoch": 43.6,
|
1167 |
+
"learning_rate": 1.0889802631578946e-05,
|
1168 |
+
"loss": 1.0456,
|
1169 |
+
"step": 15000
|
1170 |
+
},
|
1171 |
+
{
|
1172 |
+
"epoch": 43.6,
|
1173 |
+
"eval_loss": 0.323445200920105,
|
1174 |
+
"eval_runtime": 137.3413,
|
1175 |
+
"eval_samples_per_second": 33.639,
|
1176 |
+
"eval_steps_per_second": 4.208,
|
1177 |
+
"eval_wer": 0.2792632931530319,
|
1178 |
+
"step": 15000
|
1179 |
+
},
|
1180 |
+
{
|
1181 |
+
"epoch": 43.89,
|
1182 |
+
"learning_rate": 1.0396381578947367e-05,
|
1183 |
+
"loss": 1.0655,
|
1184 |
+
"step": 15100
|
1185 |
+
},
|
1186 |
+
{
|
1187 |
+
"epoch": 44.19,
|
1188 |
+
"learning_rate": 9.90296052631579e-06,
|
1189 |
+
"loss": 1.0509,
|
1190 |
+
"step": 15200
|
1191 |
+
},
|
1192 |
+
{
|
1193 |
+
"epoch": 44.48,
|
1194 |
+
"learning_rate": 9.40953947368421e-06,
|
1195 |
+
"loss": 1.0357,
|
1196 |
+
"step": 15300
|
1197 |
+
},
|
1198 |
+
{
|
1199 |
+
"epoch": 44.77,
|
1200 |
+
"learning_rate": 8.916118421052631e-06,
|
1201 |
+
"loss": 1.0467,
|
1202 |
+
"step": 15400
|
1203 |
+
},
|
1204 |
+
{
|
1205 |
+
"epoch": 45.06,
|
1206 |
+
"learning_rate": 8.422697368421051e-06,
|
1207 |
+
"loss": 1.0581,
|
1208 |
+
"step": 15500
|
1209 |
+
},
|
1210 |
+
{
|
1211 |
+
"epoch": 45.06,
|
1212 |
+
"eval_loss": 0.32200726866722107,
|
1213 |
+
"eval_runtime": 138.9222,
|
1214 |
+
"eval_samples_per_second": 33.256,
|
1215 |
+
"eval_steps_per_second": 4.161,
|
1216 |
+
"eval_wer": 0.27791715107385423,
|
1217 |
+
"step": 15500
|
1218 |
+
},
|
1219 |
+
{
|
1220 |
+
"epoch": 45.35,
|
1221 |
+
"learning_rate": 7.929276315789473e-06,
|
1222 |
+
"loss": 1.0512,
|
1223 |
+
"step": 15600
|
1224 |
+
},
|
1225 |
+
{
|
1226 |
+
"epoch": 45.64,
|
1227 |
+
"learning_rate": 7.435855263157894e-06,
|
1228 |
+
"loss": 1.0397,
|
1229 |
+
"step": 15700
|
1230 |
+
},
|
1231 |
+
{
|
1232 |
+
"epoch": 45.93,
|
1233 |
+
"learning_rate": 6.9473684210526315e-06,
|
1234 |
+
"loss": 1.0455,
|
1235 |
+
"step": 15800
|
1236 |
+
},
|
1237 |
+
{
|
1238 |
+
"epoch": 46.22,
|
1239 |
+
"learning_rate": 6.453947368421052e-06,
|
1240 |
+
"loss": 1.0347,
|
1241 |
+
"step": 15900
|
1242 |
+
},
|
1243 |
+
{
|
1244 |
+
"epoch": 46.51,
|
1245 |
+
"learning_rate": 5.9605263157894735e-06,
|
1246 |
+
"loss": 1.0406,
|
1247 |
+
"step": 16000
|
1248 |
+
},
|
1249 |
+
{
|
1250 |
+
"epoch": 46.51,
|
1251 |
+
"eval_loss": 0.32076749205589294,
|
1252 |
+
"eval_runtime": 134.5334,
|
1253 |
+
"eval_samples_per_second": 34.341,
|
1254 |
+
"eval_steps_per_second": 4.296,
|
1255 |
+
"eval_wer": 0.2762344734748822,
|
1256 |
+
"step": 16000
|
1257 |
+
},
|
1258 |
+
{
|
1259 |
+
"epoch": 46.8,
|
1260 |
+
"learning_rate": 5.467105263157894e-06,
|
1261 |
+
"loss": 1.0528,
|
1262 |
+
"step": 16100
|
1263 |
+
},
|
1264 |
+
{
|
1265 |
+
"epoch": 47.09,
|
1266 |
+
"learning_rate": 4.973684210526316e-06,
|
1267 |
+
"loss": 1.045,
|
1268 |
+
"step": 16200
|
1269 |
+
},
|
1270 |
+
{
|
1271 |
+
"epoch": 47.38,
|
1272 |
+
"learning_rate": 4.480263157894736e-06,
|
1273 |
+
"loss": 1.0129,
|
1274 |
+
"step": 16300
|
1275 |
+
},
|
1276 |
+
{
|
1277 |
+
"epoch": 47.67,
|
1278 |
+
"learning_rate": 3.986842105263157e-06,
|
1279 |
+
"loss": 1.0102,
|
1280 |
+
"step": 16400
|
1281 |
+
},
|
1282 |
+
{
|
1283 |
+
"epoch": 47.96,
|
1284 |
+
"learning_rate": 3.4934210526315787e-06,
|
1285 |
+
"loss": 1.0422,
|
1286 |
+
"step": 16500
|
1287 |
+
},
|
1288 |
+
{
|
1289 |
+
"epoch": 47.96,
|
1290 |
+
"eval_loss": 0.3183736503124237,
|
1291 |
+
"eval_runtime": 137.2442,
|
1292 |
+
"eval_samples_per_second": 33.663,
|
1293 |
+
"eval_steps_per_second": 4.211,
|
1294 |
+
"eval_wer": 0.2751942727773359,
|
1295 |
+
"step": 16500
|
1296 |
+
}
|
1297 |
+
],
|
1298 |
+
"max_steps": 17200,
|
1299 |
+
"num_train_epochs": 50,
|
1300 |
+
"total_flos": 4.869540507003347e+19,
|
1301 |
+
"trial_name": null,
|
1302 |
+
"trial_params": null
|
1303 |
+
}
|
{checkpoint-1500 β checkpoint-16500}/training_args.bin
RENAMED
File without changes
|
{checkpoint-500 β checkpoint-17000}/config.json
RENAMED
File without changes
|
{checkpoint-500 β checkpoint-17000}/optimizer.pt
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 2490362385
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1daeae5d4afe881cb6c0563fb2d056998741e52f136ae2e7acaac074cf332438
|
3 |
size 2490362385
|
{checkpoint-500 β checkpoint-17000}/preprocessor_config.json
RENAMED
File without changes
|
{checkpoint-500 β checkpoint-17000}/pytorch_model.bin
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 1262075377
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f8f1f33d70822ede911a061ebd9c8ac6adb1b85af8f06aef411beb21f64e5513
|
3 |
size 1262075377
|
{checkpoint-1500 β checkpoint-17000}/rng_state.pth
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:da448dcf17115922df19725e2f419fc357399e18ea720fe14302bdb4ceec2828
|
3 |
+
size 14567
|
{checkpoint-1000 β checkpoint-17000}/scaler.pt
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 559
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7084db4c57c4f9d27d144de3844af74c529a53119c951e7d5083f342e8946b0f
|
3 |
size 559
|
{checkpoint-1500 β checkpoint-17000}/scheduler.pt
RENAMED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 623
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:29137f29f94e1209db7d5266845567f6c8e2dcdd80c671249312f489b8a23f6f
|
3 |
size 623
|
checkpoint-17000/trainer_state.json
ADDED
@@ -0,0 +1,1342 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 49.41769398114576,
|
5 |
+
"global_step": 17000,
|
6 |
+
"is_hyper_param_search": false,
|
7 |
+
"is_local_process_zero": true,
|
8 |
+
"is_world_process_zero": true,
|
9 |
+
"log_history": [
|
10 |
+
{
|
11 |
+
"epoch": 0.29,
|
12 |
+
"learning_rate": 3.7125e-06,
|
13 |
+
"loss": 12.2713,
|
14 |
+
"step": 100
|
15 |
+
},
|
16 |
+
{
|
17 |
+
"epoch": 0.58,
|
18 |
+
"learning_rate": 7.4625e-06,
|
19 |
+
"loss": 6.2026,
|
20 |
+
"step": 200
|
21 |
+
},
|
22 |
+
{
|
23 |
+
"epoch": 0.87,
|
24 |
+
"learning_rate": 1.1212499999999998e-05,
|
25 |
+
"loss": 4.1374,
|
26 |
+
"step": 300
|
27 |
+
},
|
28 |
+
{
|
29 |
+
"epoch": 1.16,
|
30 |
+
"learning_rate": 1.49625e-05,
|
31 |
+
"loss": 3.6755,
|
32 |
+
"step": 400
|
33 |
+
},
|
34 |
+
{
|
35 |
+
"epoch": 1.45,
|
36 |
+
"learning_rate": 1.8712499999999997e-05,
|
37 |
+
"loss": 3.3332,
|
38 |
+
"step": 500
|
39 |
+
},
|
40 |
+
{
|
41 |
+
"epoch": 1.45,
|
42 |
+
"eval_loss": 3.292029857635498,
|
43 |
+
"eval_runtime": 137.2974,
|
44 |
+
"eval_samples_per_second": 33.65,
|
45 |
+
"eval_steps_per_second": 4.21,
|
46 |
+
"eval_wer": 1.0,
|
47 |
+
"step": 500
|
48 |
+
},
|
49 |
+
{
|
50 |
+
"epoch": 1.74,
|
51 |
+
"learning_rate": 2.2462499999999997e-05,
|
52 |
+
"loss": 3.1418,
|
53 |
+
"step": 600
|
54 |
+
},
|
55 |
+
{
|
56 |
+
"epoch": 2.03,
|
57 |
+
"learning_rate": 2.6212499999999997e-05,
|
58 |
+
"loss": 3.0879,
|
59 |
+
"step": 700
|
60 |
+
},
|
61 |
+
{
|
62 |
+
"epoch": 2.32,
|
63 |
+
"learning_rate": 2.99625e-05,
|
64 |
+
"loss": 3.0216,
|
65 |
+
"step": 800
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 2.61,
|
69 |
+
"learning_rate": 3.37125e-05,
|
70 |
+
"loss": 2.9595,
|
71 |
+
"step": 900
|
72 |
+
},
|
73 |
+
{
|
74 |
+
"epoch": 2.91,
|
75 |
+
"learning_rate": 3.7462499999999996e-05,
|
76 |
+
"loss": 2.9269,
|
77 |
+
"step": 1000
|
78 |
+
},
|
79 |
+
{
|
80 |
+
"epoch": 2.91,
|
81 |
+
"eval_loss": 2.941540241241455,
|
82 |
+
"eval_runtime": 178.4388,
|
83 |
+
"eval_samples_per_second": 25.891,
|
84 |
+
"eval_steps_per_second": 3.239,
|
85 |
+
"eval_wer": 0.9966346448020559,
|
86 |
+
"step": 1000
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 3.2,
|
90 |
+
"learning_rate": 4.12125e-05,
|
91 |
+
"loss": 2.914,
|
92 |
+
"step": 1100
|
93 |
+
},
|
94 |
+
{
|
95 |
+
"epoch": 3.49,
|
96 |
+
"learning_rate": 4.4962499999999995e-05,
|
97 |
+
"loss": 2.8432,
|
98 |
+
"step": 1200
|
99 |
+
},
|
100 |
+
{
|
101 |
+
"epoch": 3.78,
|
102 |
+
"learning_rate": 4.871249999999999e-05,
|
103 |
+
"loss": 2.6828,
|
104 |
+
"step": 1300
|
105 |
+
},
|
106 |
+
{
|
107 |
+
"epoch": 4.07,
|
108 |
+
"learning_rate": 5.2462499999999994e-05,
|
109 |
+
"loss": 2.355,
|
110 |
+
"step": 1400
|
111 |
+
},
|
112 |
+
{
|
113 |
+
"epoch": 4.36,
|
114 |
+
"learning_rate": 5.62125e-05,
|
115 |
+
"loss": 2.0719,
|
116 |
+
"step": 1500
|
117 |
+
},
|
118 |
+
{
|
119 |
+
"epoch": 4.36,
|
120 |
+
"eval_loss": 1.164096713066101,
|
121 |
+
"eval_runtime": 134.2981,
|
122 |
+
"eval_samples_per_second": 34.401,
|
123 |
+
"eval_steps_per_second": 4.304,
|
124 |
+
"eval_wer": 0.8507923881784251,
|
125 |
+
"step": 1500
|
126 |
+
},
|
127 |
+
{
|
128 |
+
"epoch": 4.65,
|
129 |
+
"learning_rate": 5.9962499999999994e-05,
|
130 |
+
"loss": 1.944,
|
131 |
+
"step": 1600
|
132 |
+
},
|
133 |
+
{
|
134 |
+
"epoch": 4.94,
|
135 |
+
"learning_rate": 6.37125e-05,
|
136 |
+
"loss": 1.8772,
|
137 |
+
"step": 1700
|
138 |
+
},
|
139 |
+
{
|
140 |
+
"epoch": 5.23,
|
141 |
+
"learning_rate": 6.746249999999999e-05,
|
142 |
+
"loss": 1.807,
|
143 |
+
"step": 1800
|
144 |
+
},
|
145 |
+
{
|
146 |
+
"epoch": 5.52,
|
147 |
+
"learning_rate": 7.121249999999999e-05,
|
148 |
+
"loss": 1.7612,
|
149 |
+
"step": 1900
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 5.81,
|
153 |
+
"learning_rate": 7.492499999999999e-05,
|
154 |
+
"loss": 1.7404,
|
155 |
+
"step": 2000
|
156 |
+
},
|
157 |
+
{
|
158 |
+
"epoch": 5.81,
|
159 |
+
"eval_loss": 0.7280541658401489,
|
160 |
+
"eval_runtime": 174.7873,
|
161 |
+
"eval_samples_per_second": 26.432,
|
162 |
+
"eval_steps_per_second": 3.307,
|
163 |
+
"eval_wer": 0.6846356238144772,
|
164 |
+
"step": 2000
|
165 |
+
},
|
166 |
+
{
|
167 |
+
"epoch": 6.1,
|
168 |
+
"learning_rate": 7.451644736842105e-05,
|
169 |
+
"loss": 1.6883,
|
170 |
+
"step": 2100
|
171 |
+
},
|
172 |
+
{
|
173 |
+
"epoch": 6.39,
|
174 |
+
"learning_rate": 7.402302631578947e-05,
|
175 |
+
"loss": 1.6463,
|
176 |
+
"step": 2200
|
177 |
+
},
|
178 |
+
{
|
179 |
+
"epoch": 6.68,
|
180 |
+
"learning_rate": 7.352960526315788e-05,
|
181 |
+
"loss": 1.6216,
|
182 |
+
"step": 2300
|
183 |
+
},
|
184 |
+
{
|
185 |
+
"epoch": 6.97,
|
186 |
+
"learning_rate": 7.30361842105263e-05,
|
187 |
+
"loss": 1.586,
|
188 |
+
"step": 2400
|
189 |
+
},
|
190 |
+
{
|
191 |
+
"epoch": 7.27,
|
192 |
+
"learning_rate": 7.254276315789473e-05,
|
193 |
+
"loss": 1.5921,
|
194 |
+
"step": 2500
|
195 |
+
},
|
196 |
+
{
|
197 |
+
"epoch": 7.27,
|
198 |
+
"eval_loss": 0.5885757803916931,
|
199 |
+
"eval_runtime": 137.296,
|
200 |
+
"eval_samples_per_second": 33.65,
|
201 |
+
"eval_steps_per_second": 4.21,
|
202 |
+
"eval_wer": 0.5146545921801383,
|
203 |
+
"step": 2500
|
204 |
+
},
|
205 |
+
{
|
206 |
+
"epoch": 7.56,
|
207 |
+
"learning_rate": 7.204934210526316e-05,
|
208 |
+
"loss": 1.5694,
|
209 |
+
"step": 2600
|
210 |
+
},
|
211 |
+
{
|
212 |
+
"epoch": 7.85,
|
213 |
+
"learning_rate": 7.155592105263157e-05,
|
214 |
+
"loss": 1.5279,
|
215 |
+
"step": 2700
|
216 |
+
},
|
217 |
+
{
|
218 |
+
"epoch": 8.14,
|
219 |
+
"learning_rate": 7.10625e-05,
|
220 |
+
"loss": 1.5435,
|
221 |
+
"step": 2800
|
222 |
+
},
|
223 |
+
{
|
224 |
+
"epoch": 8.43,
|
225 |
+
"learning_rate": 7.056907894736841e-05,
|
226 |
+
"loss": 1.5163,
|
227 |
+
"step": 2900
|
228 |
+
},
|
229 |
+
{
|
230 |
+
"epoch": 8.72,
|
231 |
+
"learning_rate": 7.007565789473684e-05,
|
232 |
+
"loss": 1.4941,
|
233 |
+
"step": 3000
|
234 |
+
},
|
235 |
+
{
|
236 |
+
"epoch": 8.72,
|
237 |
+
"eval_loss": 0.5183178186416626,
|
238 |
+
"eval_runtime": 134.3566,
|
239 |
+
"eval_samples_per_second": 34.386,
|
240 |
+
"eval_steps_per_second": 4.302,
|
241 |
+
"eval_wer": 0.5063023924616044,
|
242 |
+
"step": 3000
|
243 |
+
},
|
244 |
+
{
|
245 |
+
"epoch": 9.01,
|
246 |
+
"learning_rate": 6.958223684210525e-05,
|
247 |
+
"loss": 1.5061,
|
248 |
+
"step": 3100
|
249 |
+
},
|
250 |
+
{
|
251 |
+
"epoch": 9.3,
|
252 |
+
"learning_rate": 6.908881578947368e-05,
|
253 |
+
"loss": 1.4551,
|
254 |
+
"step": 3200
|
255 |
+
},
|
256 |
+
{
|
257 |
+
"epoch": 9.59,
|
258 |
+
"learning_rate": 6.859539473684209e-05,
|
259 |
+
"loss": 1.4674,
|
260 |
+
"step": 3300
|
261 |
+
},
|
262 |
+
{
|
263 |
+
"epoch": 9.88,
|
264 |
+
"learning_rate": 6.810197368421052e-05,
|
265 |
+
"loss": 1.4691,
|
266 |
+
"step": 3400
|
267 |
+
},
|
268 |
+
{
|
269 |
+
"epoch": 10.17,
|
270 |
+
"learning_rate": 6.760855263157895e-05,
|
271 |
+
"loss": 1.4486,
|
272 |
+
"step": 3500
|
273 |
+
},
|
274 |
+
{
|
275 |
+
"epoch": 10.17,
|
276 |
+
"eval_loss": 0.47492966055870056,
|
277 |
+
"eval_runtime": 135.1792,
|
278 |
+
"eval_samples_per_second": 34.177,
|
279 |
+
"eval_steps_per_second": 4.276,
|
280 |
+
"eval_wer": 0.46760080768524753,
|
281 |
+
"step": 3500
|
282 |
+
},
|
283 |
+
{
|
284 |
+
"epoch": 10.46,
|
285 |
+
"learning_rate": 6.711513157894736e-05,
|
286 |
+
"loss": 1.4274,
|
287 |
+
"step": 3600
|
288 |
+
},
|
289 |
+
{
|
290 |
+
"epoch": 10.75,
|
291 |
+
"learning_rate": 6.66266447368421e-05,
|
292 |
+
"loss": 1.4277,
|
293 |
+
"step": 3700
|
294 |
+
},
|
295 |
+
{
|
296 |
+
"epoch": 11.05,
|
297 |
+
"learning_rate": 6.613322368421052e-05,
|
298 |
+
"loss": 1.4445,
|
299 |
+
"step": 3800
|
300 |
+
},
|
301 |
+
{
|
302 |
+
"epoch": 11.34,
|
303 |
+
"learning_rate": 6.563980263157894e-05,
|
304 |
+
"loss": 1.4101,
|
305 |
+
"step": 3900
|
306 |
+
},
|
307 |
+
{
|
308 |
+
"epoch": 11.63,
|
309 |
+
"learning_rate": 6.514638157894736e-05,
|
310 |
+
"loss": 1.3899,
|
311 |
+
"step": 4000
|
312 |
+
},
|
313 |
+
{
|
314 |
+
"epoch": 11.63,
|
315 |
+
"eval_loss": 0.4564875364303589,
|
316 |
+
"eval_runtime": 134.5257,
|
317 |
+
"eval_samples_per_second": 34.343,
|
318 |
+
"eval_steps_per_second": 4.297,
|
319 |
+
"eval_wer": 0.4432172795692345,
|
320 |
+
"step": 4000
|
321 |
+
},
|
322 |
+
{
|
323 |
+
"epoch": 11.92,
|
324 |
+
"learning_rate": 6.465296052631578e-05,
|
325 |
+
"loss": 1.404,
|
326 |
+
"step": 4100
|
327 |
+
},
|
328 |
+
{
|
329 |
+
"epoch": 12.21,
|
330 |
+
"learning_rate": 6.415953947368421e-05,
|
331 |
+
"loss": 1.3911,
|
332 |
+
"step": 4200
|
333 |
+
},
|
334 |
+
{
|
335 |
+
"epoch": 12.5,
|
336 |
+
"learning_rate": 6.366611842105262e-05,
|
337 |
+
"loss": 1.3873,
|
338 |
+
"step": 4300
|
339 |
+
},
|
340 |
+
{
|
341 |
+
"epoch": 12.79,
|
342 |
+
"learning_rate": 6.317269736842105e-05,
|
343 |
+
"loss": 1.3712,
|
344 |
+
"step": 4400
|
345 |
+
},
|
346 |
+
{
|
347 |
+
"epoch": 13.08,
|
348 |
+
"learning_rate": 6.267927631578946e-05,
|
349 |
+
"loss": 1.3881,
|
350 |
+
"step": 4500
|
351 |
+
},
|
352 |
+
{
|
353 |
+
"epoch": 13.08,
|
354 |
+
"eval_loss": 0.43156012892723083,
|
355 |
+
"eval_runtime": 134.756,
|
356 |
+
"eval_samples_per_second": 34.284,
|
357 |
+
"eval_steps_per_second": 4.289,
|
358 |
+
"eval_wer": 0.42278039527626504,
|
359 |
+
"step": 4500
|
360 |
+
},
|
361 |
+
{
|
362 |
+
"epoch": 13.37,
|
363 |
+
"learning_rate": 6.218585526315789e-05,
|
364 |
+
"loss": 1.3538,
|
365 |
+
"step": 4600
|
366 |
+
},
|
367 |
+
{
|
368 |
+
"epoch": 13.66,
|
369 |
+
"learning_rate": 6.16924342105263e-05,
|
370 |
+
"loss": 1.355,
|
371 |
+
"step": 4700
|
372 |
+
},
|
373 |
+
{
|
374 |
+
"epoch": 13.95,
|
375 |
+
"learning_rate": 6.119901315789473e-05,
|
376 |
+
"loss": 1.341,
|
377 |
+
"step": 4800
|
378 |
+
},
|
379 |
+
{
|
380 |
+
"epoch": 14.24,
|
381 |
+
"learning_rate": 6.070559210526316e-05,
|
382 |
+
"loss": 1.3474,
|
383 |
+
"step": 4900
|
384 |
+
},
|
385 |
+
{
|
386 |
+
"epoch": 14.53,
|
387 |
+
"learning_rate": 6.021217105263157e-05,
|
388 |
+
"loss": 1.3572,
|
389 |
+
"step": 5000
|
390 |
+
},
|
391 |
+
{
|
392 |
+
"epoch": 14.53,
|
393 |
+
"eval_loss": 0.4194825291633606,
|
394 |
+
"eval_runtime": 134.9338,
|
395 |
+
"eval_samples_per_second": 34.239,
|
396 |
+
"eval_steps_per_second": 4.284,
|
397 |
+
"eval_wer": 0.3834363335984825,
|
398 |
+
"step": 5000
|
399 |
+
},
|
400 |
+
{
|
401 |
+
"epoch": 14.82,
|
402 |
+
"learning_rate": 5.971875e-05,
|
403 |
+
"loss": 1.3408,
|
404 |
+
"step": 5100
|
405 |
+
},
|
406 |
+
{
|
407 |
+
"epoch": 15.12,
|
408 |
+
"learning_rate": 5.922532894736842e-05,
|
409 |
+
"loss": 1.3357,
|
410 |
+
"step": 5200
|
411 |
+
},
|
412 |
+
{
|
413 |
+
"epoch": 15.41,
|
414 |
+
"learning_rate": 5.873190789473683e-05,
|
415 |
+
"loss": 1.3288,
|
416 |
+
"step": 5300
|
417 |
+
},
|
418 |
+
{
|
419 |
+
"epoch": 15.7,
|
420 |
+
"learning_rate": 5.823848684210526e-05,
|
421 |
+
"loss": 1.3258,
|
422 |
+
"step": 5400
|
423 |
+
},
|
424 |
+
{
|
425 |
+
"epoch": 15.99,
|
426 |
+
"learning_rate": 5.7745065789473685e-05,
|
427 |
+
"loss": 1.3261,
|
428 |
+
"step": 5500
|
429 |
+
},
|
430 |
+
{
|
431 |
+
"epoch": 15.99,
|
432 |
+
"eval_loss": 0.3974045217037201,
|
433 |
+
"eval_runtime": 134.8844,
|
434 |
+
"eval_samples_per_second": 34.252,
|
435 |
+
"eval_steps_per_second": 4.285,
|
436 |
+
"eval_wer": 0.3606742948051153,
|
437 |
+
"step": 5500
|
438 |
+
},
|
439 |
+
{
|
440 |
+
"epoch": 16.28,
|
441 |
+
"learning_rate": 5.72516447368421e-05,
|
442 |
+
"loss": 1.3273,
|
443 |
+
"step": 5600
|
444 |
+
},
|
445 |
+
{
|
446 |
+
"epoch": 16.57,
|
447 |
+
"learning_rate": 5.675822368421052e-05,
|
448 |
+
"loss": 1.3131,
|
449 |
+
"step": 5700
|
450 |
+
},
|
451 |
+
{
|
452 |
+
"epoch": 16.86,
|
453 |
+
"learning_rate": 5.6264802631578946e-05,
|
454 |
+
"loss": 1.2739,
|
455 |
+
"step": 5800
|
456 |
+
},
|
457 |
+
{
|
458 |
+
"epoch": 17.15,
|
459 |
+
"learning_rate": 5.577138157894736e-05,
|
460 |
+
"loss": 1.3141,
|
461 |
+
"step": 5900
|
462 |
+
},
|
463 |
+
{
|
464 |
+
"epoch": 17.44,
|
465 |
+
"learning_rate": 5.5277960526315786e-05,
|
466 |
+
"loss": 1.2809,
|
467 |
+
"step": 6000
|
468 |
+
},
|
469 |
+
{
|
470 |
+
"epoch": 17.44,
|
471 |
+
"eval_loss": 0.3844749331474304,
|
472 |
+
"eval_runtime": 136.3592,
|
473 |
+
"eval_samples_per_second": 33.881,
|
474 |
+
"eval_steps_per_second": 4.239,
|
475 |
+
"eval_wer": 0.3466621795264027,
|
476 |
+
"step": 6000
|
477 |
+
},
|
478 |
+
{
|
479 |
+
"epoch": 17.73,
|
480 |
+
"learning_rate": 5.4784539473684206e-05,
|
481 |
+
"loss": 1.2847,
|
482 |
+
"step": 6100
|
483 |
+
},
|
484 |
+
{
|
485 |
+
"epoch": 18.02,
|
486 |
+
"learning_rate": 5.4291118421052626e-05,
|
487 |
+
"loss": 1.303,
|
488 |
+
"step": 6200
|
489 |
+
},
|
490 |
+
{
|
491 |
+
"epoch": 18.31,
|
492 |
+
"learning_rate": 5.3797697368421046e-05,
|
493 |
+
"loss": 1.2733,
|
494 |
+
"step": 6300
|
495 |
+
},
|
496 |
+
{
|
497 |
+
"epoch": 18.6,
|
498 |
+
"learning_rate": 5.330427631578947e-05,
|
499 |
+
"loss": 1.2707,
|
500 |
+
"step": 6400
|
501 |
+
},
|
502 |
+
{
|
503 |
+
"epoch": 18.89,
|
504 |
+
"learning_rate": 5.2810855263157886e-05,
|
505 |
+
"loss": 1.2713,
|
506 |
+
"step": 6500
|
507 |
+
},
|
508 |
+
{
|
509 |
+
"epoch": 18.89,
|
510 |
+
"eval_loss": 0.3831607401371002,
|
511 |
+
"eval_runtime": 138.2456,
|
512 |
+
"eval_samples_per_second": 33.419,
|
513 |
+
"eval_steps_per_second": 4.181,
|
514 |
+
"eval_wer": 0.3449795019274307,
|
515 |
+
"step": 6500
|
516 |
+
},
|
517 |
+
{
|
518 |
+
"epoch": 19.19,
|
519 |
+
"learning_rate": 5.231743421052631e-05,
|
520 |
+
"loss": 1.2562,
|
521 |
+
"step": 6600
|
522 |
+
},
|
523 |
+
{
|
524 |
+
"epoch": 19.48,
|
525 |
+
"learning_rate": 5.182894736842105e-05,
|
526 |
+
"loss": 1.2721,
|
527 |
+
"step": 6700
|
528 |
+
},
|
529 |
+
{
|
530 |
+
"epoch": 19.77,
|
531 |
+
"learning_rate": 5.133552631578947e-05,
|
532 |
+
"loss": 1.2536,
|
533 |
+
"step": 6800
|
534 |
+
},
|
535 |
+
{
|
536 |
+
"epoch": 20.06,
|
537 |
+
"learning_rate": 5.084210526315789e-05,
|
538 |
+
"loss": 1.2503,
|
539 |
+
"step": 6900
|
540 |
+
},
|
541 |
+
{
|
542 |
+
"epoch": 20.35,
|
543 |
+
"learning_rate": 5.0353618421052625e-05,
|
544 |
+
"loss": 1.257,
|
545 |
+
"step": 7000
|
546 |
+
},
|
547 |
+
{
|
548 |
+
"epoch": 20.35,
|
549 |
+
"eval_loss": 0.3778809607028961,
|
550 |
+
"eval_runtime": 138.2403,
|
551 |
+
"eval_samples_per_second": 33.42,
|
552 |
+
"eval_steps_per_second": 4.181,
|
553 |
+
"eval_wer": 0.33726977911032247,
|
554 |
+
"step": 7000
|
555 |
+
},
|
556 |
+
{
|
557 |
+
"epoch": 20.64,
|
558 |
+
"learning_rate": 4.9860197368421045e-05,
|
559 |
+
"loss": 1.2557,
|
560 |
+
"step": 7100
|
561 |
+
},
|
562 |
+
{
|
563 |
+
"epoch": 20.93,
|
564 |
+
"learning_rate": 4.936677631578947e-05,
|
565 |
+
"loss": 1.2397,
|
566 |
+
"step": 7200
|
567 |
+
},
|
568 |
+
{
|
569 |
+
"epoch": 21.22,
|
570 |
+
"learning_rate": 4.8873355263157886e-05,
|
571 |
+
"loss": 1.2527,
|
572 |
+
"step": 7300
|
573 |
+
},
|
574 |
+
{
|
575 |
+
"epoch": 21.51,
|
576 |
+
"learning_rate": 4.837993421052631e-05,
|
577 |
+
"loss": 1.2273,
|
578 |
+
"step": 7400
|
579 |
+
},
|
580 |
+
{
|
581 |
+
"epoch": 21.8,
|
582 |
+
"learning_rate": 4.788651315789473e-05,
|
583 |
+
"loss": 1.2298,
|
584 |
+
"step": 7500
|
585 |
+
},
|
586 |
+
{
|
587 |
+
"epoch": 21.8,
|
588 |
+
"eval_loss": 0.3743567168712616,
|
589 |
+
"eval_runtime": 136.1375,
|
590 |
+
"eval_samples_per_second": 33.936,
|
591 |
+
"eval_steps_per_second": 4.246,
|
592 |
+
"eval_wer": 0.33913602153827327,
|
593 |
+
"step": 7500
|
594 |
+
},
|
595 |
+
{
|
596 |
+
"epoch": 22.09,
|
597 |
+
"learning_rate": 4.739309210526315e-05,
|
598 |
+
"loss": 1.2444,
|
599 |
+
"step": 7600
|
600 |
+
},
|
601 |
+
{
|
602 |
+
"epoch": 22.38,
|
603 |
+
"learning_rate": 4.689967105263157e-05,
|
604 |
+
"loss": 1.2153,
|
605 |
+
"step": 7700
|
606 |
+
},
|
607 |
+
{
|
608 |
+
"epoch": 22.67,
|
609 |
+
"learning_rate": 4.640625e-05,
|
610 |
+
"loss": 1.2327,
|
611 |
+
"step": 7800
|
612 |
+
},
|
613 |
+
{
|
614 |
+
"epoch": 22.96,
|
615 |
+
"learning_rate": 4.591282894736841e-05,
|
616 |
+
"loss": 1.2404,
|
617 |
+
"step": 7900
|
618 |
+
},
|
619 |
+
{
|
620 |
+
"epoch": 23.26,
|
621 |
+
"learning_rate": 4.541940789473684e-05,
|
622 |
+
"loss": 1.2173,
|
623 |
+
"step": 8000
|
624 |
+
},
|
625 |
+
{
|
626 |
+
"epoch": 23.26,
|
627 |
+
"eval_loss": 0.3745496869087219,
|
628 |
+
"eval_runtime": 139.3565,
|
629 |
+
"eval_samples_per_second": 33.152,
|
630 |
+
"eval_steps_per_second": 4.148,
|
631 |
+
"eval_wer": 0.32619470109527016,
|
632 |
+
"step": 8000
|
633 |
+
},
|
634 |
+
{
|
635 |
+
"epoch": 23.55,
|
636 |
+
"learning_rate": 4.492598684210526e-05,
|
637 |
+
"loss": 1.222,
|
638 |
+
"step": 8100
|
639 |
+
},
|
640 |
+
{
|
641 |
+
"epoch": 23.84,
|
642 |
+
"learning_rate": 4.443256578947369e-05,
|
643 |
+
"loss": 1.1868,
|
644 |
+
"step": 8200
|
645 |
+
},
|
646 |
+
{
|
647 |
+
"epoch": 24.13,
|
648 |
+
"learning_rate": 4.39391447368421e-05,
|
649 |
+
"loss": 1.2272,
|
650 |
+
"step": 8300
|
651 |
+
},
|
652 |
+
{
|
653 |
+
"epoch": 24.42,
|
654 |
+
"learning_rate": 4.344572368421052e-05,
|
655 |
+
"loss": 1.2212,
|
656 |
+
"step": 8400
|
657 |
+
},
|
658 |
+
{
|
659 |
+
"epoch": 24.71,
|
660 |
+
"learning_rate": 4.295230263157895e-05,
|
661 |
+
"loss": 1.1966,
|
662 |
+
"step": 8500
|
663 |
+
},
|
664 |
+
{
|
665 |
+
"epoch": 24.71,
|
666 |
+
"eval_loss": 0.36800575256347656,
|
667 |
+
"eval_runtime": 137.1685,
|
668 |
+
"eval_samples_per_second": 33.681,
|
669 |
+
"eval_steps_per_second": 4.214,
|
670 |
+
"eval_wer": 0.32411429970017747,
|
671 |
+
"step": 8500
|
672 |
+
},
|
673 |
+
{
|
674 |
+
"epoch": 25.0,
|
675 |
+
"learning_rate": 4.245888157894736e-05,
|
676 |
+
"loss": 1.2128,
|
677 |
+
"step": 8600
|
678 |
+
},
|
679 |
+
{
|
680 |
+
"epoch": 25.29,
|
681 |
+
"learning_rate": 4.196546052631579e-05,
|
682 |
+
"loss": 1.2118,
|
683 |
+
"step": 8700
|
684 |
+
},
|
685 |
+
{
|
686 |
+
"epoch": 25.58,
|
687 |
+
"learning_rate": 4.147203947368421e-05,
|
688 |
+
"loss": 1.2009,
|
689 |
+
"step": 8800
|
690 |
+
},
|
691 |
+
{
|
692 |
+
"epoch": 25.87,
|
693 |
+
"learning_rate": 4.097861842105263e-05,
|
694 |
+
"loss": 1.1684,
|
695 |
+
"step": 8900
|
696 |
+
},
|
697 |
+
{
|
698 |
+
"epoch": 26.16,
|
699 |
+
"learning_rate": 4.048519736842105e-05,
|
700 |
+
"loss": 1.1925,
|
701 |
+
"step": 9000
|
702 |
+
},
|
703 |
+
{
|
704 |
+
"epoch": 26.16,
|
705 |
+
"eval_loss": 0.3604615330696106,
|
706 |
+
"eval_runtime": 153.4534,
|
707 |
+
"eval_samples_per_second": 30.107,
|
708 |
+
"eval_steps_per_second": 3.767,
|
709 |
+
"eval_wer": 0.31713883619898425,
|
710 |
+
"step": 9000
|
711 |
+
},
|
712 |
+
{
|
713 |
+
"epoch": 26.45,
|
714 |
+
"learning_rate": 3.9991776315789475e-05,
|
715 |
+
"loss": 1.1759,
|
716 |
+
"step": 9100
|
717 |
+
},
|
718 |
+
{
|
719 |
+
"epoch": 26.74,
|
720 |
+
"learning_rate": 3.949835526315789e-05,
|
721 |
+
"loss": 1.1754,
|
722 |
+
"step": 9200
|
723 |
+
},
|
724 |
+
{
|
725 |
+
"epoch": 27.03,
|
726 |
+
"learning_rate": 3.9004934210526315e-05,
|
727 |
+
"loss": 1.198,
|
728 |
+
"step": 9300
|
729 |
+
},
|
730 |
+
{
|
731 |
+
"epoch": 27.32,
|
732 |
+
"learning_rate": 3.8511513157894735e-05,
|
733 |
+
"loss": 1.1602,
|
734 |
+
"step": 9400
|
735 |
+
},
|
736 |
+
{
|
737 |
+
"epoch": 27.61,
|
738 |
+
"learning_rate": 3.801809210526315e-05,
|
739 |
+
"loss": 1.1692,
|
740 |
+
"step": 9500
|
741 |
+
},
|
742 |
+
{
|
743 |
+
"epoch": 27.61,
|
744 |
+
"eval_loss": 0.3512294292449951,
|
745 |
+
"eval_runtime": 166.9619,
|
746 |
+
"eval_samples_per_second": 27.671,
|
747 |
+
"eval_steps_per_second": 3.462,
|
748 |
+
"eval_wer": 0.31472189928409716,
|
749 |
+
"step": 9500
|
750 |
+
},
|
751 |
+
{
|
752 |
+
"epoch": 27.91,
|
753 |
+
"learning_rate": 3.7524671052631575e-05,
|
754 |
+
"loss": 1.1495,
|
755 |
+
"step": 9600
|
756 |
+
},
|
757 |
+
{
|
758 |
+
"epoch": 28.2,
|
759 |
+
"learning_rate": 3.7031249999999995e-05,
|
760 |
+
"loss": 1.1722,
|
761 |
+
"step": 9700
|
762 |
+
},
|
763 |
+
{
|
764 |
+
"epoch": 28.49,
|
765 |
+
"learning_rate": 3.6537828947368416e-05,
|
766 |
+
"loss": 1.1443,
|
767 |
+
"step": 9800
|
768 |
+
},
|
769 |
+
{
|
770 |
+
"epoch": 28.78,
|
771 |
+
"learning_rate": 3.6044407894736836e-05,
|
772 |
+
"loss": 1.1535,
|
773 |
+
"step": 9900
|
774 |
+
},
|
775 |
+
{
|
776 |
+
"epoch": 29.07,
|
777 |
+
"learning_rate": 3.555098684210526e-05,
|
778 |
+
"loss": 1.1704,
|
779 |
+
"step": 10000
|
780 |
+
},
|
781 |
+
{
|
782 |
+
"epoch": 29.07,
|
783 |
+
"eval_loss": 0.3532196581363678,
|
784 |
+
"eval_runtime": 134.5452,
|
785 |
+
"eval_samples_per_second": 34.338,
|
786 |
+
"eval_steps_per_second": 4.296,
|
787 |
+
"eval_wer": 0.30976564890167047,
|
788 |
+
"step": 10000
|
789 |
+
},
|
790 |
+
{
|
791 |
+
"epoch": 29.36,
|
792 |
+
"learning_rate": 3.505756578947368e-05,
|
793 |
+
"loss": 1.1519,
|
794 |
+
"step": 10100
|
795 |
+
},
|
796 |
+
{
|
797 |
+
"epoch": 29.65,
|
798 |
+
"learning_rate": 3.45641447368421e-05,
|
799 |
+
"loss": 1.1507,
|
800 |
+
"step": 10200
|
801 |
+
},
|
802 |
+
{
|
803 |
+
"epoch": 29.94,
|
804 |
+
"learning_rate": 3.407072368421052e-05,
|
805 |
+
"loss": 1.1517,
|
806 |
+
"step": 10300
|
807 |
+
},
|
808 |
+
{
|
809 |
+
"epoch": 30.23,
|
810 |
+
"learning_rate": 3.357730263157894e-05,
|
811 |
+
"loss": 1.1447,
|
812 |
+
"step": 10400
|
813 |
+
},
|
814 |
+
{
|
815 |
+
"epoch": 30.52,
|
816 |
+
"learning_rate": 3.308388157894737e-05,
|
817 |
+
"loss": 1.1595,
|
818 |
+
"step": 10500
|
819 |
+
},
|
820 |
+
{
|
821 |
+
"epoch": 30.52,
|
822 |
+
"eval_loss": 0.34246256947517395,
|
823 |
+
"eval_runtime": 159.7393,
|
824 |
+
"eval_samples_per_second": 28.922,
|
825 |
+
"eval_steps_per_second": 3.618,
|
826 |
+
"eval_wer": 0.3038915743743499,
|
827 |
+
"step": 10500
|
828 |
+
},
|
829 |
+
{
|
830 |
+
"epoch": 30.81,
|
831 |
+
"learning_rate": 3.259046052631579e-05,
|
832 |
+
"loss": 1.1451,
|
833 |
+
"step": 10600
|
834 |
+
},
|
835 |
+
{
|
836 |
+
"epoch": 31.1,
|
837 |
+
"learning_rate": 3.2097039473684203e-05,
|
838 |
+
"loss": 1.1304,
|
839 |
+
"step": 10700
|
840 |
+
},
|
841 |
+
{
|
842 |
+
"epoch": 31.39,
|
843 |
+
"learning_rate": 3.160361842105263e-05,
|
844 |
+
"loss": 1.1316,
|
845 |
+
"step": 10800
|
846 |
+
},
|
847 |
+
{
|
848 |
+
"epoch": 31.68,
|
849 |
+
"learning_rate": 3.111019736842105e-05,
|
850 |
+
"loss": 1.1301,
|
851 |
+
"step": 10900
|
852 |
+
},
|
853 |
+
{
|
854 |
+
"epoch": 31.97,
|
855 |
+
"learning_rate": 3.061677631578947e-05,
|
856 |
+
"loss": 1.1433,
|
857 |
+
"step": 11000
|
858 |
+
},
|
859 |
+
{
|
860 |
+
"epoch": 31.97,
|
861 |
+
"eval_loss": 0.35683709383010864,
|
862 |
+
"eval_runtime": 158.7503,
|
863 |
+
"eval_samples_per_second": 29.102,
|
864 |
+
"eval_steps_per_second": 3.641,
|
865 |
+
"eval_wer": 0.30257602643333537,
|
866 |
+
"step": 11000
|
867 |
+
},
|
868 |
+
{
|
869 |
+
"epoch": 32.27,
|
870 |
+
"learning_rate": 3.0123355263157894e-05,
|
871 |
+
"loss": 1.1337,
|
872 |
+
"step": 11100
|
873 |
+
},
|
874 |
+
{
|
875 |
+
"epoch": 32.56,
|
876 |
+
"learning_rate": 2.9629934210526314e-05,
|
877 |
+
"loss": 1.1227,
|
878 |
+
"step": 11200
|
879 |
+
},
|
880 |
+
{
|
881 |
+
"epoch": 32.85,
|
882 |
+
"learning_rate": 2.913651315789473e-05,
|
883 |
+
"loss": 1.1281,
|
884 |
+
"step": 11300
|
885 |
+
},
|
886 |
+
{
|
887 |
+
"epoch": 33.14,
|
888 |
+
"learning_rate": 2.8643092105263154e-05,
|
889 |
+
"loss": 1.1203,
|
890 |
+
"step": 11400
|
891 |
+
},
|
892 |
+
{
|
893 |
+
"epoch": 33.43,
|
894 |
+
"learning_rate": 2.8149671052631574e-05,
|
895 |
+
"loss": 1.1295,
|
896 |
+
"step": 11500
|
897 |
+
},
|
898 |
+
{
|
899 |
+
"epoch": 33.43,
|
900 |
+
"eval_loss": 0.34605443477630615,
|
901 |
+
"eval_runtime": 160.1749,
|
902 |
+
"eval_samples_per_second": 28.843,
|
903 |
+
"eval_steps_per_second": 3.609,
|
904 |
+
"eval_wer": 0.2992106712353913,
|
905 |
+
"step": 11500
|
906 |
+
},
|
907 |
+
{
|
908 |
+
"epoch": 33.72,
|
909 |
+
"learning_rate": 2.7656249999999998e-05,
|
910 |
+
"loss": 1.0915,
|
911 |
+
"step": 11600
|
912 |
+
},
|
913 |
+
{
|
914 |
+
"epoch": 34.01,
|
915 |
+
"learning_rate": 2.7162828947368418e-05,
|
916 |
+
"loss": 1.1312,
|
917 |
+
"step": 11700
|
918 |
+
},
|
919 |
+
{
|
920 |
+
"epoch": 34.3,
|
921 |
+
"learning_rate": 2.6669407894736838e-05,
|
922 |
+
"loss": 1.1146,
|
923 |
+
"step": 11800
|
924 |
+
},
|
925 |
+
{
|
926 |
+
"epoch": 34.59,
|
927 |
+
"learning_rate": 2.6175986842105262e-05,
|
928 |
+
"loss": 1.1184,
|
929 |
+
"step": 11900
|
930 |
+
},
|
931 |
+
{
|
932 |
+
"epoch": 34.88,
|
933 |
+
"learning_rate": 2.56875e-05,
|
934 |
+
"loss": 1.1131,
|
935 |
+
"step": 12000
|
936 |
+
},
|
937 |
+
{
|
938 |
+
"epoch": 34.88,
|
939 |
+
"eval_loss": 0.3348712623119354,
|
940 |
+
"eval_runtime": 157.6811,
|
941 |
+
"eval_samples_per_second": 29.3,
|
942 |
+
"eval_steps_per_second": 3.666,
|
943 |
+
"eval_wer": 0.29422382671480146,
|
944 |
+
"step": 12000
|
945 |
+
},
|
946 |
+
{
|
947 |
+
"epoch": 35.17,
|
948 |
+
"learning_rate": 2.5194078947368418e-05,
|
949 |
+
"loss": 1.1025,
|
950 |
+
"step": 12100
|
951 |
+
},
|
952 |
+
{
|
953 |
+
"epoch": 35.46,
|
954 |
+
"learning_rate": 2.4700657894736838e-05,
|
955 |
+
"loss": 1.1069,
|
956 |
+
"step": 12200
|
957 |
+
},
|
958 |
+
{
|
959 |
+
"epoch": 35.75,
|
960 |
+
"learning_rate": 2.420723684210526e-05,
|
961 |
+
"loss": 1.076,
|
962 |
+
"step": 12300
|
963 |
+
},
|
964 |
+
{
|
965 |
+
"epoch": 36.05,
|
966 |
+
"learning_rate": 2.371381578947368e-05,
|
967 |
+
"loss": 1.111,
|
968 |
+
"step": 12400
|
969 |
+
},
|
970 |
+
{
|
971 |
+
"epoch": 36.34,
|
972 |
+
"learning_rate": 2.32203947368421e-05,
|
973 |
+
"loss": 1.1015,
|
974 |
+
"step": 12500
|
975 |
+
},
|
976 |
+
{
|
977 |
+
"epoch": 36.34,
|
978 |
+
"eval_loss": 0.3378337025642395,
|
979 |
+
"eval_runtime": 160.9202,
|
980 |
+
"eval_samples_per_second": 28.71,
|
981 |
+
"eval_steps_per_second": 3.592,
|
982 |
+
"eval_wer": 0.29612066328091535,
|
983 |
+
"step": 12500
|
984 |
+
},
|
985 |
+
{
|
986 |
+
"epoch": 36.63,
|
987 |
+
"learning_rate": 2.2726973684210525e-05,
|
988 |
+
"loss": 1.1047,
|
989 |
+
"step": 12600
|
990 |
+
},
|
991 |
+
{
|
992 |
+
"epoch": 36.92,
|
993 |
+
"learning_rate": 2.2233552631578945e-05,
|
994 |
+
"loss": 1.102,
|
995 |
+
"step": 12700
|
996 |
+
},
|
997 |
+
{
|
998 |
+
"epoch": 37.21,
|
999 |
+
"learning_rate": 2.1740131578947365e-05,
|
1000 |
+
"loss": 1.095,
|
1001 |
+
"step": 12800
|
1002 |
+
},
|
1003 |
+
{
|
1004 |
+
"epoch": 37.5,
|
1005 |
+
"learning_rate": 2.124671052631579e-05,
|
1006 |
+
"loss": 1.0709,
|
1007 |
+
"step": 12900
|
1008 |
+
},
|
1009 |
+
{
|
1010 |
+
"epoch": 37.79,
|
1011 |
+
"learning_rate": 2.075328947368421e-05,
|
1012 |
+
"loss": 1.0835,
|
1013 |
+
"step": 13000
|
1014 |
+
},
|
1015 |
+
{
|
1016 |
+
"epoch": 37.79,
|
1017 |
+
"eval_loss": 0.3281959593296051,
|
1018 |
+
"eval_runtime": 159.3279,
|
1019 |
+
"eval_samples_per_second": 28.997,
|
1020 |
+
"eval_steps_per_second": 3.628,
|
1021 |
+
"eval_wer": 0.2865141038976932,
|
1022 |
+
"step": 13000
|
1023 |
+
},
|
1024 |
+
{
|
1025 |
+
"epoch": 38.08,
|
1026 |
+
"learning_rate": 2.0259868421052632e-05,
|
1027 |
+
"loss": 1.0846,
|
1028 |
+
"step": 13100
|
1029 |
+
},
|
1030 |
+
{
|
1031 |
+
"epoch": 38.37,
|
1032 |
+
"learning_rate": 1.9766447368421053e-05,
|
1033 |
+
"loss": 1.0933,
|
1034 |
+
"step": 13200
|
1035 |
+
},
|
1036 |
+
{
|
1037 |
+
"epoch": 38.66,
|
1038 |
+
"learning_rate": 1.927302631578947e-05,
|
1039 |
+
"loss": 1.0661,
|
1040 |
+
"step": 13300
|
1041 |
+
},
|
1042 |
+
{
|
1043 |
+
"epoch": 38.95,
|
1044 |
+
"learning_rate": 1.8779605263157893e-05,
|
1045 |
+
"loss": 1.0902,
|
1046 |
+
"step": 13400
|
1047 |
+
},
|
1048 |
+
{
|
1049 |
+
"epoch": 39.24,
|
1050 |
+
"learning_rate": 1.8286184210526313e-05,
|
1051 |
+
"loss": 1.083,
|
1052 |
+
"step": 13500
|
1053 |
+
},
|
1054 |
+
{
|
1055 |
+
"epoch": 39.24,
|
1056 |
+
"eval_loss": 0.3181643486022949,
|
1057 |
+
"eval_runtime": 158.3527,
|
1058 |
+
"eval_samples_per_second": 29.175,
|
1059 |
+
"eval_steps_per_second": 3.65,
|
1060 |
+
"eval_wer": 0.2825674600746497,
|
1061 |
+
"step": 13500
|
1062 |
+
},
|
1063 |
+
{
|
1064 |
+
"epoch": 39.53,
|
1065 |
+
"learning_rate": 1.7792763157894736e-05,
|
1066 |
+
"loss": 1.0548,
|
1067 |
+
"step": 13600
|
1068 |
+
},
|
1069 |
+
{
|
1070 |
+
"epoch": 39.82,
|
1071 |
+
"learning_rate": 1.7299342105263156e-05,
|
1072 |
+
"loss": 1.0735,
|
1073 |
+
"step": 13700
|
1074 |
+
},
|
1075 |
+
{
|
1076 |
+
"epoch": 40.12,
|
1077 |
+
"learning_rate": 1.6805921052631577e-05,
|
1078 |
+
"loss": 1.085,
|
1079 |
+
"step": 13800
|
1080 |
+
},
|
1081 |
+
{
|
1082 |
+
"epoch": 40.41,
|
1083 |
+
"learning_rate": 1.63125e-05,
|
1084 |
+
"loss": 1.0531,
|
1085 |
+
"step": 13900
|
1086 |
+
},
|
1087 |
+
{
|
1088 |
+
"epoch": 40.7,
|
1089 |
+
"learning_rate": 1.581907894736842e-05,
|
1090 |
+
"loss": 1.0819,
|
1091 |
+
"step": 14000
|
1092 |
+
},
|
1093 |
+
{
|
1094 |
+
"epoch": 40.7,
|
1095 |
+
"eval_loss": 0.32643991708755493,
|
1096 |
+
"eval_runtime": 159.1099,
|
1097 |
+
"eval_samples_per_second": 29.037,
|
1098 |
+
"eval_steps_per_second": 3.633,
|
1099 |
+
"eval_wer": 0.28504558526586304,
|
1100 |
+
"step": 14000
|
1101 |
+
},
|
1102 |
+
{
|
1103 |
+
"epoch": 40.99,
|
1104 |
+
"learning_rate": 1.532565789473684e-05,
|
1105 |
+
"loss": 1.0679,
|
1106 |
+
"step": 14100
|
1107 |
+
},
|
1108 |
+
{
|
1109 |
+
"epoch": 41.28,
|
1110 |
+
"learning_rate": 1.4832236842105262e-05,
|
1111 |
+
"loss": 1.0773,
|
1112 |
+
"step": 14200
|
1113 |
+
},
|
1114 |
+
{
|
1115 |
+
"epoch": 41.57,
|
1116 |
+
"learning_rate": 1.4338815789473682e-05,
|
1117 |
+
"loss": 1.0587,
|
1118 |
+
"step": 14300
|
1119 |
+
},
|
1120 |
+
{
|
1121 |
+
"epoch": 41.86,
|
1122 |
+
"learning_rate": 1.3845394736842104e-05,
|
1123 |
+
"loss": 1.0287,
|
1124 |
+
"step": 14400
|
1125 |
+
},
|
1126 |
+
{
|
1127 |
+
"epoch": 42.15,
|
1128 |
+
"learning_rate": 1.3351973684210524e-05,
|
1129 |
+
"loss": 1.072,
|
1130 |
+
"step": 14500
|
1131 |
+
},
|
1132 |
+
{
|
1133 |
+
"epoch": 42.15,
|
1134 |
+
"eval_loss": 0.32792535424232483,
|
1135 |
+
"eval_runtime": 155.2521,
|
1136 |
+
"eval_samples_per_second": 29.758,
|
1137 |
+
"eval_steps_per_second": 3.723,
|
1138 |
+
"eval_wer": 0.2817108242060821,
|
1139 |
+
"step": 14500
|
1140 |
+
},
|
1141 |
+
{
|
1142 |
+
"epoch": 42.44,
|
1143 |
+
"learning_rate": 1.2863486842105262e-05,
|
1144 |
+
"loss": 1.0508,
|
1145 |
+
"step": 14600
|
1146 |
+
},
|
1147 |
+
{
|
1148 |
+
"epoch": 42.73,
|
1149 |
+
"learning_rate": 1.2370065789473684e-05,
|
1150 |
+
"loss": 1.0408,
|
1151 |
+
"step": 14700
|
1152 |
+
},
|
1153 |
+
{
|
1154 |
+
"epoch": 43.02,
|
1155 |
+
"learning_rate": 1.1876644736842105e-05,
|
1156 |
+
"loss": 1.0725,
|
1157 |
+
"step": 14800
|
1158 |
+
},
|
1159 |
+
{
|
1160 |
+
"epoch": 43.31,
|
1161 |
+
"learning_rate": 1.1383223684210525e-05,
|
1162 |
+
"loss": 1.0381,
|
1163 |
+
"step": 14900
|
1164 |
+
},
|
1165 |
+
{
|
1166 |
+
"epoch": 43.6,
|
1167 |
+
"learning_rate": 1.0889802631578946e-05,
|
1168 |
+
"loss": 1.0456,
|
1169 |
+
"step": 15000
|
1170 |
+
},
|
1171 |
+
{
|
1172 |
+
"epoch": 43.6,
|
1173 |
+
"eval_loss": 0.323445200920105,
|
1174 |
+
"eval_runtime": 137.3413,
|
1175 |
+
"eval_samples_per_second": 33.639,
|
1176 |
+
"eval_steps_per_second": 4.208,
|
1177 |
+
"eval_wer": 0.2792632931530319,
|
1178 |
+
"step": 15000
|
1179 |
+
},
|
1180 |
+
{
|
1181 |
+
"epoch": 43.89,
|
1182 |
+
"learning_rate": 1.0396381578947367e-05,
|
1183 |
+
"loss": 1.0655,
|
1184 |
+
"step": 15100
|
1185 |
+
},
|
1186 |
+
{
|
1187 |
+
"epoch": 44.19,
|
1188 |
+
"learning_rate": 9.90296052631579e-06,
|
1189 |
+
"loss": 1.0509,
|
1190 |
+
"step": 15200
|
1191 |
+
},
|
1192 |
+
{
|
1193 |
+
"epoch": 44.48,
|
1194 |
+
"learning_rate": 9.40953947368421e-06,
|
1195 |
+
"loss": 1.0357,
|
1196 |
+
"step": 15300
|
1197 |
+
},
|
1198 |
+
{
|
1199 |
+
"epoch": 44.77,
|
1200 |
+
"learning_rate": 8.916118421052631e-06,
|
1201 |
+
"loss": 1.0467,
|
1202 |
+
"step": 15400
|
1203 |
+
},
|
1204 |
+
{
|
1205 |
+
"epoch": 45.06,
|
1206 |
+
"learning_rate": 8.422697368421051e-06,
|
1207 |
+
"loss": 1.0581,
|
1208 |
+
"step": 15500
|
1209 |
+
},
|
1210 |
+
{
|
1211 |
+
"epoch": 45.06,
|
1212 |
+
"eval_loss": 0.32200726866722107,
|
1213 |
+
"eval_runtime": 138.9222,
|
1214 |
+
"eval_samples_per_second": 33.256,
|
1215 |
+
"eval_steps_per_second": 4.161,
|
1216 |
+
"eval_wer": 0.27791715107385423,
|
1217 |
+
"step": 15500
|
1218 |
+
},
|
1219 |
+
{
|
1220 |
+
"epoch": 45.35,
|
1221 |
+
"learning_rate": 7.929276315789473e-06,
|
1222 |
+
"loss": 1.0512,
|
1223 |
+
"step": 15600
|
1224 |
+
},
|
1225 |
+
{
|
1226 |
+
"epoch": 45.64,
|
1227 |
+
"learning_rate": 7.435855263157894e-06,
|
1228 |
+
"loss": 1.0397,
|
1229 |
+
"step": 15700
|
1230 |
+
},
|
1231 |
+
{
|
1232 |
+
"epoch": 45.93,
|
1233 |
+
"learning_rate": 6.9473684210526315e-06,
|
1234 |
+
"loss": 1.0455,
|
1235 |
+
"step": 15800
|
1236 |
+
},
|
1237 |
+
{
|
1238 |
+
"epoch": 46.22,
|
1239 |
+
"learning_rate": 6.453947368421052e-06,
|
1240 |
+
"loss": 1.0347,
|
1241 |
+
"step": 15900
|
1242 |
+
},
|
1243 |
+
{
|
1244 |
+
"epoch": 46.51,
|
1245 |
+
"learning_rate": 5.9605263157894735e-06,
|
1246 |
+
"loss": 1.0406,
|
1247 |
+
"step": 16000
|
1248 |
+
},
|
1249 |
+
{
|
1250 |
+
"epoch": 46.51,
|
1251 |
+
"eval_loss": 0.32076749205589294,
|
1252 |
+
"eval_runtime": 134.5334,
|
1253 |
+
"eval_samples_per_second": 34.341,
|
1254 |
+
"eval_steps_per_second": 4.296,
|
1255 |
+
"eval_wer": 0.2762344734748822,
|
1256 |
+
"step": 16000
|
1257 |
+
},
|
1258 |
+
{
|
1259 |
+
"epoch": 46.8,
|
1260 |
+
"learning_rate": 5.467105263157894e-06,
|
1261 |
+
"loss": 1.0528,
|
1262 |
+
"step": 16100
|
1263 |
+
},
|
1264 |
+
{
|
1265 |
+
"epoch": 47.09,
|
1266 |
+
"learning_rate": 4.973684210526316e-06,
|
1267 |
+
"loss": 1.045,
|
1268 |
+
"step": 16200
|
1269 |
+
},
|
1270 |
+
{
|
1271 |
+
"epoch": 47.38,
|
1272 |
+
"learning_rate": 4.480263157894736e-06,
|
1273 |
+
"loss": 1.0129,
|
1274 |
+
"step": 16300
|
1275 |
+
},
|
1276 |
+
{
|
1277 |
+
"epoch": 47.67,
|
1278 |
+
"learning_rate": 3.986842105263157e-06,
|
1279 |
+
"loss": 1.0102,
|
1280 |
+
"step": 16400
|
1281 |
+
},
|
1282 |
+
{
|
1283 |
+
"epoch": 47.96,
|
1284 |
+
"learning_rate": 3.4934210526315787e-06,
|
1285 |
+
"loss": 1.0422,
|
1286 |
+
"step": 16500
|
1287 |
+
},
|
1288 |
+
{
|
1289 |
+
"epoch": 47.96,
|
1290 |
+
"eval_loss": 0.3183736503124237,
|
1291 |
+
"eval_runtime": 137.2442,
|
1292 |
+
"eval_samples_per_second": 33.663,
|
1293 |
+
"eval_steps_per_second": 4.211,
|
1294 |
+
"eval_wer": 0.2751942727773359,
|
1295 |
+
"step": 16500
|
1296 |
+
},
|
1297 |
+
{
|
1298 |
+
"epoch": 48.26,
|
1299 |
+
"learning_rate": 2.9999999999999997e-06,
|
1300 |
+
"loss": 1.0323,
|
1301 |
+
"step": 16600
|
1302 |
+
},
|
1303 |
+
{
|
1304 |
+
"epoch": 48.55,
|
1305 |
+
"learning_rate": 2.506578947368421e-06,
|
1306 |
+
"loss": 1.0292,
|
1307 |
+
"step": 16700
|
1308 |
+
},
|
1309 |
+
{
|
1310 |
+
"epoch": 48.84,
|
1311 |
+
"learning_rate": 2.013157894736842e-06,
|
1312 |
+
"loss": 1.0378,
|
1313 |
+
"step": 16800
|
1314 |
+
},
|
1315 |
+
{
|
1316 |
+
"epoch": 49.13,
|
1317 |
+
"learning_rate": 1.519736842105263e-06,
|
1318 |
+
"loss": 1.035,
|
1319 |
+
"step": 16900
|
1320 |
+
},
|
1321 |
+
{
|
1322 |
+
"epoch": 49.42,
|
1323 |
+
"learning_rate": 1.0263157894736841e-06,
|
1324 |
+
"loss": 1.0099,
|
1325 |
+
"step": 17000
|
1326 |
+
},
|
1327 |
+
{
|
1328 |
+
"epoch": 49.42,
|
1329 |
+
"eval_loss": 0.31810811161994934,
|
1330 |
+
"eval_runtime": 134.533,
|
1331 |
+
"eval_samples_per_second": 34.341,
|
1332 |
+
"eval_steps_per_second": 4.296,
|
1333 |
+
"eval_wer": 0.2734810010402007,
|
1334 |
+
"step": 17000
|
1335 |
+
}
|
1336 |
+
],
|
1337 |
+
"max_steps": 17200,
|
1338 |
+
"num_train_epochs": 50,
|
1339 |
+
"total_flos": 5.016617790448189e+19,
|
1340 |
+
"trial_name": null,
|
1341 |
+
"trial_params": null
|
1342 |
+
}
|
{checkpoint-500 β checkpoint-17000}/training_args.bin
RENAMED
File without changes
|
checkpoint-500/trainer_state.json
DELETED
@@ -1,55 +0,0 @@
|
|
1 |
-
{
|
2 |
-
"best_metric": null,
|
3 |
-
"best_model_checkpoint": null,
|
4 |
-
"epoch": 1.4525018129079044,
|
5 |
-
"global_step": 500,
|
6 |
-
"is_hyper_param_search": false,
|
7 |
-
"is_local_process_zero": true,
|
8 |
-
"is_world_process_zero": true,
|
9 |
-
"log_history": [
|
10 |
-
{
|
11 |
-
"epoch": 0.29,
|
12 |
-
"learning_rate": 3.7125e-06,
|
13 |
-
"loss": 12.2713,
|
14 |
-
"step": 100
|
15 |
-
},
|
16 |
-
{
|
17 |
-
"epoch": 0.58,
|
18 |
-
"learning_rate": 7.4625e-06,
|
19 |
-
"loss": 6.2026,
|
20 |
-
"step": 200
|
21 |
-
},
|
22 |
-
{
|
23 |
-
"epoch": 0.87,
|
24 |
-
"learning_rate": 1.1212499999999998e-05,
|
25 |
-
"loss": 4.1374,
|
26 |
-
"step": 300
|
27 |
-
},
|
28 |
-
{
|
29 |
-
"epoch": 1.16,
|
30 |
-
"learning_rate": 1.49625e-05,
|
31 |
-
"loss": 3.6755,
|
32 |
-
"step": 400
|
33 |
-
},
|
34 |
-
{
|
35 |
-
"epoch": 1.45,
|
36 |
-
"learning_rate": 1.8712499999999997e-05,
|
37 |
-
"loss": 3.3332,
|
38 |
-
"step": 500
|
39 |
-
},
|
40 |
-
{
|
41 |
-
"epoch": 1.45,
|
42 |
-
"eval_loss": 3.292029857635498,
|
43 |
-
"eval_runtime": 137.2974,
|
44 |
-
"eval_samples_per_second": 33.65,
|
45 |
-
"eval_steps_per_second": 4.21,
|
46 |
-
"eval_wer": 1.0,
|
47 |
-
"step": 500
|
48 |
-
}
|
49 |
-
],
|
50 |
-
"max_steps": 17200,
|
51 |
-
"num_train_epochs": 50,
|
52 |
-
"total_flos": 1.4827194756605722e+18,
|
53 |
-
"trial_name": null,
|
54 |
-
"trial_params": null
|
55 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
eval_results.json
CHANGED
@@ -1,9 +1,9 @@
|
|
1 |
{
|
2 |
-
"epoch":
|
3 |
-
"eval_loss":
|
4 |
-
"eval_runtime":
|
5 |
"eval_samples": 4620,
|
6 |
-
"eval_samples_per_second":
|
7 |
-
"eval_steps_per_second": 4.
|
8 |
-
"eval_wer": 0.
|
9 |
}
|
|
|
1 |
{
|
2 |
+
"epoch": 50.0,
|
3 |
+
"eval_loss": 0.31790396571159363,
|
4 |
+
"eval_runtime": 136.0793,
|
5 |
"eval_samples": 4620,
|
6 |
+
"eval_samples_per_second": 33.951,
|
7 |
+
"eval_steps_per_second": 4.248,
|
8 |
+
"eval_wer": 0.2734810010402007
|
9 |
}
|
pytorch_model.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 1262075377
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e501ae3712d27da72be4224776458113a2e4b0d236983057ecd2414a0d777c3e
|
3 |
size 1262075377
|
train_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
-
"epoch":
|
3 |
-
"train_loss":
|
4 |
-
"train_runtime":
|
5 |
"train_samples": 11030,
|
6 |
-
"train_samples_per_second": 18.
|
7 |
-
"train_steps_per_second": 0.
|
8 |
}
|
|
|
1 |
{
|
2 |
+
"epoch": 50.0,
|
3 |
+
"train_loss": 1.4555730460410894,
|
4 |
+
"train_runtime": 29339.4334,
|
5 |
"train_samples": 11030,
|
6 |
+
"train_samples_per_second": 18.797,
|
7 |
+
"train_steps_per_second": 0.586
|
8 |
}
|
trainer_state.json
CHANGED
@@ -1,154 +1,1363 @@
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
-
"epoch":
|
5 |
-
"global_step":
|
6 |
"is_hyper_param_search": false,
|
7 |
"is_local_process_zero": true,
|
8 |
"is_world_process_zero": true,
|
9 |
"log_history": [
|
10 |
{
|
11 |
"epoch": 0.29,
|
12 |
-
"learning_rate":
|
13 |
-
"loss":
|
14 |
"step": 100
|
15 |
},
|
16 |
{
|
17 |
"epoch": 0.58,
|
18 |
-
"learning_rate":
|
19 |
-
"loss":
|
20 |
"step": 200
|
21 |
},
|
22 |
{
|
23 |
"epoch": 0.87,
|
24 |
-
"learning_rate":
|
25 |
-
"loss":
|
26 |
"step": 300
|
27 |
},
|
28 |
{
|
29 |
"epoch": 1.16,
|
30 |
-
"learning_rate":
|
31 |
-
"loss":
|
32 |
"step": 400
|
33 |
},
|
34 |
{
|
35 |
"epoch": 1.45,
|
36 |
-
"learning_rate":
|
37 |
-
"loss":
|
38 |
"step": 500
|
39 |
},
|
40 |
{
|
41 |
"epoch": 1.45,
|
42 |
-
"eval_loss":
|
43 |
-
"eval_runtime":
|
44 |
-
"eval_samples_per_second":
|
45 |
-
"eval_steps_per_second": 4.
|
46 |
"eval_wer": 1.0,
|
47 |
"step": 500
|
48 |
},
|
49 |
{
|
50 |
"epoch": 1.74,
|
51 |
-
"learning_rate":
|
52 |
-
"loss":
|
53 |
"step": 600
|
54 |
},
|
55 |
{
|
56 |
"epoch": 2.03,
|
57 |
-
"learning_rate":
|
58 |
-
"loss":
|
59 |
"step": 700
|
60 |
},
|
61 |
{
|
62 |
"epoch": 2.32,
|
63 |
-
"learning_rate":
|
64 |
-
"loss":
|
65 |
"step": 800
|
66 |
},
|
67 |
{
|
68 |
"epoch": 2.61,
|
69 |
-
"learning_rate": 3.
|
70 |
-
"loss": 2.
|
71 |
"step": 900
|
72 |
},
|
73 |
{
|
74 |
"epoch": 2.91,
|
75 |
-
"learning_rate": 3.
|
76 |
-
"loss": 2.
|
77 |
"step": 1000
|
78 |
},
|
79 |
{
|
80 |
"epoch": 2.91,
|
81 |
-
"eval_loss": 2.
|
82 |
-
"eval_runtime":
|
83 |
-
"eval_samples_per_second":
|
84 |
-
"eval_steps_per_second": 3.
|
85 |
-
"eval_wer":
|
86 |
"step": 1000
|
87 |
},
|
88 |
{
|
89 |
"epoch": 3.2,
|
90 |
-
"learning_rate":
|
91 |
-
"loss": 2.
|
92 |
"step": 1100
|
93 |
},
|
94 |
{
|
95 |
"epoch": 3.49,
|
96 |
-
"learning_rate":
|
97 |
-
"loss": 2.
|
98 |
"step": 1200
|
99 |
},
|
100 |
{
|
101 |
"epoch": 3.78,
|
102 |
-
"learning_rate":
|
103 |
-
"loss": 2.
|
104 |
"step": 1300
|
105 |
},
|
106 |
{
|
107 |
"epoch": 4.07,
|
108 |
-
"learning_rate":
|
109 |
-
"loss": 2.
|
110 |
"step": 1400
|
111 |
},
|
112 |
{
|
113 |
"epoch": 4.36,
|
114 |
-
"learning_rate":
|
115 |
-
"loss": 2.
|
116 |
"step": 1500
|
117 |
},
|
118 |
{
|
119 |
"epoch": 4.36,
|
120 |
-
"eval_loss": 1.
|
121 |
-
"eval_runtime":
|
122 |
-
"eval_samples_per_second":
|
123 |
-
"eval_steps_per_second": 4.
|
124 |
-
"eval_wer": 0.
|
125 |
"step": 1500
|
126 |
},
|
127 |
{
|
128 |
"epoch": 4.65,
|
129 |
-
"learning_rate": 5.
|
130 |
-
"loss":
|
131 |
"step": 1600
|
132 |
},
|
133 |
{
|
134 |
"epoch": 4.94,
|
135 |
-
"learning_rate":
|
136 |
-
"loss":
|
137 |
"step": 1700
|
138 |
},
|
139 |
{
|
140 |
-
"epoch": 5.
|
141 |
-
"
|
142 |
-
"
|
143 |
-
"
|
144 |
-
|
145 |
-
|
146 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
147 |
}
|
148 |
],
|
149 |
-
"max_steps":
|
150 |
-
"num_train_epochs":
|
151 |
-
"total_flos": 5.
|
152 |
"trial_name": null,
|
153 |
"trial_params": null
|
154 |
}
|
|
|
1 |
{
|
2 |
"best_metric": null,
|
3 |
"best_model_checkpoint": null,
|
4 |
+
"epoch": 49.99782451051487,
|
5 |
+
"global_step": 17200,
|
6 |
"is_hyper_param_search": false,
|
7 |
"is_local_process_zero": true,
|
8 |
"is_world_process_zero": true,
|
9 |
"log_history": [
|
10 |
{
|
11 |
"epoch": 0.29,
|
12 |
+
"learning_rate": 3.7125e-06,
|
13 |
+
"loss": 12.2713,
|
14 |
"step": 100
|
15 |
},
|
16 |
{
|
17 |
"epoch": 0.58,
|
18 |
+
"learning_rate": 7.4625e-06,
|
19 |
+
"loss": 6.2026,
|
20 |
"step": 200
|
21 |
},
|
22 |
{
|
23 |
"epoch": 0.87,
|
24 |
+
"learning_rate": 1.1212499999999998e-05,
|
25 |
+
"loss": 4.1374,
|
26 |
"step": 300
|
27 |
},
|
28 |
{
|
29 |
"epoch": 1.16,
|
30 |
+
"learning_rate": 1.49625e-05,
|
31 |
+
"loss": 3.6755,
|
32 |
"step": 400
|
33 |
},
|
34 |
{
|
35 |
"epoch": 1.45,
|
36 |
+
"learning_rate": 1.8712499999999997e-05,
|
37 |
+
"loss": 3.3332,
|
38 |
"step": 500
|
39 |
},
|
40 |
{
|
41 |
"epoch": 1.45,
|
42 |
+
"eval_loss": 3.292029857635498,
|
43 |
+
"eval_runtime": 137.2974,
|
44 |
+
"eval_samples_per_second": 33.65,
|
45 |
+
"eval_steps_per_second": 4.21,
|
46 |
"eval_wer": 1.0,
|
47 |
"step": 500
|
48 |
},
|
49 |
{
|
50 |
"epoch": 1.74,
|
51 |
+
"learning_rate": 2.2462499999999997e-05,
|
52 |
+
"loss": 3.1418,
|
53 |
"step": 600
|
54 |
},
|
55 |
{
|
56 |
"epoch": 2.03,
|
57 |
+
"learning_rate": 2.6212499999999997e-05,
|
58 |
+
"loss": 3.0879,
|
59 |
"step": 700
|
60 |
},
|
61 |
{
|
62 |
"epoch": 2.32,
|
63 |
+
"learning_rate": 2.99625e-05,
|
64 |
+
"loss": 3.0216,
|
65 |
"step": 800
|
66 |
},
|
67 |
{
|
68 |
"epoch": 2.61,
|
69 |
+
"learning_rate": 3.37125e-05,
|
70 |
+
"loss": 2.9595,
|
71 |
"step": 900
|
72 |
},
|
73 |
{
|
74 |
"epoch": 2.91,
|
75 |
+
"learning_rate": 3.7462499999999996e-05,
|
76 |
+
"loss": 2.9269,
|
77 |
"step": 1000
|
78 |
},
|
79 |
{
|
80 |
"epoch": 2.91,
|
81 |
+
"eval_loss": 2.941540241241455,
|
82 |
+
"eval_runtime": 178.4388,
|
83 |
+
"eval_samples_per_second": 25.891,
|
84 |
+
"eval_steps_per_second": 3.239,
|
85 |
+
"eval_wer": 0.9966346448020559,
|
86 |
"step": 1000
|
87 |
},
|
88 |
{
|
89 |
"epoch": 3.2,
|
90 |
+
"learning_rate": 4.12125e-05,
|
91 |
+
"loss": 2.914,
|
92 |
"step": 1100
|
93 |
},
|
94 |
{
|
95 |
"epoch": 3.49,
|
96 |
+
"learning_rate": 4.4962499999999995e-05,
|
97 |
+
"loss": 2.8432,
|
98 |
"step": 1200
|
99 |
},
|
100 |
{
|
101 |
"epoch": 3.78,
|
102 |
+
"learning_rate": 4.871249999999999e-05,
|
103 |
+
"loss": 2.6828,
|
104 |
"step": 1300
|
105 |
},
|
106 |
{
|
107 |
"epoch": 4.07,
|
108 |
+
"learning_rate": 5.2462499999999994e-05,
|
109 |
+
"loss": 2.355,
|
110 |
"step": 1400
|
111 |
},
|
112 |
{
|
113 |
"epoch": 4.36,
|
114 |
+
"learning_rate": 5.62125e-05,
|
115 |
+
"loss": 2.0719,
|
116 |
"step": 1500
|
117 |
},
|
118 |
{
|
119 |
"epoch": 4.36,
|
120 |
+
"eval_loss": 1.164096713066101,
|
121 |
+
"eval_runtime": 134.2981,
|
122 |
+
"eval_samples_per_second": 34.401,
|
123 |
+
"eval_steps_per_second": 4.304,
|
124 |
+
"eval_wer": 0.8507923881784251,
|
125 |
"step": 1500
|
126 |
},
|
127 |
{
|
128 |
"epoch": 4.65,
|
129 |
+
"learning_rate": 5.9962499999999994e-05,
|
130 |
+
"loss": 1.944,
|
131 |
"step": 1600
|
132 |
},
|
133 |
{
|
134 |
"epoch": 4.94,
|
135 |
+
"learning_rate": 6.37125e-05,
|
136 |
+
"loss": 1.8772,
|
137 |
"step": 1700
|
138 |
},
|
139 |
{
|
140 |
+
"epoch": 5.23,
|
141 |
+
"learning_rate": 6.746249999999999e-05,
|
142 |
+
"loss": 1.807,
|
143 |
+
"step": 1800
|
144 |
+
},
|
145 |
+
{
|
146 |
+
"epoch": 5.52,
|
147 |
+
"learning_rate": 7.121249999999999e-05,
|
148 |
+
"loss": 1.7612,
|
149 |
+
"step": 1900
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 5.81,
|
153 |
+
"learning_rate": 7.492499999999999e-05,
|
154 |
+
"loss": 1.7404,
|
155 |
+
"step": 2000
|
156 |
+
},
|
157 |
+
{
|
158 |
+
"epoch": 5.81,
|
159 |
+
"eval_loss": 0.7280541658401489,
|
160 |
+
"eval_runtime": 174.7873,
|
161 |
+
"eval_samples_per_second": 26.432,
|
162 |
+
"eval_steps_per_second": 3.307,
|
163 |
+
"eval_wer": 0.6846356238144772,
|
164 |
+
"step": 2000
|
165 |
+
},
|
166 |
+
{
|
167 |
+
"epoch": 6.1,
|
168 |
+
"learning_rate": 7.451644736842105e-05,
|
169 |
+
"loss": 1.6883,
|
170 |
+
"step": 2100
|
171 |
+
},
|
172 |
+
{
|
173 |
+
"epoch": 6.39,
|
174 |
+
"learning_rate": 7.402302631578947e-05,
|
175 |
+
"loss": 1.6463,
|
176 |
+
"step": 2200
|
177 |
+
},
|
178 |
+
{
|
179 |
+
"epoch": 6.68,
|
180 |
+
"learning_rate": 7.352960526315788e-05,
|
181 |
+
"loss": 1.6216,
|
182 |
+
"step": 2300
|
183 |
+
},
|
184 |
+
{
|
185 |
+
"epoch": 6.97,
|
186 |
+
"learning_rate": 7.30361842105263e-05,
|
187 |
+
"loss": 1.586,
|
188 |
+
"step": 2400
|
189 |
+
},
|
190 |
+
{
|
191 |
+
"epoch": 7.27,
|
192 |
+
"learning_rate": 7.254276315789473e-05,
|
193 |
+
"loss": 1.5921,
|
194 |
+
"step": 2500
|
195 |
+
},
|
196 |
+
{
|
197 |
+
"epoch": 7.27,
|
198 |
+
"eval_loss": 0.5885757803916931,
|
199 |
+
"eval_runtime": 137.296,
|
200 |
+
"eval_samples_per_second": 33.65,
|
201 |
+
"eval_steps_per_second": 4.21,
|
202 |
+
"eval_wer": 0.5146545921801383,
|
203 |
+
"step": 2500
|
204 |
+
},
|
205 |
+
{
|
206 |
+
"epoch": 7.56,
|
207 |
+
"learning_rate": 7.204934210526316e-05,
|
208 |
+
"loss": 1.5694,
|
209 |
+
"step": 2600
|
210 |
+
},
|
211 |
+
{
|
212 |
+
"epoch": 7.85,
|
213 |
+
"learning_rate": 7.155592105263157e-05,
|
214 |
+
"loss": 1.5279,
|
215 |
+
"step": 2700
|
216 |
+
},
|
217 |
+
{
|
218 |
+
"epoch": 8.14,
|
219 |
+
"learning_rate": 7.10625e-05,
|
220 |
+
"loss": 1.5435,
|
221 |
+
"step": 2800
|
222 |
+
},
|
223 |
+
{
|
224 |
+
"epoch": 8.43,
|
225 |
+
"learning_rate": 7.056907894736841e-05,
|
226 |
+
"loss": 1.5163,
|
227 |
+
"step": 2900
|
228 |
+
},
|
229 |
+
{
|
230 |
+
"epoch": 8.72,
|
231 |
+
"learning_rate": 7.007565789473684e-05,
|
232 |
+
"loss": 1.4941,
|
233 |
+
"step": 3000
|
234 |
+
},
|
235 |
+
{
|
236 |
+
"epoch": 8.72,
|
237 |
+
"eval_loss": 0.5183178186416626,
|
238 |
+
"eval_runtime": 134.3566,
|
239 |
+
"eval_samples_per_second": 34.386,
|
240 |
+
"eval_steps_per_second": 4.302,
|
241 |
+
"eval_wer": 0.5063023924616044,
|
242 |
+
"step": 3000
|
243 |
+
},
|
244 |
+
{
|
245 |
+
"epoch": 9.01,
|
246 |
+
"learning_rate": 6.958223684210525e-05,
|
247 |
+
"loss": 1.5061,
|
248 |
+
"step": 3100
|
249 |
+
},
|
250 |
+
{
|
251 |
+
"epoch": 9.3,
|
252 |
+
"learning_rate": 6.908881578947368e-05,
|
253 |
+
"loss": 1.4551,
|
254 |
+
"step": 3200
|
255 |
+
},
|
256 |
+
{
|
257 |
+
"epoch": 9.59,
|
258 |
+
"learning_rate": 6.859539473684209e-05,
|
259 |
+
"loss": 1.4674,
|
260 |
+
"step": 3300
|
261 |
+
},
|
262 |
+
{
|
263 |
+
"epoch": 9.88,
|
264 |
+
"learning_rate": 6.810197368421052e-05,
|
265 |
+
"loss": 1.4691,
|
266 |
+
"step": 3400
|
267 |
+
},
|
268 |
+
{
|
269 |
+
"epoch": 10.17,
|
270 |
+
"learning_rate": 6.760855263157895e-05,
|
271 |
+
"loss": 1.4486,
|
272 |
+
"step": 3500
|
273 |
+
},
|
274 |
+
{
|
275 |
+
"epoch": 10.17,
|
276 |
+
"eval_loss": 0.47492966055870056,
|
277 |
+
"eval_runtime": 135.1792,
|
278 |
+
"eval_samples_per_second": 34.177,
|
279 |
+
"eval_steps_per_second": 4.276,
|
280 |
+
"eval_wer": 0.46760080768524753,
|
281 |
+
"step": 3500
|
282 |
+
},
|
283 |
+
{
|
284 |
+
"epoch": 10.46,
|
285 |
+
"learning_rate": 6.711513157894736e-05,
|
286 |
+
"loss": 1.4274,
|
287 |
+
"step": 3600
|
288 |
+
},
|
289 |
+
{
|
290 |
+
"epoch": 10.75,
|
291 |
+
"learning_rate": 6.66266447368421e-05,
|
292 |
+
"loss": 1.4277,
|
293 |
+
"step": 3700
|
294 |
+
},
|
295 |
+
{
|
296 |
+
"epoch": 11.05,
|
297 |
+
"learning_rate": 6.613322368421052e-05,
|
298 |
+
"loss": 1.4445,
|
299 |
+
"step": 3800
|
300 |
+
},
|
301 |
+
{
|
302 |
+
"epoch": 11.34,
|
303 |
+
"learning_rate": 6.563980263157894e-05,
|
304 |
+
"loss": 1.4101,
|
305 |
+
"step": 3900
|
306 |
+
},
|
307 |
+
{
|
308 |
+
"epoch": 11.63,
|
309 |
+
"learning_rate": 6.514638157894736e-05,
|
310 |
+
"loss": 1.3899,
|
311 |
+
"step": 4000
|
312 |
+
},
|
313 |
+
{
|
314 |
+
"epoch": 11.63,
|
315 |
+
"eval_loss": 0.4564875364303589,
|
316 |
+
"eval_runtime": 134.5257,
|
317 |
+
"eval_samples_per_second": 34.343,
|
318 |
+
"eval_steps_per_second": 4.297,
|
319 |
+
"eval_wer": 0.4432172795692345,
|
320 |
+
"step": 4000
|
321 |
+
},
|
322 |
+
{
|
323 |
+
"epoch": 11.92,
|
324 |
+
"learning_rate": 6.465296052631578e-05,
|
325 |
+
"loss": 1.404,
|
326 |
+
"step": 4100
|
327 |
+
},
|
328 |
+
{
|
329 |
+
"epoch": 12.21,
|
330 |
+
"learning_rate": 6.415953947368421e-05,
|
331 |
+
"loss": 1.3911,
|
332 |
+
"step": 4200
|
333 |
+
},
|
334 |
+
{
|
335 |
+
"epoch": 12.5,
|
336 |
+
"learning_rate": 6.366611842105262e-05,
|
337 |
+
"loss": 1.3873,
|
338 |
+
"step": 4300
|
339 |
+
},
|
340 |
+
{
|
341 |
+
"epoch": 12.79,
|
342 |
+
"learning_rate": 6.317269736842105e-05,
|
343 |
+
"loss": 1.3712,
|
344 |
+
"step": 4400
|
345 |
+
},
|
346 |
+
{
|
347 |
+
"epoch": 13.08,
|
348 |
+
"learning_rate": 6.267927631578946e-05,
|
349 |
+
"loss": 1.3881,
|
350 |
+
"step": 4500
|
351 |
+
},
|
352 |
+
{
|
353 |
+
"epoch": 13.08,
|
354 |
+
"eval_loss": 0.43156012892723083,
|
355 |
+
"eval_runtime": 134.756,
|
356 |
+
"eval_samples_per_second": 34.284,
|
357 |
+
"eval_steps_per_second": 4.289,
|
358 |
+
"eval_wer": 0.42278039527626504,
|
359 |
+
"step": 4500
|
360 |
+
},
|
361 |
+
{
|
362 |
+
"epoch": 13.37,
|
363 |
+
"learning_rate": 6.218585526315789e-05,
|
364 |
+
"loss": 1.3538,
|
365 |
+
"step": 4600
|
366 |
+
},
|
367 |
+
{
|
368 |
+
"epoch": 13.66,
|
369 |
+
"learning_rate": 6.16924342105263e-05,
|
370 |
+
"loss": 1.355,
|
371 |
+
"step": 4700
|
372 |
+
},
|
373 |
+
{
|
374 |
+
"epoch": 13.95,
|
375 |
+
"learning_rate": 6.119901315789473e-05,
|
376 |
+
"loss": 1.341,
|
377 |
+
"step": 4800
|
378 |
+
},
|
379 |
+
{
|
380 |
+
"epoch": 14.24,
|
381 |
+
"learning_rate": 6.070559210526316e-05,
|
382 |
+
"loss": 1.3474,
|
383 |
+
"step": 4900
|
384 |
+
},
|
385 |
+
{
|
386 |
+
"epoch": 14.53,
|
387 |
+
"learning_rate": 6.021217105263157e-05,
|
388 |
+
"loss": 1.3572,
|
389 |
+
"step": 5000
|
390 |
+
},
|
391 |
+
{
|
392 |
+
"epoch": 14.53,
|
393 |
+
"eval_loss": 0.4194825291633606,
|
394 |
+
"eval_runtime": 134.9338,
|
395 |
+
"eval_samples_per_second": 34.239,
|
396 |
+
"eval_steps_per_second": 4.284,
|
397 |
+
"eval_wer": 0.3834363335984825,
|
398 |
+
"step": 5000
|
399 |
+
},
|
400 |
+
{
|
401 |
+
"epoch": 14.82,
|
402 |
+
"learning_rate": 5.971875e-05,
|
403 |
+
"loss": 1.3408,
|
404 |
+
"step": 5100
|
405 |
+
},
|
406 |
+
{
|
407 |
+
"epoch": 15.12,
|
408 |
+
"learning_rate": 5.922532894736842e-05,
|
409 |
+
"loss": 1.3357,
|
410 |
+
"step": 5200
|
411 |
+
},
|
412 |
+
{
|
413 |
+
"epoch": 15.41,
|
414 |
+
"learning_rate": 5.873190789473683e-05,
|
415 |
+
"loss": 1.3288,
|
416 |
+
"step": 5300
|
417 |
+
},
|
418 |
+
{
|
419 |
+
"epoch": 15.7,
|
420 |
+
"learning_rate": 5.823848684210526e-05,
|
421 |
+
"loss": 1.3258,
|
422 |
+
"step": 5400
|
423 |
+
},
|
424 |
+
{
|
425 |
+
"epoch": 15.99,
|
426 |
+
"learning_rate": 5.7745065789473685e-05,
|
427 |
+
"loss": 1.3261,
|
428 |
+
"step": 5500
|
429 |
+
},
|
430 |
+
{
|
431 |
+
"epoch": 15.99,
|
432 |
+
"eval_loss": 0.3974045217037201,
|
433 |
+
"eval_runtime": 134.8844,
|
434 |
+
"eval_samples_per_second": 34.252,
|
435 |
+
"eval_steps_per_second": 4.285,
|
436 |
+
"eval_wer": 0.3606742948051153,
|
437 |
+
"step": 5500
|
438 |
+
},
|
439 |
+
{
|
440 |
+
"epoch": 16.28,
|
441 |
+
"learning_rate": 5.72516447368421e-05,
|
442 |
+
"loss": 1.3273,
|
443 |
+
"step": 5600
|
444 |
+
},
|
445 |
+
{
|
446 |
+
"epoch": 16.57,
|
447 |
+
"learning_rate": 5.675822368421052e-05,
|
448 |
+
"loss": 1.3131,
|
449 |
+
"step": 5700
|
450 |
+
},
|
451 |
+
{
|
452 |
+
"epoch": 16.86,
|
453 |
+
"learning_rate": 5.6264802631578946e-05,
|
454 |
+
"loss": 1.2739,
|
455 |
+
"step": 5800
|
456 |
+
},
|
457 |
+
{
|
458 |
+
"epoch": 17.15,
|
459 |
+
"learning_rate": 5.577138157894736e-05,
|
460 |
+
"loss": 1.3141,
|
461 |
+
"step": 5900
|
462 |
+
},
|
463 |
+
{
|
464 |
+
"epoch": 17.44,
|
465 |
+
"learning_rate": 5.5277960526315786e-05,
|
466 |
+
"loss": 1.2809,
|
467 |
+
"step": 6000
|
468 |
+
},
|
469 |
+
{
|
470 |
+
"epoch": 17.44,
|
471 |
+
"eval_loss": 0.3844749331474304,
|
472 |
+
"eval_runtime": 136.3592,
|
473 |
+
"eval_samples_per_second": 33.881,
|
474 |
+
"eval_steps_per_second": 4.239,
|
475 |
+
"eval_wer": 0.3466621795264027,
|
476 |
+
"step": 6000
|
477 |
+
},
|
478 |
+
{
|
479 |
+
"epoch": 17.73,
|
480 |
+
"learning_rate": 5.4784539473684206e-05,
|
481 |
+
"loss": 1.2847,
|
482 |
+
"step": 6100
|
483 |
+
},
|
484 |
+
{
|
485 |
+
"epoch": 18.02,
|
486 |
+
"learning_rate": 5.4291118421052626e-05,
|
487 |
+
"loss": 1.303,
|
488 |
+
"step": 6200
|
489 |
+
},
|
490 |
+
{
|
491 |
+
"epoch": 18.31,
|
492 |
+
"learning_rate": 5.3797697368421046e-05,
|
493 |
+
"loss": 1.2733,
|
494 |
+
"step": 6300
|
495 |
+
},
|
496 |
+
{
|
497 |
+
"epoch": 18.6,
|
498 |
+
"learning_rate": 5.330427631578947e-05,
|
499 |
+
"loss": 1.2707,
|
500 |
+
"step": 6400
|
501 |
+
},
|
502 |
+
{
|
503 |
+
"epoch": 18.89,
|
504 |
+
"learning_rate": 5.2810855263157886e-05,
|
505 |
+
"loss": 1.2713,
|
506 |
+
"step": 6500
|
507 |
+
},
|
508 |
+
{
|
509 |
+
"epoch": 18.89,
|
510 |
+
"eval_loss": 0.3831607401371002,
|
511 |
+
"eval_runtime": 138.2456,
|
512 |
+
"eval_samples_per_second": 33.419,
|
513 |
+
"eval_steps_per_second": 4.181,
|
514 |
+
"eval_wer": 0.3449795019274307,
|
515 |
+
"step": 6500
|
516 |
+
},
|
517 |
+
{
|
518 |
+
"epoch": 19.19,
|
519 |
+
"learning_rate": 5.231743421052631e-05,
|
520 |
+
"loss": 1.2562,
|
521 |
+
"step": 6600
|
522 |
+
},
|
523 |
+
{
|
524 |
+
"epoch": 19.48,
|
525 |
+
"learning_rate": 5.182894736842105e-05,
|
526 |
+
"loss": 1.2721,
|
527 |
+
"step": 6700
|
528 |
+
},
|
529 |
+
{
|
530 |
+
"epoch": 19.77,
|
531 |
+
"learning_rate": 5.133552631578947e-05,
|
532 |
+
"loss": 1.2536,
|
533 |
+
"step": 6800
|
534 |
+
},
|
535 |
+
{
|
536 |
+
"epoch": 20.06,
|
537 |
+
"learning_rate": 5.084210526315789e-05,
|
538 |
+
"loss": 1.2503,
|
539 |
+
"step": 6900
|
540 |
+
},
|
541 |
+
{
|
542 |
+
"epoch": 20.35,
|
543 |
+
"learning_rate": 5.0353618421052625e-05,
|
544 |
+
"loss": 1.257,
|
545 |
+
"step": 7000
|
546 |
+
},
|
547 |
+
{
|
548 |
+
"epoch": 20.35,
|
549 |
+
"eval_loss": 0.3778809607028961,
|
550 |
+
"eval_runtime": 138.2403,
|
551 |
+
"eval_samples_per_second": 33.42,
|
552 |
+
"eval_steps_per_second": 4.181,
|
553 |
+
"eval_wer": 0.33726977911032247,
|
554 |
+
"step": 7000
|
555 |
+
},
|
556 |
+
{
|
557 |
+
"epoch": 20.64,
|
558 |
+
"learning_rate": 4.9860197368421045e-05,
|
559 |
+
"loss": 1.2557,
|
560 |
+
"step": 7100
|
561 |
+
},
|
562 |
+
{
|
563 |
+
"epoch": 20.93,
|
564 |
+
"learning_rate": 4.936677631578947e-05,
|
565 |
+
"loss": 1.2397,
|
566 |
+
"step": 7200
|
567 |
+
},
|
568 |
+
{
|
569 |
+
"epoch": 21.22,
|
570 |
+
"learning_rate": 4.8873355263157886e-05,
|
571 |
+
"loss": 1.2527,
|
572 |
+
"step": 7300
|
573 |
+
},
|
574 |
+
{
|
575 |
+
"epoch": 21.51,
|
576 |
+
"learning_rate": 4.837993421052631e-05,
|
577 |
+
"loss": 1.2273,
|
578 |
+
"step": 7400
|
579 |
+
},
|
580 |
+
{
|
581 |
+
"epoch": 21.8,
|
582 |
+
"learning_rate": 4.788651315789473e-05,
|
583 |
+
"loss": 1.2298,
|
584 |
+
"step": 7500
|
585 |
+
},
|
586 |
+
{
|
587 |
+
"epoch": 21.8,
|
588 |
+
"eval_loss": 0.3743567168712616,
|
589 |
+
"eval_runtime": 136.1375,
|
590 |
+
"eval_samples_per_second": 33.936,
|
591 |
+
"eval_steps_per_second": 4.246,
|
592 |
+
"eval_wer": 0.33913602153827327,
|
593 |
+
"step": 7500
|
594 |
+
},
|
595 |
+
{
|
596 |
+
"epoch": 22.09,
|
597 |
+
"learning_rate": 4.739309210526315e-05,
|
598 |
+
"loss": 1.2444,
|
599 |
+
"step": 7600
|
600 |
+
},
|
601 |
+
{
|
602 |
+
"epoch": 22.38,
|
603 |
+
"learning_rate": 4.689967105263157e-05,
|
604 |
+
"loss": 1.2153,
|
605 |
+
"step": 7700
|
606 |
+
},
|
607 |
+
{
|
608 |
+
"epoch": 22.67,
|
609 |
+
"learning_rate": 4.640625e-05,
|
610 |
+
"loss": 1.2327,
|
611 |
+
"step": 7800
|
612 |
+
},
|
613 |
+
{
|
614 |
+
"epoch": 22.96,
|
615 |
+
"learning_rate": 4.591282894736841e-05,
|
616 |
+
"loss": 1.2404,
|
617 |
+
"step": 7900
|
618 |
+
},
|
619 |
+
{
|
620 |
+
"epoch": 23.26,
|
621 |
+
"learning_rate": 4.541940789473684e-05,
|
622 |
+
"loss": 1.2173,
|
623 |
+
"step": 8000
|
624 |
+
},
|
625 |
+
{
|
626 |
+
"epoch": 23.26,
|
627 |
+
"eval_loss": 0.3745496869087219,
|
628 |
+
"eval_runtime": 139.3565,
|
629 |
+
"eval_samples_per_second": 33.152,
|
630 |
+
"eval_steps_per_second": 4.148,
|
631 |
+
"eval_wer": 0.32619470109527016,
|
632 |
+
"step": 8000
|
633 |
+
},
|
634 |
+
{
|
635 |
+
"epoch": 23.55,
|
636 |
+
"learning_rate": 4.492598684210526e-05,
|
637 |
+
"loss": 1.222,
|
638 |
+
"step": 8100
|
639 |
+
},
|
640 |
+
{
|
641 |
+
"epoch": 23.84,
|
642 |
+
"learning_rate": 4.443256578947369e-05,
|
643 |
+
"loss": 1.1868,
|
644 |
+
"step": 8200
|
645 |
+
},
|
646 |
+
{
|
647 |
+
"epoch": 24.13,
|
648 |
+
"learning_rate": 4.39391447368421e-05,
|
649 |
+
"loss": 1.2272,
|
650 |
+
"step": 8300
|
651 |
+
},
|
652 |
+
{
|
653 |
+
"epoch": 24.42,
|
654 |
+
"learning_rate": 4.344572368421052e-05,
|
655 |
+
"loss": 1.2212,
|
656 |
+
"step": 8400
|
657 |
+
},
|
658 |
+
{
|
659 |
+
"epoch": 24.71,
|
660 |
+
"learning_rate": 4.295230263157895e-05,
|
661 |
+
"loss": 1.1966,
|
662 |
+
"step": 8500
|
663 |
+
},
|
664 |
+
{
|
665 |
+
"epoch": 24.71,
|
666 |
+
"eval_loss": 0.36800575256347656,
|
667 |
+
"eval_runtime": 137.1685,
|
668 |
+
"eval_samples_per_second": 33.681,
|
669 |
+
"eval_steps_per_second": 4.214,
|
670 |
+
"eval_wer": 0.32411429970017747,
|
671 |
+
"step": 8500
|
672 |
+
},
|
673 |
+
{
|
674 |
+
"epoch": 25.0,
|
675 |
+
"learning_rate": 4.245888157894736e-05,
|
676 |
+
"loss": 1.2128,
|
677 |
+
"step": 8600
|
678 |
+
},
|
679 |
+
{
|
680 |
+
"epoch": 25.29,
|
681 |
+
"learning_rate": 4.196546052631579e-05,
|
682 |
+
"loss": 1.2118,
|
683 |
+
"step": 8700
|
684 |
+
},
|
685 |
+
{
|
686 |
+
"epoch": 25.58,
|
687 |
+
"learning_rate": 4.147203947368421e-05,
|
688 |
+
"loss": 1.2009,
|
689 |
+
"step": 8800
|
690 |
+
},
|
691 |
+
{
|
692 |
+
"epoch": 25.87,
|
693 |
+
"learning_rate": 4.097861842105263e-05,
|
694 |
+
"loss": 1.1684,
|
695 |
+
"step": 8900
|
696 |
+
},
|
697 |
+
{
|
698 |
+
"epoch": 26.16,
|
699 |
+
"learning_rate": 4.048519736842105e-05,
|
700 |
+
"loss": 1.1925,
|
701 |
+
"step": 9000
|
702 |
+
},
|
703 |
+
{
|
704 |
+
"epoch": 26.16,
|
705 |
+
"eval_loss": 0.3604615330696106,
|
706 |
+
"eval_runtime": 153.4534,
|
707 |
+
"eval_samples_per_second": 30.107,
|
708 |
+
"eval_steps_per_second": 3.767,
|
709 |
+
"eval_wer": 0.31713883619898425,
|
710 |
+
"step": 9000
|
711 |
+
},
|
712 |
+
{
|
713 |
+
"epoch": 26.45,
|
714 |
+
"learning_rate": 3.9991776315789475e-05,
|
715 |
+
"loss": 1.1759,
|
716 |
+
"step": 9100
|
717 |
+
},
|
718 |
+
{
|
719 |
+
"epoch": 26.74,
|
720 |
+
"learning_rate": 3.949835526315789e-05,
|
721 |
+
"loss": 1.1754,
|
722 |
+
"step": 9200
|
723 |
+
},
|
724 |
+
{
|
725 |
+
"epoch": 27.03,
|
726 |
+
"learning_rate": 3.9004934210526315e-05,
|
727 |
+
"loss": 1.198,
|
728 |
+
"step": 9300
|
729 |
+
},
|
730 |
+
{
|
731 |
+
"epoch": 27.32,
|
732 |
+
"learning_rate": 3.8511513157894735e-05,
|
733 |
+
"loss": 1.1602,
|
734 |
+
"step": 9400
|
735 |
+
},
|
736 |
+
{
|
737 |
+
"epoch": 27.61,
|
738 |
+
"learning_rate": 3.801809210526315e-05,
|
739 |
+
"loss": 1.1692,
|
740 |
+
"step": 9500
|
741 |
+
},
|
742 |
+
{
|
743 |
+
"epoch": 27.61,
|
744 |
+
"eval_loss": 0.3512294292449951,
|
745 |
+
"eval_runtime": 166.9619,
|
746 |
+
"eval_samples_per_second": 27.671,
|
747 |
+
"eval_steps_per_second": 3.462,
|
748 |
+
"eval_wer": 0.31472189928409716,
|
749 |
+
"step": 9500
|
750 |
+
},
|
751 |
+
{
|
752 |
+
"epoch": 27.91,
|
753 |
+
"learning_rate": 3.7524671052631575e-05,
|
754 |
+
"loss": 1.1495,
|
755 |
+
"step": 9600
|
756 |
+
},
|
757 |
+
{
|
758 |
+
"epoch": 28.2,
|
759 |
+
"learning_rate": 3.7031249999999995e-05,
|
760 |
+
"loss": 1.1722,
|
761 |
+
"step": 9700
|
762 |
+
},
|
763 |
+
{
|
764 |
+
"epoch": 28.49,
|
765 |
+
"learning_rate": 3.6537828947368416e-05,
|
766 |
+
"loss": 1.1443,
|
767 |
+
"step": 9800
|
768 |
+
},
|
769 |
+
{
|
770 |
+
"epoch": 28.78,
|
771 |
+
"learning_rate": 3.6044407894736836e-05,
|
772 |
+
"loss": 1.1535,
|
773 |
+
"step": 9900
|
774 |
+
},
|
775 |
+
{
|
776 |
+
"epoch": 29.07,
|
777 |
+
"learning_rate": 3.555098684210526e-05,
|
778 |
+
"loss": 1.1704,
|
779 |
+
"step": 10000
|
780 |
+
},
|
781 |
+
{
|
782 |
+
"epoch": 29.07,
|
783 |
+
"eval_loss": 0.3532196581363678,
|
784 |
+
"eval_runtime": 134.5452,
|
785 |
+
"eval_samples_per_second": 34.338,
|
786 |
+
"eval_steps_per_second": 4.296,
|
787 |
+
"eval_wer": 0.30976564890167047,
|
788 |
+
"step": 10000
|
789 |
+
},
|
790 |
+
{
|
791 |
+
"epoch": 29.36,
|
792 |
+
"learning_rate": 3.505756578947368e-05,
|
793 |
+
"loss": 1.1519,
|
794 |
+
"step": 10100
|
795 |
+
},
|
796 |
+
{
|
797 |
+
"epoch": 29.65,
|
798 |
+
"learning_rate": 3.45641447368421e-05,
|
799 |
+
"loss": 1.1507,
|
800 |
+
"step": 10200
|
801 |
+
},
|
802 |
+
{
|
803 |
+
"epoch": 29.94,
|
804 |
+
"learning_rate": 3.407072368421052e-05,
|
805 |
+
"loss": 1.1517,
|
806 |
+
"step": 10300
|
807 |
+
},
|
808 |
+
{
|
809 |
+
"epoch": 30.23,
|
810 |
+
"learning_rate": 3.357730263157894e-05,
|
811 |
+
"loss": 1.1447,
|
812 |
+
"step": 10400
|
813 |
+
},
|
814 |
+
{
|
815 |
+
"epoch": 30.52,
|
816 |
+
"learning_rate": 3.308388157894737e-05,
|
817 |
+
"loss": 1.1595,
|
818 |
+
"step": 10500
|
819 |
+
},
|
820 |
+
{
|
821 |
+
"epoch": 30.52,
|
822 |
+
"eval_loss": 0.34246256947517395,
|
823 |
+
"eval_runtime": 159.7393,
|
824 |
+
"eval_samples_per_second": 28.922,
|
825 |
+
"eval_steps_per_second": 3.618,
|
826 |
+
"eval_wer": 0.3038915743743499,
|
827 |
+
"step": 10500
|
828 |
+
},
|
829 |
+
{
|
830 |
+
"epoch": 30.81,
|
831 |
+
"learning_rate": 3.259046052631579e-05,
|
832 |
+
"loss": 1.1451,
|
833 |
+
"step": 10600
|
834 |
+
},
|
835 |
+
{
|
836 |
+
"epoch": 31.1,
|
837 |
+
"learning_rate": 3.2097039473684203e-05,
|
838 |
+
"loss": 1.1304,
|
839 |
+
"step": 10700
|
840 |
+
},
|
841 |
+
{
|
842 |
+
"epoch": 31.39,
|
843 |
+
"learning_rate": 3.160361842105263e-05,
|
844 |
+
"loss": 1.1316,
|
845 |
+
"step": 10800
|
846 |
+
},
|
847 |
+
{
|
848 |
+
"epoch": 31.68,
|
849 |
+
"learning_rate": 3.111019736842105e-05,
|
850 |
+
"loss": 1.1301,
|
851 |
+
"step": 10900
|
852 |
+
},
|
853 |
+
{
|
854 |
+
"epoch": 31.97,
|
855 |
+
"learning_rate": 3.061677631578947e-05,
|
856 |
+
"loss": 1.1433,
|
857 |
+
"step": 11000
|
858 |
+
},
|
859 |
+
{
|
860 |
+
"epoch": 31.97,
|
861 |
+
"eval_loss": 0.35683709383010864,
|
862 |
+
"eval_runtime": 158.7503,
|
863 |
+
"eval_samples_per_second": 29.102,
|
864 |
+
"eval_steps_per_second": 3.641,
|
865 |
+
"eval_wer": 0.30257602643333537,
|
866 |
+
"step": 11000
|
867 |
+
},
|
868 |
+
{
|
869 |
+
"epoch": 32.27,
|
870 |
+
"learning_rate": 3.0123355263157894e-05,
|
871 |
+
"loss": 1.1337,
|
872 |
+
"step": 11100
|
873 |
+
},
|
874 |
+
{
|
875 |
+
"epoch": 32.56,
|
876 |
+
"learning_rate": 2.9629934210526314e-05,
|
877 |
+
"loss": 1.1227,
|
878 |
+
"step": 11200
|
879 |
+
},
|
880 |
+
{
|
881 |
+
"epoch": 32.85,
|
882 |
+
"learning_rate": 2.913651315789473e-05,
|
883 |
+
"loss": 1.1281,
|
884 |
+
"step": 11300
|
885 |
+
},
|
886 |
+
{
|
887 |
+
"epoch": 33.14,
|
888 |
+
"learning_rate": 2.8643092105263154e-05,
|
889 |
+
"loss": 1.1203,
|
890 |
+
"step": 11400
|
891 |
+
},
|
892 |
+
{
|
893 |
+
"epoch": 33.43,
|
894 |
+
"learning_rate": 2.8149671052631574e-05,
|
895 |
+
"loss": 1.1295,
|
896 |
+
"step": 11500
|
897 |
+
},
|
898 |
+
{
|
899 |
+
"epoch": 33.43,
|
900 |
+
"eval_loss": 0.34605443477630615,
|
901 |
+
"eval_runtime": 160.1749,
|
902 |
+
"eval_samples_per_second": 28.843,
|
903 |
+
"eval_steps_per_second": 3.609,
|
904 |
+
"eval_wer": 0.2992106712353913,
|
905 |
+
"step": 11500
|
906 |
+
},
|
907 |
+
{
|
908 |
+
"epoch": 33.72,
|
909 |
+
"learning_rate": 2.7656249999999998e-05,
|
910 |
+
"loss": 1.0915,
|
911 |
+
"step": 11600
|
912 |
+
},
|
913 |
+
{
|
914 |
+
"epoch": 34.01,
|
915 |
+
"learning_rate": 2.7162828947368418e-05,
|
916 |
+
"loss": 1.1312,
|
917 |
+
"step": 11700
|
918 |
+
},
|
919 |
+
{
|
920 |
+
"epoch": 34.3,
|
921 |
+
"learning_rate": 2.6669407894736838e-05,
|
922 |
+
"loss": 1.1146,
|
923 |
+
"step": 11800
|
924 |
+
},
|
925 |
+
{
|
926 |
+
"epoch": 34.59,
|
927 |
+
"learning_rate": 2.6175986842105262e-05,
|
928 |
+
"loss": 1.1184,
|
929 |
+
"step": 11900
|
930 |
+
},
|
931 |
+
{
|
932 |
+
"epoch": 34.88,
|
933 |
+
"learning_rate": 2.56875e-05,
|
934 |
+
"loss": 1.1131,
|
935 |
+
"step": 12000
|
936 |
+
},
|
937 |
+
{
|
938 |
+
"epoch": 34.88,
|
939 |
+
"eval_loss": 0.3348712623119354,
|
940 |
+
"eval_runtime": 157.6811,
|
941 |
+
"eval_samples_per_second": 29.3,
|
942 |
+
"eval_steps_per_second": 3.666,
|
943 |
+
"eval_wer": 0.29422382671480146,
|
944 |
+
"step": 12000
|
945 |
+
},
|
946 |
+
{
|
947 |
+
"epoch": 35.17,
|
948 |
+
"learning_rate": 2.5194078947368418e-05,
|
949 |
+
"loss": 1.1025,
|
950 |
+
"step": 12100
|
951 |
+
},
|
952 |
+
{
|
953 |
+
"epoch": 35.46,
|
954 |
+
"learning_rate": 2.4700657894736838e-05,
|
955 |
+
"loss": 1.1069,
|
956 |
+
"step": 12200
|
957 |
+
},
|
958 |
+
{
|
959 |
+
"epoch": 35.75,
|
960 |
+
"learning_rate": 2.420723684210526e-05,
|
961 |
+
"loss": 1.076,
|
962 |
+
"step": 12300
|
963 |
+
},
|
964 |
+
{
|
965 |
+
"epoch": 36.05,
|
966 |
+
"learning_rate": 2.371381578947368e-05,
|
967 |
+
"loss": 1.111,
|
968 |
+
"step": 12400
|
969 |
+
},
|
970 |
+
{
|
971 |
+
"epoch": 36.34,
|
972 |
+
"learning_rate": 2.32203947368421e-05,
|
973 |
+
"loss": 1.1015,
|
974 |
+
"step": 12500
|
975 |
+
},
|
976 |
+
{
|
977 |
+
"epoch": 36.34,
|
978 |
+
"eval_loss": 0.3378337025642395,
|
979 |
+
"eval_runtime": 160.9202,
|
980 |
+
"eval_samples_per_second": 28.71,
|
981 |
+
"eval_steps_per_second": 3.592,
|
982 |
+
"eval_wer": 0.29612066328091535,
|
983 |
+
"step": 12500
|
984 |
+
},
|
985 |
+
{
|
986 |
+
"epoch": 36.63,
|
987 |
+
"learning_rate": 2.2726973684210525e-05,
|
988 |
+
"loss": 1.1047,
|
989 |
+
"step": 12600
|
990 |
+
},
|
991 |
+
{
|
992 |
+
"epoch": 36.92,
|
993 |
+
"learning_rate": 2.2233552631578945e-05,
|
994 |
+
"loss": 1.102,
|
995 |
+
"step": 12700
|
996 |
+
},
|
997 |
+
{
|
998 |
+
"epoch": 37.21,
|
999 |
+
"learning_rate": 2.1740131578947365e-05,
|
1000 |
+
"loss": 1.095,
|
1001 |
+
"step": 12800
|
1002 |
+
},
|
1003 |
+
{
|
1004 |
+
"epoch": 37.5,
|
1005 |
+
"learning_rate": 2.124671052631579e-05,
|
1006 |
+
"loss": 1.0709,
|
1007 |
+
"step": 12900
|
1008 |
+
},
|
1009 |
+
{
|
1010 |
+
"epoch": 37.79,
|
1011 |
+
"learning_rate": 2.075328947368421e-05,
|
1012 |
+
"loss": 1.0835,
|
1013 |
+
"step": 13000
|
1014 |
+
},
|
1015 |
+
{
|
1016 |
+
"epoch": 37.79,
|
1017 |
+
"eval_loss": 0.3281959593296051,
|
1018 |
+
"eval_runtime": 159.3279,
|
1019 |
+
"eval_samples_per_second": 28.997,
|
1020 |
+
"eval_steps_per_second": 3.628,
|
1021 |
+
"eval_wer": 0.2865141038976932,
|
1022 |
+
"step": 13000
|
1023 |
+
},
|
1024 |
+
{
|
1025 |
+
"epoch": 38.08,
|
1026 |
+
"learning_rate": 2.0259868421052632e-05,
|
1027 |
+
"loss": 1.0846,
|
1028 |
+
"step": 13100
|
1029 |
+
},
|
1030 |
+
{
|
1031 |
+
"epoch": 38.37,
|
1032 |
+
"learning_rate": 1.9766447368421053e-05,
|
1033 |
+
"loss": 1.0933,
|
1034 |
+
"step": 13200
|
1035 |
+
},
|
1036 |
+
{
|
1037 |
+
"epoch": 38.66,
|
1038 |
+
"learning_rate": 1.927302631578947e-05,
|
1039 |
+
"loss": 1.0661,
|
1040 |
+
"step": 13300
|
1041 |
+
},
|
1042 |
+
{
|
1043 |
+
"epoch": 38.95,
|
1044 |
+
"learning_rate": 1.8779605263157893e-05,
|
1045 |
+
"loss": 1.0902,
|
1046 |
+
"step": 13400
|
1047 |
+
},
|
1048 |
+
{
|
1049 |
+
"epoch": 39.24,
|
1050 |
+
"learning_rate": 1.8286184210526313e-05,
|
1051 |
+
"loss": 1.083,
|
1052 |
+
"step": 13500
|
1053 |
+
},
|
1054 |
+
{
|
1055 |
+
"epoch": 39.24,
|
1056 |
+
"eval_loss": 0.3181643486022949,
|
1057 |
+
"eval_runtime": 158.3527,
|
1058 |
+
"eval_samples_per_second": 29.175,
|
1059 |
+
"eval_steps_per_second": 3.65,
|
1060 |
+
"eval_wer": 0.2825674600746497,
|
1061 |
+
"step": 13500
|
1062 |
+
},
|
1063 |
+
{
|
1064 |
+
"epoch": 39.53,
|
1065 |
+
"learning_rate": 1.7792763157894736e-05,
|
1066 |
+
"loss": 1.0548,
|
1067 |
+
"step": 13600
|
1068 |
+
},
|
1069 |
+
{
|
1070 |
+
"epoch": 39.82,
|
1071 |
+
"learning_rate": 1.7299342105263156e-05,
|
1072 |
+
"loss": 1.0735,
|
1073 |
+
"step": 13700
|
1074 |
+
},
|
1075 |
+
{
|
1076 |
+
"epoch": 40.12,
|
1077 |
+
"learning_rate": 1.6805921052631577e-05,
|
1078 |
+
"loss": 1.085,
|
1079 |
+
"step": 13800
|
1080 |
+
},
|
1081 |
+
{
|
1082 |
+
"epoch": 40.41,
|
1083 |
+
"learning_rate": 1.63125e-05,
|
1084 |
+
"loss": 1.0531,
|
1085 |
+
"step": 13900
|
1086 |
+
},
|
1087 |
+
{
|
1088 |
+
"epoch": 40.7,
|
1089 |
+
"learning_rate": 1.581907894736842e-05,
|
1090 |
+
"loss": 1.0819,
|
1091 |
+
"step": 14000
|
1092 |
+
},
|
1093 |
+
{
|
1094 |
+
"epoch": 40.7,
|
1095 |
+
"eval_loss": 0.32643991708755493,
|
1096 |
+
"eval_runtime": 159.1099,
|
1097 |
+
"eval_samples_per_second": 29.037,
|
1098 |
+
"eval_steps_per_second": 3.633,
|
1099 |
+
"eval_wer": 0.28504558526586304,
|
1100 |
+
"step": 14000
|
1101 |
+
},
|
1102 |
+
{
|
1103 |
+
"epoch": 40.99,
|
1104 |
+
"learning_rate": 1.532565789473684e-05,
|
1105 |
+
"loss": 1.0679,
|
1106 |
+
"step": 14100
|
1107 |
+
},
|
1108 |
+
{
|
1109 |
+
"epoch": 41.28,
|
1110 |
+
"learning_rate": 1.4832236842105262e-05,
|
1111 |
+
"loss": 1.0773,
|
1112 |
+
"step": 14200
|
1113 |
+
},
|
1114 |
+
{
|
1115 |
+
"epoch": 41.57,
|
1116 |
+
"learning_rate": 1.4338815789473682e-05,
|
1117 |
+
"loss": 1.0587,
|
1118 |
+
"step": 14300
|
1119 |
+
},
|
1120 |
+
{
|
1121 |
+
"epoch": 41.86,
|
1122 |
+
"learning_rate": 1.3845394736842104e-05,
|
1123 |
+
"loss": 1.0287,
|
1124 |
+
"step": 14400
|
1125 |
+
},
|
1126 |
+
{
|
1127 |
+
"epoch": 42.15,
|
1128 |
+
"learning_rate": 1.3351973684210524e-05,
|
1129 |
+
"loss": 1.072,
|
1130 |
+
"step": 14500
|
1131 |
+
},
|
1132 |
+
{
|
1133 |
+
"epoch": 42.15,
|
1134 |
+
"eval_loss": 0.32792535424232483,
|
1135 |
+
"eval_runtime": 155.2521,
|
1136 |
+
"eval_samples_per_second": 29.758,
|
1137 |
+
"eval_steps_per_second": 3.723,
|
1138 |
+
"eval_wer": 0.2817108242060821,
|
1139 |
+
"step": 14500
|
1140 |
+
},
|
1141 |
+
{
|
1142 |
+
"epoch": 42.44,
|
1143 |
+
"learning_rate": 1.2863486842105262e-05,
|
1144 |
+
"loss": 1.0508,
|
1145 |
+
"step": 14600
|
1146 |
+
},
|
1147 |
+
{
|
1148 |
+
"epoch": 42.73,
|
1149 |
+
"learning_rate": 1.2370065789473684e-05,
|
1150 |
+
"loss": 1.0408,
|
1151 |
+
"step": 14700
|
1152 |
+
},
|
1153 |
+
{
|
1154 |
+
"epoch": 43.02,
|
1155 |
+
"learning_rate": 1.1876644736842105e-05,
|
1156 |
+
"loss": 1.0725,
|
1157 |
+
"step": 14800
|
1158 |
+
},
|
1159 |
+
{
|
1160 |
+
"epoch": 43.31,
|
1161 |
+
"learning_rate": 1.1383223684210525e-05,
|
1162 |
+
"loss": 1.0381,
|
1163 |
+
"step": 14900
|
1164 |
+
},
|
1165 |
+
{
|
1166 |
+
"epoch": 43.6,
|
1167 |
+
"learning_rate": 1.0889802631578946e-05,
|
1168 |
+
"loss": 1.0456,
|
1169 |
+
"step": 15000
|
1170 |
+
},
|
1171 |
+
{
|
1172 |
+
"epoch": 43.6,
|
1173 |
+
"eval_loss": 0.323445200920105,
|
1174 |
+
"eval_runtime": 137.3413,
|
1175 |
+
"eval_samples_per_second": 33.639,
|
1176 |
+
"eval_steps_per_second": 4.208,
|
1177 |
+
"eval_wer": 0.2792632931530319,
|
1178 |
+
"step": 15000
|
1179 |
+
},
|
1180 |
+
{
|
1181 |
+
"epoch": 43.89,
|
1182 |
+
"learning_rate": 1.0396381578947367e-05,
|
1183 |
+
"loss": 1.0655,
|
1184 |
+
"step": 15100
|
1185 |
+
},
|
1186 |
+
{
|
1187 |
+
"epoch": 44.19,
|
1188 |
+
"learning_rate": 9.90296052631579e-06,
|
1189 |
+
"loss": 1.0509,
|
1190 |
+
"step": 15200
|
1191 |
+
},
|
1192 |
+
{
|
1193 |
+
"epoch": 44.48,
|
1194 |
+
"learning_rate": 9.40953947368421e-06,
|
1195 |
+
"loss": 1.0357,
|
1196 |
+
"step": 15300
|
1197 |
+
},
|
1198 |
+
{
|
1199 |
+
"epoch": 44.77,
|
1200 |
+
"learning_rate": 8.916118421052631e-06,
|
1201 |
+
"loss": 1.0467,
|
1202 |
+
"step": 15400
|
1203 |
+
},
|
1204 |
+
{
|
1205 |
+
"epoch": 45.06,
|
1206 |
+
"learning_rate": 8.422697368421051e-06,
|
1207 |
+
"loss": 1.0581,
|
1208 |
+
"step": 15500
|
1209 |
+
},
|
1210 |
+
{
|
1211 |
+
"epoch": 45.06,
|
1212 |
+
"eval_loss": 0.32200726866722107,
|
1213 |
+
"eval_runtime": 138.9222,
|
1214 |
+
"eval_samples_per_second": 33.256,
|
1215 |
+
"eval_steps_per_second": 4.161,
|
1216 |
+
"eval_wer": 0.27791715107385423,
|
1217 |
+
"step": 15500
|
1218 |
+
},
|
1219 |
+
{
|
1220 |
+
"epoch": 45.35,
|
1221 |
+
"learning_rate": 7.929276315789473e-06,
|
1222 |
+
"loss": 1.0512,
|
1223 |
+
"step": 15600
|
1224 |
+
},
|
1225 |
+
{
|
1226 |
+
"epoch": 45.64,
|
1227 |
+
"learning_rate": 7.435855263157894e-06,
|
1228 |
+
"loss": 1.0397,
|
1229 |
+
"step": 15700
|
1230 |
+
},
|
1231 |
+
{
|
1232 |
+
"epoch": 45.93,
|
1233 |
+
"learning_rate": 6.9473684210526315e-06,
|
1234 |
+
"loss": 1.0455,
|
1235 |
+
"step": 15800
|
1236 |
+
},
|
1237 |
+
{
|
1238 |
+
"epoch": 46.22,
|
1239 |
+
"learning_rate": 6.453947368421052e-06,
|
1240 |
+
"loss": 1.0347,
|
1241 |
+
"step": 15900
|
1242 |
+
},
|
1243 |
+
{
|
1244 |
+
"epoch": 46.51,
|
1245 |
+
"learning_rate": 5.9605263157894735e-06,
|
1246 |
+
"loss": 1.0406,
|
1247 |
+
"step": 16000
|
1248 |
+
},
|
1249 |
+
{
|
1250 |
+
"epoch": 46.51,
|
1251 |
+
"eval_loss": 0.32076749205589294,
|
1252 |
+
"eval_runtime": 134.5334,
|
1253 |
+
"eval_samples_per_second": 34.341,
|
1254 |
+
"eval_steps_per_second": 4.296,
|
1255 |
+
"eval_wer": 0.2762344734748822,
|
1256 |
+
"step": 16000
|
1257 |
+
},
|
1258 |
+
{
|
1259 |
+
"epoch": 46.8,
|
1260 |
+
"learning_rate": 5.467105263157894e-06,
|
1261 |
+
"loss": 1.0528,
|
1262 |
+
"step": 16100
|
1263 |
+
},
|
1264 |
+
{
|
1265 |
+
"epoch": 47.09,
|
1266 |
+
"learning_rate": 4.973684210526316e-06,
|
1267 |
+
"loss": 1.045,
|
1268 |
+
"step": 16200
|
1269 |
+
},
|
1270 |
+
{
|
1271 |
+
"epoch": 47.38,
|
1272 |
+
"learning_rate": 4.480263157894736e-06,
|
1273 |
+
"loss": 1.0129,
|
1274 |
+
"step": 16300
|
1275 |
+
},
|
1276 |
+
{
|
1277 |
+
"epoch": 47.67,
|
1278 |
+
"learning_rate": 3.986842105263157e-06,
|
1279 |
+
"loss": 1.0102,
|
1280 |
+
"step": 16400
|
1281 |
+
},
|
1282 |
+
{
|
1283 |
+
"epoch": 47.96,
|
1284 |
+
"learning_rate": 3.4934210526315787e-06,
|
1285 |
+
"loss": 1.0422,
|
1286 |
+
"step": 16500
|
1287 |
+
},
|
1288 |
+
{
|
1289 |
+
"epoch": 47.96,
|
1290 |
+
"eval_loss": 0.3183736503124237,
|
1291 |
+
"eval_runtime": 137.2442,
|
1292 |
+
"eval_samples_per_second": 33.663,
|
1293 |
+
"eval_steps_per_second": 4.211,
|
1294 |
+
"eval_wer": 0.2751942727773359,
|
1295 |
+
"step": 16500
|
1296 |
+
},
|
1297 |
+
{
|
1298 |
+
"epoch": 48.26,
|
1299 |
+
"learning_rate": 2.9999999999999997e-06,
|
1300 |
+
"loss": 1.0323,
|
1301 |
+
"step": 16600
|
1302 |
+
},
|
1303 |
+
{
|
1304 |
+
"epoch": 48.55,
|
1305 |
+
"learning_rate": 2.506578947368421e-06,
|
1306 |
+
"loss": 1.0292,
|
1307 |
+
"step": 16700
|
1308 |
+
},
|
1309 |
+
{
|
1310 |
+
"epoch": 48.84,
|
1311 |
+
"learning_rate": 2.013157894736842e-06,
|
1312 |
+
"loss": 1.0378,
|
1313 |
+
"step": 16800
|
1314 |
+
},
|
1315 |
+
{
|
1316 |
+
"epoch": 49.13,
|
1317 |
+
"learning_rate": 1.519736842105263e-06,
|
1318 |
+
"loss": 1.035,
|
1319 |
+
"step": 16900
|
1320 |
+
},
|
1321 |
+
{
|
1322 |
+
"epoch": 49.42,
|
1323 |
+
"learning_rate": 1.0263157894736841e-06,
|
1324 |
+
"loss": 1.0099,
|
1325 |
+
"step": 17000
|
1326 |
+
},
|
1327 |
+
{
|
1328 |
+
"epoch": 49.42,
|
1329 |
+
"eval_loss": 0.31810811161994934,
|
1330 |
+
"eval_runtime": 134.533,
|
1331 |
+
"eval_samples_per_second": 34.341,
|
1332 |
+
"eval_steps_per_second": 4.296,
|
1333 |
+
"eval_wer": 0.2734810010402007,
|
1334 |
+
"step": 17000
|
1335 |
+
},
|
1336 |
+
{
|
1337 |
+
"epoch": 49.71,
|
1338 |
+
"learning_rate": 5.328947368421052e-07,
|
1339 |
+
"loss": 1.0338,
|
1340 |
+
"step": 17100
|
1341 |
+
},
|
1342 |
+
{
|
1343 |
+
"epoch": 50.0,
|
1344 |
+
"learning_rate": 3.947368421052631e-08,
|
1345 |
+
"loss": 1.03,
|
1346 |
+
"step": 17200
|
1347 |
+
},
|
1348 |
+
{
|
1349 |
+
"epoch": 50.0,
|
1350 |
+
"step": 17200,
|
1351 |
+
"total_flos": 5.074861809449127e+19,
|
1352 |
+
"train_loss": 1.4555730460410894,
|
1353 |
+
"train_runtime": 29339.4334,
|
1354 |
+
"train_samples_per_second": 18.797,
|
1355 |
+
"train_steps_per_second": 0.586
|
1356 |
}
|
1357 |
],
|
1358 |
+
"max_steps": 17200,
|
1359 |
+
"num_train_epochs": 50,
|
1360 |
+
"total_flos": 5.074861809449127e+19,
|
1361 |
"trial_name": null,
|
1362 |
"trial_params": null
|
1363 |
}
|