Commit ·
8f173c0
1
Parent(s): 7bab2fb
Copied from Google Drive
Browse files- checkpoint-1000/config.json +42 -0
- checkpoint-1000/optimizer.pt +3 -0
- checkpoint-1000/preprocessor_config.json +0 -0
- checkpoint-1000/pytorch_model.bin +3 -0
- checkpoint-1000/rng_state.pth +3 -0
- checkpoint-1000/scaler.pt +3 -0
- checkpoint-1000/scheduler.pt +3 -0
- checkpoint-1000/trainer_state.json +265 -0
- checkpoint-1000/training_args.bin +3 -0
- checkpoint-2000/config.json +42 -0
- checkpoint-2000/optimizer.pt +3 -0
- checkpoint-2000/preprocessor_config.json +0 -0
- checkpoint-2000/pytorch_model.bin +3 -0
- checkpoint-2000/rng_state.pth +3 -0
- checkpoint-2000/scaler.pt +3 -0
- checkpoint-2000/scheduler.pt +3 -0
- checkpoint-2000/trainer_state.json +514 -0
- checkpoint-2000/training_args.bin +3 -0
- checkpoint-3000/config.json +42 -0
- checkpoint-3000/optimizer.pt +3 -0
- checkpoint-3000/preprocessor_config.json +0 -0
- checkpoint-3000/pytorch_model.bin +3 -0
- checkpoint-3000/rng_state.pth +3 -0
- checkpoint-3000/scaler.pt +3 -0
- checkpoint-3000/scheduler.pt +3 -0
- checkpoint-3000/trainer_state.json +763 -0
- checkpoint-3000/training_args.bin +3 -0
- checkpoint-4000/config.json +42 -0
- checkpoint-4000/optimizer.pt +3 -0
- checkpoint-4000/preprocessor_config.json +0 -0
- checkpoint-4000/pytorch_model.bin +3 -0
- checkpoint-4000/rng_state.pth +3 -0
- checkpoint-4000/scaler.pt +3 -0
- checkpoint-4000/scheduler.pt +3 -0
- checkpoint-4000/trainer_state.json +1012 -0
- checkpoint-4000/training_args.bin +3 -0
checkpoint-1000/config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "openai/whisper-tiny",
|
| 3 |
+
"activation_dropout": 0.0,
|
| 4 |
+
"activation_function": "gelu",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"WhisperForConditionalGeneration"
|
| 7 |
+
],
|
| 8 |
+
"attention_dropout": 0.0,
|
| 9 |
+
"begin_suppress_tokens": [
|
| 10 |
+
220,
|
| 11 |
+
50257
|
| 12 |
+
],
|
| 13 |
+
"bos_token_id": 50257,
|
| 14 |
+
"d_model": 384,
|
| 15 |
+
"decoder_attention_heads": 8,
|
| 16 |
+
"decoder_ffn_dim": 1536,
|
| 17 |
+
"decoder_layerdrop": 0.0,
|
| 18 |
+
"decoder_layers": 4,
|
| 19 |
+
"decoder_start_token_id": 50258,
|
| 20 |
+
"dropout": 0.0,
|
| 21 |
+
"encoder_attention_heads": 8,
|
| 22 |
+
"encoder_ffn_dim": 1536,
|
| 23 |
+
"encoder_layerdrop": 0.0,
|
| 24 |
+
"encoder_layers": 4,
|
| 25 |
+
"eos_token_id": 50257,
|
| 26 |
+
"forced_decoder_ids": null,
|
| 27 |
+
"init_std": 0.02,
|
| 28 |
+
"is_encoder_decoder": true,
|
| 29 |
+
"max_length": 448,
|
| 30 |
+
"max_source_positions": 1500,
|
| 31 |
+
"max_target_positions": 448,
|
| 32 |
+
"model_type": "whisper",
|
| 33 |
+
"num_hidden_layers": 4,
|
| 34 |
+
"num_mel_bins": 80,
|
| 35 |
+
"pad_token_id": 50257,
|
| 36 |
+
"scale_embedding": false,
|
| 37 |
+
"suppress_tokens": [],
|
| 38 |
+
"torch_dtype": "float32",
|
| 39 |
+
"transformers_version": "4.26.0.dev0",
|
| 40 |
+
"use_cache": false,
|
| 41 |
+
"vocab_size": 51865
|
| 42 |
+
}
|
checkpoint-1000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eb07e156a0ddf11d6e46f8bac8a486b68a198d76804407a3786611914a70f89f
|
| 3 |
+
size 302181925
|
checkpoint-1000/preprocessor_config.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-1000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ddfdb2074a2bea75797d98df4bfeda206815b76cb3da84151b788f7b9102b759
|
| 3 |
+
size 151097331
|
checkpoint-1000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:09fe5b3dc696a6166acee045746eb26fd72da66482b5937bce83d0523a2a2bc8
|
| 3 |
+
size 14503
|
checkpoint-1000/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:daae308645d159c91abc2157feb4f3a1445aeddeca2c9aceb077ac427964e138
|
| 3 |
+
size 559
|
checkpoint-1000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dba7c66c33b2d7ae78e5feba71d5825b9e394072c9adfd7097871538a4cb84a4
|
| 3 |
+
size 623
|
checkpoint-1000/trainer_state.json
ADDED
|
@@ -0,0 +1,265 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 644.20410427066,
|
| 3 |
+
"best_model_checkpoint": "./whisper-tiny-it-7/checkpoint-1000",
|
| 4 |
+
"epoch": 0.9541984732824428,
|
| 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": 0.02,
|
| 12 |
+
"learning_rate": 4e-08,
|
| 13 |
+
"loss": 8.8525,
|
| 14 |
+
"step": 25
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.05,
|
| 18 |
+
"learning_rate": 9e-08,
|
| 19 |
+
"loss": 8.857,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.07,
|
| 24 |
+
"learning_rate": 1.4e-07,
|
| 25 |
+
"loss": 8.6107,
|
| 26 |
+
"step": 75
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.1,
|
| 30 |
+
"learning_rate": 1.8999999999999998e-07,
|
| 31 |
+
"loss": 8.1659,
|
| 32 |
+
"step": 100
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.12,
|
| 36 |
+
"learning_rate": 2.4e-07,
|
| 37 |
+
"loss": 7.7359,
|
| 38 |
+
"step": 125
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.14,
|
| 42 |
+
"learning_rate": 2.9e-07,
|
| 43 |
+
"loss": 7.2816,
|
| 44 |
+
"step": 150
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.17,
|
| 48 |
+
"learning_rate": 3.4000000000000003e-07,
|
| 49 |
+
"loss": 6.9829,
|
| 50 |
+
"step": 175
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.19,
|
| 54 |
+
"learning_rate": 3.8999999999999997e-07,
|
| 55 |
+
"loss": 6.5257,
|
| 56 |
+
"step": 200
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.21,
|
| 60 |
+
"learning_rate": 4.3999999999999997e-07,
|
| 61 |
+
"loss": 6.1594,
|
| 62 |
+
"step": 225
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.24,
|
| 66 |
+
"learning_rate": 4.9e-07,
|
| 67 |
+
"loss": 5.9187,
|
| 68 |
+
"step": 250
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.26,
|
| 72 |
+
"learning_rate": 5.4e-07,
|
| 73 |
+
"loss": 5.7123,
|
| 74 |
+
"step": 275
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.29,
|
| 78 |
+
"learning_rate": 5.9e-07,
|
| 79 |
+
"loss": 5.4899,
|
| 80 |
+
"step": 300
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.31,
|
| 84 |
+
"learning_rate": 6.4e-07,
|
| 85 |
+
"loss": 5.1635,
|
| 86 |
+
"step": 325
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.33,
|
| 90 |
+
"learning_rate": 6.9e-07,
|
| 91 |
+
"loss": 4.9477,
|
| 92 |
+
"step": 350
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.36,
|
| 96 |
+
"learning_rate": 7.4e-07,
|
| 97 |
+
"loss": 4.8882,
|
| 98 |
+
"step": 375
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.38,
|
| 102 |
+
"learning_rate": 7.9e-07,
|
| 103 |
+
"loss": 4.7571,
|
| 104 |
+
"step": 400
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.41,
|
| 108 |
+
"learning_rate": 8.399999999999999e-07,
|
| 109 |
+
"loss": 4.7402,
|
| 110 |
+
"step": 425
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.43,
|
| 114 |
+
"learning_rate": 8.9e-07,
|
| 115 |
+
"loss": 4.6719,
|
| 116 |
+
"step": 450
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.45,
|
| 120 |
+
"learning_rate": 9.399999999999999e-07,
|
| 121 |
+
"loss": 4.5627,
|
| 122 |
+
"step": 475
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.48,
|
| 126 |
+
"learning_rate": 9.9e-07,
|
| 127 |
+
"loss": 4.6078,
|
| 128 |
+
"step": 500
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.5,
|
| 132 |
+
"learning_rate": 9.942857142857144e-07,
|
| 133 |
+
"loss": 4.5315,
|
| 134 |
+
"step": 525
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.52,
|
| 138 |
+
"learning_rate": 9.871428571428572e-07,
|
| 139 |
+
"loss": 4.4441,
|
| 140 |
+
"step": 550
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.55,
|
| 144 |
+
"learning_rate": 9.8e-07,
|
| 145 |
+
"loss": 4.4877,
|
| 146 |
+
"step": 575
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.57,
|
| 150 |
+
"learning_rate": 9.72857142857143e-07,
|
| 151 |
+
"loss": 4.4972,
|
| 152 |
+
"step": 600
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.6,
|
| 156 |
+
"learning_rate": 9.657142857142857e-07,
|
| 157 |
+
"loss": 4.4206,
|
| 158 |
+
"step": 625
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.62,
|
| 162 |
+
"learning_rate": 9.585714285714285e-07,
|
| 163 |
+
"loss": 4.4121,
|
| 164 |
+
"step": 650
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.64,
|
| 168 |
+
"learning_rate": 9.514285714285713e-07,
|
| 169 |
+
"loss": 4.3059,
|
| 170 |
+
"step": 675
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.67,
|
| 174 |
+
"learning_rate": 9.442857142857142e-07,
|
| 175 |
+
"loss": 4.3611,
|
| 176 |
+
"step": 700
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.69,
|
| 180 |
+
"learning_rate": 9.371428571428571e-07,
|
| 181 |
+
"loss": 4.343,
|
| 182 |
+
"step": 725
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.72,
|
| 186 |
+
"learning_rate": 9.3e-07,
|
| 187 |
+
"loss": 4.3366,
|
| 188 |
+
"step": 750
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.74,
|
| 192 |
+
"learning_rate": 9.228571428571428e-07,
|
| 193 |
+
"loss": 4.3313,
|
| 194 |
+
"step": 775
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.76,
|
| 198 |
+
"learning_rate": 9.157142857142856e-07,
|
| 199 |
+
"loss": 4.317,
|
| 200 |
+
"step": 800
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.79,
|
| 204 |
+
"learning_rate": 9.085714285714285e-07,
|
| 205 |
+
"loss": 4.2746,
|
| 206 |
+
"step": 825
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.81,
|
| 210 |
+
"learning_rate": 9.014285714285714e-07,
|
| 211 |
+
"loss": 4.2817,
|
| 212 |
+
"step": 850
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.83,
|
| 216 |
+
"learning_rate": 8.942857142857142e-07,
|
| 217 |
+
"loss": 4.2529,
|
| 218 |
+
"step": 875
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.86,
|
| 222 |
+
"learning_rate": 8.871428571428571e-07,
|
| 223 |
+
"loss": 4.248,
|
| 224 |
+
"step": 900
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.88,
|
| 228 |
+
"learning_rate": 8.799999999999999e-07,
|
| 229 |
+
"loss": 4.2731,
|
| 230 |
+
"step": 925
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.91,
|
| 234 |
+
"learning_rate": 8.728571428571428e-07,
|
| 235 |
+
"loss": 4.2709,
|
| 236 |
+
"step": 950
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.93,
|
| 240 |
+
"learning_rate": 8.657142857142857e-07,
|
| 241 |
+
"loss": 4.2254,
|
| 242 |
+
"step": 975
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.95,
|
| 246 |
+
"learning_rate": 8.585714285714285e-07,
|
| 247 |
+
"loss": 4.204,
|
| 248 |
+
"step": 1000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.95,
|
| 252 |
+
"eval_loss": 4.17011833190918,
|
| 253 |
+
"eval_runtime": 1604.0964,
|
| 254 |
+
"eval_samples_per_second": 0.935,
|
| 255 |
+
"eval_steps_per_second": 0.117,
|
| 256 |
+
"eval_wer": 644.20410427066,
|
| 257 |
+
"step": 1000
|
| 258 |
+
}
|
| 259 |
+
],
|
| 260 |
+
"max_steps": 4000,
|
| 261 |
+
"num_train_epochs": 4,
|
| 262 |
+
"total_flos": 3.9390216192e+17,
|
| 263 |
+
"trial_name": null,
|
| 264 |
+
"trial_params": null
|
| 265 |
+
}
|
checkpoint-1000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bef1dc6124409479b93eb5f081a655debb2a0dea8030612856467bf2b8c0655b
|
| 3 |
+
size 3503
|
checkpoint-2000/config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "openai/whisper-tiny",
|
| 3 |
+
"activation_dropout": 0.0,
|
| 4 |
+
"activation_function": "gelu",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"WhisperForConditionalGeneration"
|
| 7 |
+
],
|
| 8 |
+
"attention_dropout": 0.0,
|
| 9 |
+
"begin_suppress_tokens": [
|
| 10 |
+
220,
|
| 11 |
+
50257
|
| 12 |
+
],
|
| 13 |
+
"bos_token_id": 50257,
|
| 14 |
+
"d_model": 384,
|
| 15 |
+
"decoder_attention_heads": 8,
|
| 16 |
+
"decoder_ffn_dim": 1536,
|
| 17 |
+
"decoder_layerdrop": 0.0,
|
| 18 |
+
"decoder_layers": 4,
|
| 19 |
+
"decoder_start_token_id": 50258,
|
| 20 |
+
"dropout": 0.0,
|
| 21 |
+
"encoder_attention_heads": 8,
|
| 22 |
+
"encoder_ffn_dim": 1536,
|
| 23 |
+
"encoder_layerdrop": 0.0,
|
| 24 |
+
"encoder_layers": 4,
|
| 25 |
+
"eos_token_id": 50257,
|
| 26 |
+
"forced_decoder_ids": null,
|
| 27 |
+
"init_std": 0.02,
|
| 28 |
+
"is_encoder_decoder": true,
|
| 29 |
+
"max_length": 448,
|
| 30 |
+
"max_source_positions": 1500,
|
| 31 |
+
"max_target_positions": 448,
|
| 32 |
+
"model_type": "whisper",
|
| 33 |
+
"num_hidden_layers": 4,
|
| 34 |
+
"num_mel_bins": 80,
|
| 35 |
+
"pad_token_id": 50257,
|
| 36 |
+
"scale_embedding": false,
|
| 37 |
+
"suppress_tokens": [],
|
| 38 |
+
"torch_dtype": "float32",
|
| 39 |
+
"transformers_version": "4.26.0.dev0",
|
| 40 |
+
"use_cache": false,
|
| 41 |
+
"vocab_size": 51865
|
| 42 |
+
}
|
checkpoint-2000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2e3c15d25fe5c7133e45a35c06be20de9e38f175f0e21146e28a4e334307e52d
|
| 3 |
+
size 302181925
|
checkpoint-2000/preprocessor_config.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-2000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8ab2db4049809d534335bbd067148b413312ac3bbc68e4d1223042e367f893a1
|
| 3 |
+
size 151097331
|
checkpoint-2000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:502b85493d6890e92fba64ab91a9ca82398737312b5b53407f70d09b9b0670e6
|
| 3 |
+
size 14503
|
checkpoint-2000/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:55950e8917904a21fc5d9fb13f05004c45267242b39ae8b98076a6e1f6177765
|
| 3 |
+
size 559
|
checkpoint-2000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e95661e70bf11906828528ff91c0a61cfddce7436b7a777a781b337d628aa3d3
|
| 3 |
+
size 623
|
checkpoint-2000/trainer_state.json
ADDED
|
@@ -0,0 +1,514 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 160.23294509151413,
|
| 3 |
+
"best_model_checkpoint": "./whisper-tiny-it-7/checkpoint-2000",
|
| 4 |
+
"epoch": 1.9083969465648853,
|
| 5 |
+
"global_step": 2000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.02,
|
| 12 |
+
"learning_rate": 4e-08,
|
| 13 |
+
"loss": 8.8525,
|
| 14 |
+
"step": 25
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.05,
|
| 18 |
+
"learning_rate": 9e-08,
|
| 19 |
+
"loss": 8.857,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.07,
|
| 24 |
+
"learning_rate": 1.4e-07,
|
| 25 |
+
"loss": 8.6107,
|
| 26 |
+
"step": 75
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.1,
|
| 30 |
+
"learning_rate": 1.8999999999999998e-07,
|
| 31 |
+
"loss": 8.1659,
|
| 32 |
+
"step": 100
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.12,
|
| 36 |
+
"learning_rate": 2.4e-07,
|
| 37 |
+
"loss": 7.7359,
|
| 38 |
+
"step": 125
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.14,
|
| 42 |
+
"learning_rate": 2.9e-07,
|
| 43 |
+
"loss": 7.2816,
|
| 44 |
+
"step": 150
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.17,
|
| 48 |
+
"learning_rate": 3.4000000000000003e-07,
|
| 49 |
+
"loss": 6.9829,
|
| 50 |
+
"step": 175
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.19,
|
| 54 |
+
"learning_rate": 3.8999999999999997e-07,
|
| 55 |
+
"loss": 6.5257,
|
| 56 |
+
"step": 200
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.21,
|
| 60 |
+
"learning_rate": 4.3999999999999997e-07,
|
| 61 |
+
"loss": 6.1594,
|
| 62 |
+
"step": 225
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.24,
|
| 66 |
+
"learning_rate": 4.9e-07,
|
| 67 |
+
"loss": 5.9187,
|
| 68 |
+
"step": 250
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.26,
|
| 72 |
+
"learning_rate": 5.4e-07,
|
| 73 |
+
"loss": 5.7123,
|
| 74 |
+
"step": 275
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.29,
|
| 78 |
+
"learning_rate": 5.9e-07,
|
| 79 |
+
"loss": 5.4899,
|
| 80 |
+
"step": 300
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.31,
|
| 84 |
+
"learning_rate": 6.4e-07,
|
| 85 |
+
"loss": 5.1635,
|
| 86 |
+
"step": 325
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.33,
|
| 90 |
+
"learning_rate": 6.9e-07,
|
| 91 |
+
"loss": 4.9477,
|
| 92 |
+
"step": 350
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.36,
|
| 96 |
+
"learning_rate": 7.4e-07,
|
| 97 |
+
"loss": 4.8882,
|
| 98 |
+
"step": 375
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.38,
|
| 102 |
+
"learning_rate": 7.9e-07,
|
| 103 |
+
"loss": 4.7571,
|
| 104 |
+
"step": 400
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.41,
|
| 108 |
+
"learning_rate": 8.399999999999999e-07,
|
| 109 |
+
"loss": 4.7402,
|
| 110 |
+
"step": 425
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.43,
|
| 114 |
+
"learning_rate": 8.9e-07,
|
| 115 |
+
"loss": 4.6719,
|
| 116 |
+
"step": 450
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.45,
|
| 120 |
+
"learning_rate": 9.399999999999999e-07,
|
| 121 |
+
"loss": 4.5627,
|
| 122 |
+
"step": 475
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.48,
|
| 126 |
+
"learning_rate": 9.9e-07,
|
| 127 |
+
"loss": 4.6078,
|
| 128 |
+
"step": 500
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.5,
|
| 132 |
+
"learning_rate": 9.942857142857144e-07,
|
| 133 |
+
"loss": 4.5315,
|
| 134 |
+
"step": 525
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.52,
|
| 138 |
+
"learning_rate": 9.871428571428572e-07,
|
| 139 |
+
"loss": 4.4441,
|
| 140 |
+
"step": 550
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.55,
|
| 144 |
+
"learning_rate": 9.8e-07,
|
| 145 |
+
"loss": 4.4877,
|
| 146 |
+
"step": 575
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.57,
|
| 150 |
+
"learning_rate": 9.72857142857143e-07,
|
| 151 |
+
"loss": 4.4972,
|
| 152 |
+
"step": 600
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.6,
|
| 156 |
+
"learning_rate": 9.657142857142857e-07,
|
| 157 |
+
"loss": 4.4206,
|
| 158 |
+
"step": 625
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.62,
|
| 162 |
+
"learning_rate": 9.585714285714285e-07,
|
| 163 |
+
"loss": 4.4121,
|
| 164 |
+
"step": 650
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.64,
|
| 168 |
+
"learning_rate": 9.514285714285713e-07,
|
| 169 |
+
"loss": 4.3059,
|
| 170 |
+
"step": 675
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.67,
|
| 174 |
+
"learning_rate": 9.442857142857142e-07,
|
| 175 |
+
"loss": 4.3611,
|
| 176 |
+
"step": 700
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.69,
|
| 180 |
+
"learning_rate": 9.371428571428571e-07,
|
| 181 |
+
"loss": 4.343,
|
| 182 |
+
"step": 725
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.72,
|
| 186 |
+
"learning_rate": 9.3e-07,
|
| 187 |
+
"loss": 4.3366,
|
| 188 |
+
"step": 750
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.74,
|
| 192 |
+
"learning_rate": 9.228571428571428e-07,
|
| 193 |
+
"loss": 4.3313,
|
| 194 |
+
"step": 775
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.76,
|
| 198 |
+
"learning_rate": 9.157142857142856e-07,
|
| 199 |
+
"loss": 4.317,
|
| 200 |
+
"step": 800
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.79,
|
| 204 |
+
"learning_rate": 9.085714285714285e-07,
|
| 205 |
+
"loss": 4.2746,
|
| 206 |
+
"step": 825
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.81,
|
| 210 |
+
"learning_rate": 9.014285714285714e-07,
|
| 211 |
+
"loss": 4.2817,
|
| 212 |
+
"step": 850
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.83,
|
| 216 |
+
"learning_rate": 8.942857142857142e-07,
|
| 217 |
+
"loss": 4.2529,
|
| 218 |
+
"step": 875
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.86,
|
| 222 |
+
"learning_rate": 8.871428571428571e-07,
|
| 223 |
+
"loss": 4.248,
|
| 224 |
+
"step": 900
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.88,
|
| 228 |
+
"learning_rate": 8.799999999999999e-07,
|
| 229 |
+
"loss": 4.2731,
|
| 230 |
+
"step": 925
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.91,
|
| 234 |
+
"learning_rate": 8.728571428571428e-07,
|
| 235 |
+
"loss": 4.2709,
|
| 236 |
+
"step": 950
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.93,
|
| 240 |
+
"learning_rate": 8.657142857142857e-07,
|
| 241 |
+
"loss": 4.2254,
|
| 242 |
+
"step": 975
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.95,
|
| 246 |
+
"learning_rate": 8.585714285714285e-07,
|
| 247 |
+
"loss": 4.204,
|
| 248 |
+
"step": 1000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.95,
|
| 252 |
+
"eval_loss": 4.17011833190918,
|
| 253 |
+
"eval_runtime": 1604.0964,
|
| 254 |
+
"eval_samples_per_second": 0.935,
|
| 255 |
+
"eval_steps_per_second": 0.117,
|
| 256 |
+
"eval_wer": 644.20410427066,
|
| 257 |
+
"step": 1000
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 0.98,
|
| 261 |
+
"learning_rate": 8.514285714285714e-07,
|
| 262 |
+
"loss": 4.2357,
|
| 263 |
+
"step": 1025
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 1.0,
|
| 267 |
+
"learning_rate": 8.442857142857142e-07,
|
| 268 |
+
"loss": 4.1855,
|
| 269 |
+
"step": 1050
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 1.03,
|
| 273 |
+
"learning_rate": 8.371428571428571e-07,
|
| 274 |
+
"loss": 4.1227,
|
| 275 |
+
"step": 1075
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 1.05,
|
| 279 |
+
"learning_rate": 8.299999999999999e-07,
|
| 280 |
+
"loss": 4.1665,
|
| 281 |
+
"step": 1100
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 1.07,
|
| 285 |
+
"learning_rate": 8.228571428571428e-07,
|
| 286 |
+
"loss": 4.1772,
|
| 287 |
+
"step": 1125
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 1.1,
|
| 291 |
+
"learning_rate": 8.157142857142856e-07,
|
| 292 |
+
"loss": 4.1451,
|
| 293 |
+
"step": 1150
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 1.12,
|
| 297 |
+
"learning_rate": 8.085714285714285e-07,
|
| 298 |
+
"loss": 4.0833,
|
| 299 |
+
"step": 1175
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 1.15,
|
| 303 |
+
"learning_rate": 8.014285714285713e-07,
|
| 304 |
+
"loss": 4.1437,
|
| 305 |
+
"step": 1200
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 1.17,
|
| 309 |
+
"learning_rate": 7.942857142857142e-07,
|
| 310 |
+
"loss": 4.1286,
|
| 311 |
+
"step": 1225
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 1.19,
|
| 315 |
+
"learning_rate": 7.871428571428571e-07,
|
| 316 |
+
"loss": 4.1125,
|
| 317 |
+
"step": 1250
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 1.22,
|
| 321 |
+
"learning_rate": 7.799999999999999e-07,
|
| 322 |
+
"loss": 4.0927,
|
| 323 |
+
"step": 1275
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 1.24,
|
| 327 |
+
"learning_rate": 7.728571428571428e-07,
|
| 328 |
+
"loss": 4.1013,
|
| 329 |
+
"step": 1300
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 1.26,
|
| 333 |
+
"learning_rate": 7.657142857142856e-07,
|
| 334 |
+
"loss": 4.0453,
|
| 335 |
+
"step": 1325
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 1.29,
|
| 339 |
+
"learning_rate": 7.585714285714285e-07,
|
| 340 |
+
"loss": 4.1317,
|
| 341 |
+
"step": 1350
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 1.31,
|
| 345 |
+
"learning_rate": 7.514285714285715e-07,
|
| 346 |
+
"loss": 4.0638,
|
| 347 |
+
"step": 1375
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 1.34,
|
| 351 |
+
"learning_rate": 7.442857142857143e-07,
|
| 352 |
+
"loss": 4.1043,
|
| 353 |
+
"step": 1400
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 1.36,
|
| 357 |
+
"learning_rate": 7.371428571428571e-07,
|
| 358 |
+
"loss": 4.0761,
|
| 359 |
+
"step": 1425
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 1.38,
|
| 363 |
+
"learning_rate": 7.3e-07,
|
| 364 |
+
"loss": 3.9902,
|
| 365 |
+
"step": 1450
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 1.41,
|
| 369 |
+
"learning_rate": 7.228571428571429e-07,
|
| 370 |
+
"loss": 4.0019,
|
| 371 |
+
"step": 1475
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 1.43,
|
| 375 |
+
"learning_rate": 7.157142857142858e-07,
|
| 376 |
+
"loss": 4.0683,
|
| 377 |
+
"step": 1500
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 1.46,
|
| 381 |
+
"learning_rate": 7.085714285714285e-07,
|
| 382 |
+
"loss": 3.9523,
|
| 383 |
+
"step": 1525
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 1.48,
|
| 387 |
+
"learning_rate": 7.014285714285714e-07,
|
| 388 |
+
"loss": 3.9421,
|
| 389 |
+
"step": 1550
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 1.5,
|
| 393 |
+
"learning_rate": 6.942857142857143e-07,
|
| 394 |
+
"loss": 3.7698,
|
| 395 |
+
"step": 1575
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 1.53,
|
| 399 |
+
"learning_rate": 6.871428571428572e-07,
|
| 400 |
+
"loss": 3.7565,
|
| 401 |
+
"step": 1600
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 1.55,
|
| 405 |
+
"learning_rate": 6.800000000000001e-07,
|
| 406 |
+
"loss": 3.6593,
|
| 407 |
+
"step": 1625
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 1.57,
|
| 411 |
+
"learning_rate": 6.728571428571428e-07,
|
| 412 |
+
"loss": 3.4785,
|
| 413 |
+
"step": 1650
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 1.6,
|
| 417 |
+
"learning_rate": 6.657142857142857e-07,
|
| 418 |
+
"loss": 3.4676,
|
| 419 |
+
"step": 1675
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 1.62,
|
| 423 |
+
"learning_rate": 6.585714285714286e-07,
|
| 424 |
+
"loss": 3.3216,
|
| 425 |
+
"step": 1700
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 1.65,
|
| 429 |
+
"learning_rate": 6.514285714285715e-07,
|
| 430 |
+
"loss": 3.2336,
|
| 431 |
+
"step": 1725
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 1.67,
|
| 435 |
+
"learning_rate": 6.442857142857142e-07,
|
| 436 |
+
"loss": 3.1206,
|
| 437 |
+
"step": 1750
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 1.69,
|
| 441 |
+
"learning_rate": 6.371428571428571e-07,
|
| 442 |
+
"loss": 3.0662,
|
| 443 |
+
"step": 1775
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 1.72,
|
| 447 |
+
"learning_rate": 6.3e-07,
|
| 448 |
+
"loss": 2.972,
|
| 449 |
+
"step": 1800
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 1.74,
|
| 453 |
+
"learning_rate": 6.228571428571429e-07,
|
| 454 |
+
"loss": 2.8847,
|
| 455 |
+
"step": 1825
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 1.77,
|
| 459 |
+
"learning_rate": 6.157142857142858e-07,
|
| 460 |
+
"loss": 2.8173,
|
| 461 |
+
"step": 1850
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 1.79,
|
| 465 |
+
"learning_rate": 6.085714285714285e-07,
|
| 466 |
+
"loss": 2.7043,
|
| 467 |
+
"step": 1875
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 1.81,
|
| 471 |
+
"learning_rate": 6.014285714285714e-07,
|
| 472 |
+
"loss": 2.7055,
|
| 473 |
+
"step": 1900
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 1.84,
|
| 477 |
+
"learning_rate": 5.942857142857143e-07,
|
| 478 |
+
"loss": 2.7277,
|
| 479 |
+
"step": 1925
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 1.86,
|
| 483 |
+
"learning_rate": 5.871428571428572e-07,
|
| 484 |
+
"loss": 2.6155,
|
| 485 |
+
"step": 1950
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 1.88,
|
| 489 |
+
"learning_rate": 5.8e-07,
|
| 490 |
+
"loss": 2.533,
|
| 491 |
+
"step": 1975
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 1.91,
|
| 495 |
+
"learning_rate": 5.728571428571428e-07,
|
| 496 |
+
"loss": 2.4631,
|
| 497 |
+
"step": 2000
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 1.91,
|
| 501 |
+
"eval_loss": 2.7998619079589844,
|
| 502 |
+
"eval_runtime": 865.3409,
|
| 503 |
+
"eval_samples_per_second": 1.733,
|
| 504 |
+
"eval_steps_per_second": 0.217,
|
| 505 |
+
"eval_wer": 160.23294509151413,
|
| 506 |
+
"step": 2000
|
| 507 |
+
}
|
| 508 |
+
],
|
| 509 |
+
"max_steps": 4000,
|
| 510 |
+
"num_train_epochs": 4,
|
| 511 |
+
"total_flos": 7.8763199164416e+17,
|
| 512 |
+
"trial_name": null,
|
| 513 |
+
"trial_params": null
|
| 514 |
+
}
|
checkpoint-2000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bef1dc6124409479b93eb5f081a655debb2a0dea8030612856467bf2b8c0655b
|
| 3 |
+
size 3503
|
checkpoint-3000/config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "openai/whisper-tiny",
|
| 3 |
+
"activation_dropout": 0.0,
|
| 4 |
+
"activation_function": "gelu",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"WhisperForConditionalGeneration"
|
| 7 |
+
],
|
| 8 |
+
"attention_dropout": 0.0,
|
| 9 |
+
"begin_suppress_tokens": [
|
| 10 |
+
220,
|
| 11 |
+
50257
|
| 12 |
+
],
|
| 13 |
+
"bos_token_id": 50257,
|
| 14 |
+
"d_model": 384,
|
| 15 |
+
"decoder_attention_heads": 8,
|
| 16 |
+
"decoder_ffn_dim": 1536,
|
| 17 |
+
"decoder_layerdrop": 0.0,
|
| 18 |
+
"decoder_layers": 4,
|
| 19 |
+
"decoder_start_token_id": 50258,
|
| 20 |
+
"dropout": 0.0,
|
| 21 |
+
"encoder_attention_heads": 8,
|
| 22 |
+
"encoder_ffn_dim": 1536,
|
| 23 |
+
"encoder_layerdrop": 0.0,
|
| 24 |
+
"encoder_layers": 4,
|
| 25 |
+
"eos_token_id": 50257,
|
| 26 |
+
"forced_decoder_ids": null,
|
| 27 |
+
"init_std": 0.02,
|
| 28 |
+
"is_encoder_decoder": true,
|
| 29 |
+
"max_length": 448,
|
| 30 |
+
"max_source_positions": 1500,
|
| 31 |
+
"max_target_positions": 448,
|
| 32 |
+
"model_type": "whisper",
|
| 33 |
+
"num_hidden_layers": 4,
|
| 34 |
+
"num_mel_bins": 80,
|
| 35 |
+
"pad_token_id": 50257,
|
| 36 |
+
"scale_embedding": false,
|
| 37 |
+
"suppress_tokens": [],
|
| 38 |
+
"torch_dtype": "float32",
|
| 39 |
+
"transformers_version": "4.26.0.dev0",
|
| 40 |
+
"use_cache": false,
|
| 41 |
+
"vocab_size": 51865
|
| 42 |
+
}
|
checkpoint-3000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c8ee53c2f6c48bacbbcbfae6efe5ae8162f7126c14f5320539cb7a635e259db4
|
| 3 |
+
size 302181925
|
checkpoint-3000/preprocessor_config.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-3000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e808d2818378b62ca2123d88aaaa59521a23ba282738ce79ffefd4c505236dd3
|
| 3 |
+
size 151097331
|
checkpoint-3000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8675a5cbeeab7063b1173e0ed6380233c24f95ee110cfe9c65a75d5ad3bffa65
|
| 3 |
+
size 14503
|
checkpoint-3000/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6c1d33dc95e779d3b74ba8c487374b98679f5268cf1a246a10c7de96feaf43e7
|
| 3 |
+
size 559
|
checkpoint-3000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c772c5b30f625fadf9210cd825fdd8d70e783dc4f6cb4f19bd8c6324af0774a4
|
| 3 |
+
size 623
|
checkpoint-3000/trainer_state.json
ADDED
|
@@ -0,0 +1,763 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 103.72296173044926,
|
| 3 |
+
"best_model_checkpoint": "./whisper-tiny-it-7/checkpoint-3000",
|
| 4 |
+
"epoch": 2.8625954198473282,
|
| 5 |
+
"global_step": 3000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.02,
|
| 12 |
+
"learning_rate": 4e-08,
|
| 13 |
+
"loss": 8.8525,
|
| 14 |
+
"step": 25
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.05,
|
| 18 |
+
"learning_rate": 9e-08,
|
| 19 |
+
"loss": 8.857,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.07,
|
| 24 |
+
"learning_rate": 1.4e-07,
|
| 25 |
+
"loss": 8.6107,
|
| 26 |
+
"step": 75
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.1,
|
| 30 |
+
"learning_rate": 1.8999999999999998e-07,
|
| 31 |
+
"loss": 8.1659,
|
| 32 |
+
"step": 100
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.12,
|
| 36 |
+
"learning_rate": 2.4e-07,
|
| 37 |
+
"loss": 7.7359,
|
| 38 |
+
"step": 125
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.14,
|
| 42 |
+
"learning_rate": 2.9e-07,
|
| 43 |
+
"loss": 7.2816,
|
| 44 |
+
"step": 150
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.17,
|
| 48 |
+
"learning_rate": 3.4000000000000003e-07,
|
| 49 |
+
"loss": 6.9829,
|
| 50 |
+
"step": 175
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.19,
|
| 54 |
+
"learning_rate": 3.8999999999999997e-07,
|
| 55 |
+
"loss": 6.5257,
|
| 56 |
+
"step": 200
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.21,
|
| 60 |
+
"learning_rate": 4.3999999999999997e-07,
|
| 61 |
+
"loss": 6.1594,
|
| 62 |
+
"step": 225
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.24,
|
| 66 |
+
"learning_rate": 4.9e-07,
|
| 67 |
+
"loss": 5.9187,
|
| 68 |
+
"step": 250
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.26,
|
| 72 |
+
"learning_rate": 5.4e-07,
|
| 73 |
+
"loss": 5.7123,
|
| 74 |
+
"step": 275
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.29,
|
| 78 |
+
"learning_rate": 5.9e-07,
|
| 79 |
+
"loss": 5.4899,
|
| 80 |
+
"step": 300
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.31,
|
| 84 |
+
"learning_rate": 6.4e-07,
|
| 85 |
+
"loss": 5.1635,
|
| 86 |
+
"step": 325
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.33,
|
| 90 |
+
"learning_rate": 6.9e-07,
|
| 91 |
+
"loss": 4.9477,
|
| 92 |
+
"step": 350
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.36,
|
| 96 |
+
"learning_rate": 7.4e-07,
|
| 97 |
+
"loss": 4.8882,
|
| 98 |
+
"step": 375
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.38,
|
| 102 |
+
"learning_rate": 7.9e-07,
|
| 103 |
+
"loss": 4.7571,
|
| 104 |
+
"step": 400
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.41,
|
| 108 |
+
"learning_rate": 8.399999999999999e-07,
|
| 109 |
+
"loss": 4.7402,
|
| 110 |
+
"step": 425
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.43,
|
| 114 |
+
"learning_rate": 8.9e-07,
|
| 115 |
+
"loss": 4.6719,
|
| 116 |
+
"step": 450
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.45,
|
| 120 |
+
"learning_rate": 9.399999999999999e-07,
|
| 121 |
+
"loss": 4.5627,
|
| 122 |
+
"step": 475
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.48,
|
| 126 |
+
"learning_rate": 9.9e-07,
|
| 127 |
+
"loss": 4.6078,
|
| 128 |
+
"step": 500
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.5,
|
| 132 |
+
"learning_rate": 9.942857142857144e-07,
|
| 133 |
+
"loss": 4.5315,
|
| 134 |
+
"step": 525
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.52,
|
| 138 |
+
"learning_rate": 9.871428571428572e-07,
|
| 139 |
+
"loss": 4.4441,
|
| 140 |
+
"step": 550
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.55,
|
| 144 |
+
"learning_rate": 9.8e-07,
|
| 145 |
+
"loss": 4.4877,
|
| 146 |
+
"step": 575
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.57,
|
| 150 |
+
"learning_rate": 9.72857142857143e-07,
|
| 151 |
+
"loss": 4.4972,
|
| 152 |
+
"step": 600
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.6,
|
| 156 |
+
"learning_rate": 9.657142857142857e-07,
|
| 157 |
+
"loss": 4.4206,
|
| 158 |
+
"step": 625
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.62,
|
| 162 |
+
"learning_rate": 9.585714285714285e-07,
|
| 163 |
+
"loss": 4.4121,
|
| 164 |
+
"step": 650
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.64,
|
| 168 |
+
"learning_rate": 9.514285714285713e-07,
|
| 169 |
+
"loss": 4.3059,
|
| 170 |
+
"step": 675
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.67,
|
| 174 |
+
"learning_rate": 9.442857142857142e-07,
|
| 175 |
+
"loss": 4.3611,
|
| 176 |
+
"step": 700
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.69,
|
| 180 |
+
"learning_rate": 9.371428571428571e-07,
|
| 181 |
+
"loss": 4.343,
|
| 182 |
+
"step": 725
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.72,
|
| 186 |
+
"learning_rate": 9.3e-07,
|
| 187 |
+
"loss": 4.3366,
|
| 188 |
+
"step": 750
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.74,
|
| 192 |
+
"learning_rate": 9.228571428571428e-07,
|
| 193 |
+
"loss": 4.3313,
|
| 194 |
+
"step": 775
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.76,
|
| 198 |
+
"learning_rate": 9.157142857142856e-07,
|
| 199 |
+
"loss": 4.317,
|
| 200 |
+
"step": 800
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.79,
|
| 204 |
+
"learning_rate": 9.085714285714285e-07,
|
| 205 |
+
"loss": 4.2746,
|
| 206 |
+
"step": 825
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.81,
|
| 210 |
+
"learning_rate": 9.014285714285714e-07,
|
| 211 |
+
"loss": 4.2817,
|
| 212 |
+
"step": 850
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.83,
|
| 216 |
+
"learning_rate": 8.942857142857142e-07,
|
| 217 |
+
"loss": 4.2529,
|
| 218 |
+
"step": 875
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.86,
|
| 222 |
+
"learning_rate": 8.871428571428571e-07,
|
| 223 |
+
"loss": 4.248,
|
| 224 |
+
"step": 900
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.88,
|
| 228 |
+
"learning_rate": 8.799999999999999e-07,
|
| 229 |
+
"loss": 4.2731,
|
| 230 |
+
"step": 925
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.91,
|
| 234 |
+
"learning_rate": 8.728571428571428e-07,
|
| 235 |
+
"loss": 4.2709,
|
| 236 |
+
"step": 950
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.93,
|
| 240 |
+
"learning_rate": 8.657142857142857e-07,
|
| 241 |
+
"loss": 4.2254,
|
| 242 |
+
"step": 975
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.95,
|
| 246 |
+
"learning_rate": 8.585714285714285e-07,
|
| 247 |
+
"loss": 4.204,
|
| 248 |
+
"step": 1000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.95,
|
| 252 |
+
"eval_loss": 4.17011833190918,
|
| 253 |
+
"eval_runtime": 1604.0964,
|
| 254 |
+
"eval_samples_per_second": 0.935,
|
| 255 |
+
"eval_steps_per_second": 0.117,
|
| 256 |
+
"eval_wer": 644.20410427066,
|
| 257 |
+
"step": 1000
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 0.98,
|
| 261 |
+
"learning_rate": 8.514285714285714e-07,
|
| 262 |
+
"loss": 4.2357,
|
| 263 |
+
"step": 1025
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 1.0,
|
| 267 |
+
"learning_rate": 8.442857142857142e-07,
|
| 268 |
+
"loss": 4.1855,
|
| 269 |
+
"step": 1050
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 1.03,
|
| 273 |
+
"learning_rate": 8.371428571428571e-07,
|
| 274 |
+
"loss": 4.1227,
|
| 275 |
+
"step": 1075
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 1.05,
|
| 279 |
+
"learning_rate": 8.299999999999999e-07,
|
| 280 |
+
"loss": 4.1665,
|
| 281 |
+
"step": 1100
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 1.07,
|
| 285 |
+
"learning_rate": 8.228571428571428e-07,
|
| 286 |
+
"loss": 4.1772,
|
| 287 |
+
"step": 1125
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 1.1,
|
| 291 |
+
"learning_rate": 8.157142857142856e-07,
|
| 292 |
+
"loss": 4.1451,
|
| 293 |
+
"step": 1150
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 1.12,
|
| 297 |
+
"learning_rate": 8.085714285714285e-07,
|
| 298 |
+
"loss": 4.0833,
|
| 299 |
+
"step": 1175
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 1.15,
|
| 303 |
+
"learning_rate": 8.014285714285713e-07,
|
| 304 |
+
"loss": 4.1437,
|
| 305 |
+
"step": 1200
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 1.17,
|
| 309 |
+
"learning_rate": 7.942857142857142e-07,
|
| 310 |
+
"loss": 4.1286,
|
| 311 |
+
"step": 1225
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 1.19,
|
| 315 |
+
"learning_rate": 7.871428571428571e-07,
|
| 316 |
+
"loss": 4.1125,
|
| 317 |
+
"step": 1250
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 1.22,
|
| 321 |
+
"learning_rate": 7.799999999999999e-07,
|
| 322 |
+
"loss": 4.0927,
|
| 323 |
+
"step": 1275
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 1.24,
|
| 327 |
+
"learning_rate": 7.728571428571428e-07,
|
| 328 |
+
"loss": 4.1013,
|
| 329 |
+
"step": 1300
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 1.26,
|
| 333 |
+
"learning_rate": 7.657142857142856e-07,
|
| 334 |
+
"loss": 4.0453,
|
| 335 |
+
"step": 1325
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 1.29,
|
| 339 |
+
"learning_rate": 7.585714285714285e-07,
|
| 340 |
+
"loss": 4.1317,
|
| 341 |
+
"step": 1350
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 1.31,
|
| 345 |
+
"learning_rate": 7.514285714285715e-07,
|
| 346 |
+
"loss": 4.0638,
|
| 347 |
+
"step": 1375
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 1.34,
|
| 351 |
+
"learning_rate": 7.442857142857143e-07,
|
| 352 |
+
"loss": 4.1043,
|
| 353 |
+
"step": 1400
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 1.36,
|
| 357 |
+
"learning_rate": 7.371428571428571e-07,
|
| 358 |
+
"loss": 4.0761,
|
| 359 |
+
"step": 1425
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 1.38,
|
| 363 |
+
"learning_rate": 7.3e-07,
|
| 364 |
+
"loss": 3.9902,
|
| 365 |
+
"step": 1450
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 1.41,
|
| 369 |
+
"learning_rate": 7.228571428571429e-07,
|
| 370 |
+
"loss": 4.0019,
|
| 371 |
+
"step": 1475
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 1.43,
|
| 375 |
+
"learning_rate": 7.157142857142858e-07,
|
| 376 |
+
"loss": 4.0683,
|
| 377 |
+
"step": 1500
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 1.46,
|
| 381 |
+
"learning_rate": 7.085714285714285e-07,
|
| 382 |
+
"loss": 3.9523,
|
| 383 |
+
"step": 1525
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 1.48,
|
| 387 |
+
"learning_rate": 7.014285714285714e-07,
|
| 388 |
+
"loss": 3.9421,
|
| 389 |
+
"step": 1550
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 1.5,
|
| 393 |
+
"learning_rate": 6.942857142857143e-07,
|
| 394 |
+
"loss": 3.7698,
|
| 395 |
+
"step": 1575
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 1.53,
|
| 399 |
+
"learning_rate": 6.871428571428572e-07,
|
| 400 |
+
"loss": 3.7565,
|
| 401 |
+
"step": 1600
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 1.55,
|
| 405 |
+
"learning_rate": 6.800000000000001e-07,
|
| 406 |
+
"loss": 3.6593,
|
| 407 |
+
"step": 1625
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 1.57,
|
| 411 |
+
"learning_rate": 6.728571428571428e-07,
|
| 412 |
+
"loss": 3.4785,
|
| 413 |
+
"step": 1650
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 1.6,
|
| 417 |
+
"learning_rate": 6.657142857142857e-07,
|
| 418 |
+
"loss": 3.4676,
|
| 419 |
+
"step": 1675
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 1.62,
|
| 423 |
+
"learning_rate": 6.585714285714286e-07,
|
| 424 |
+
"loss": 3.3216,
|
| 425 |
+
"step": 1700
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 1.65,
|
| 429 |
+
"learning_rate": 6.514285714285715e-07,
|
| 430 |
+
"loss": 3.2336,
|
| 431 |
+
"step": 1725
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 1.67,
|
| 435 |
+
"learning_rate": 6.442857142857142e-07,
|
| 436 |
+
"loss": 3.1206,
|
| 437 |
+
"step": 1750
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 1.69,
|
| 441 |
+
"learning_rate": 6.371428571428571e-07,
|
| 442 |
+
"loss": 3.0662,
|
| 443 |
+
"step": 1775
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 1.72,
|
| 447 |
+
"learning_rate": 6.3e-07,
|
| 448 |
+
"loss": 2.972,
|
| 449 |
+
"step": 1800
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 1.74,
|
| 453 |
+
"learning_rate": 6.228571428571429e-07,
|
| 454 |
+
"loss": 2.8847,
|
| 455 |
+
"step": 1825
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 1.77,
|
| 459 |
+
"learning_rate": 6.157142857142858e-07,
|
| 460 |
+
"loss": 2.8173,
|
| 461 |
+
"step": 1850
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 1.79,
|
| 465 |
+
"learning_rate": 6.085714285714285e-07,
|
| 466 |
+
"loss": 2.7043,
|
| 467 |
+
"step": 1875
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 1.81,
|
| 471 |
+
"learning_rate": 6.014285714285714e-07,
|
| 472 |
+
"loss": 2.7055,
|
| 473 |
+
"step": 1900
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 1.84,
|
| 477 |
+
"learning_rate": 5.942857142857143e-07,
|
| 478 |
+
"loss": 2.7277,
|
| 479 |
+
"step": 1925
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 1.86,
|
| 483 |
+
"learning_rate": 5.871428571428572e-07,
|
| 484 |
+
"loss": 2.6155,
|
| 485 |
+
"step": 1950
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 1.88,
|
| 489 |
+
"learning_rate": 5.8e-07,
|
| 490 |
+
"loss": 2.533,
|
| 491 |
+
"step": 1975
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 1.91,
|
| 495 |
+
"learning_rate": 5.728571428571428e-07,
|
| 496 |
+
"loss": 2.4631,
|
| 497 |
+
"step": 2000
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 1.91,
|
| 501 |
+
"eval_loss": 2.7998619079589844,
|
| 502 |
+
"eval_runtime": 865.3409,
|
| 503 |
+
"eval_samples_per_second": 1.733,
|
| 504 |
+
"eval_steps_per_second": 0.217,
|
| 505 |
+
"eval_wer": 160.23294509151413,
|
| 506 |
+
"step": 2000
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 1.93,
|
| 510 |
+
"learning_rate": 5.657142857142857e-07,
|
| 511 |
+
"loss": 2.4711,
|
| 512 |
+
"step": 2025
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 1.96,
|
| 516 |
+
"learning_rate": 5.585714285714286e-07,
|
| 517 |
+
"loss": 2.3802,
|
| 518 |
+
"step": 2050
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 1.98,
|
| 522 |
+
"learning_rate": 5.514285714285714e-07,
|
| 523 |
+
"loss": 2.4496,
|
| 524 |
+
"step": 2075
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 2.0,
|
| 528 |
+
"learning_rate": 5.442857142857142e-07,
|
| 529 |
+
"loss": 2.3227,
|
| 530 |
+
"step": 2100
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 2.03,
|
| 534 |
+
"learning_rate": 5.371428571428571e-07,
|
| 535 |
+
"loss": 2.2921,
|
| 536 |
+
"step": 2125
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 2.05,
|
| 540 |
+
"learning_rate": 5.3e-07,
|
| 541 |
+
"loss": 2.2983,
|
| 542 |
+
"step": 2150
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 2.08,
|
| 546 |
+
"learning_rate": 5.228571428571429e-07,
|
| 547 |
+
"loss": 2.279,
|
| 548 |
+
"step": 2175
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 2.1,
|
| 552 |
+
"learning_rate": 5.157142857142857e-07,
|
| 553 |
+
"loss": 2.298,
|
| 554 |
+
"step": 2200
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 2.12,
|
| 558 |
+
"learning_rate": 5.085714285714285e-07,
|
| 559 |
+
"loss": 2.2966,
|
| 560 |
+
"step": 2225
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 2.15,
|
| 564 |
+
"learning_rate": 5.014285714285714e-07,
|
| 565 |
+
"loss": 2.2024,
|
| 566 |
+
"step": 2250
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 2.17,
|
| 570 |
+
"learning_rate": 4.942857142857143e-07,
|
| 571 |
+
"loss": 2.2506,
|
| 572 |
+
"step": 2275
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 2.19,
|
| 576 |
+
"learning_rate": 4.871428571428572e-07,
|
| 577 |
+
"loss": 2.1831,
|
| 578 |
+
"step": 2300
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 2.22,
|
| 582 |
+
"learning_rate": 4.8e-07,
|
| 583 |
+
"loss": 2.1786,
|
| 584 |
+
"step": 2325
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 2.24,
|
| 588 |
+
"learning_rate": 4.7285714285714284e-07,
|
| 589 |
+
"loss": 2.1795,
|
| 590 |
+
"step": 2350
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 2.27,
|
| 594 |
+
"learning_rate": 4.6571428571428566e-07,
|
| 595 |
+
"loss": 2.1797,
|
| 596 |
+
"step": 2375
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 2.29,
|
| 600 |
+
"learning_rate": 4.5857142857142854e-07,
|
| 601 |
+
"loss": 2.1381,
|
| 602 |
+
"step": 2400
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 2.31,
|
| 606 |
+
"learning_rate": 4.5142857142857137e-07,
|
| 607 |
+
"loss": 2.0612,
|
| 608 |
+
"step": 2425
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 2.34,
|
| 612 |
+
"learning_rate": 4.4428571428571425e-07,
|
| 613 |
+
"loss": 2.0958,
|
| 614 |
+
"step": 2450
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 2.36,
|
| 618 |
+
"learning_rate": 4.3714285714285713e-07,
|
| 619 |
+
"loss": 2.0547,
|
| 620 |
+
"step": 2475
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 2.39,
|
| 624 |
+
"learning_rate": 4.2999999999999996e-07,
|
| 625 |
+
"loss": 2.0996,
|
| 626 |
+
"step": 2500
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 2.41,
|
| 630 |
+
"learning_rate": 4.2285714285714284e-07,
|
| 631 |
+
"loss": 2.0421,
|
| 632 |
+
"step": 2525
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 2.43,
|
| 636 |
+
"learning_rate": 4.1571428571428567e-07,
|
| 637 |
+
"loss": 1.9812,
|
| 638 |
+
"step": 2550
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 2.46,
|
| 642 |
+
"learning_rate": 4.0857142857142855e-07,
|
| 643 |
+
"loss": 2.0337,
|
| 644 |
+
"step": 2575
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 2.48,
|
| 648 |
+
"learning_rate": 4.014285714285714e-07,
|
| 649 |
+
"loss": 2.0123,
|
| 650 |
+
"step": 2600
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 2.5,
|
| 654 |
+
"learning_rate": 3.9428571428571425e-07,
|
| 655 |
+
"loss": 2.0056,
|
| 656 |
+
"step": 2625
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 2.53,
|
| 660 |
+
"learning_rate": 3.871428571428571e-07,
|
| 661 |
+
"loss": 1.9827,
|
| 662 |
+
"step": 2650
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 2.55,
|
| 666 |
+
"learning_rate": 3.7999999999999996e-07,
|
| 667 |
+
"loss": 2.0326,
|
| 668 |
+
"step": 2675
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 2.58,
|
| 672 |
+
"learning_rate": 3.728571428571428e-07,
|
| 673 |
+
"loss": 1.9789,
|
| 674 |
+
"step": 2700
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 2.6,
|
| 678 |
+
"learning_rate": 3.657142857142857e-07,
|
| 679 |
+
"loss": 1.9974,
|
| 680 |
+
"step": 2725
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 2.62,
|
| 684 |
+
"learning_rate": 3.585714285714286e-07,
|
| 685 |
+
"loss": 1.9701,
|
| 686 |
+
"step": 2750
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 2.65,
|
| 690 |
+
"learning_rate": 3.5142857142857143e-07,
|
| 691 |
+
"loss": 1.9797,
|
| 692 |
+
"step": 2775
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 2.67,
|
| 696 |
+
"learning_rate": 3.442857142857143e-07,
|
| 697 |
+
"loss": 1.93,
|
| 698 |
+
"step": 2800
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 2.7,
|
| 702 |
+
"learning_rate": 3.3714285714285714e-07,
|
| 703 |
+
"loss": 1.9068,
|
| 704 |
+
"step": 2825
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 2.72,
|
| 708 |
+
"learning_rate": 3.3e-07,
|
| 709 |
+
"loss": 1.9442,
|
| 710 |
+
"step": 2850
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 2.74,
|
| 714 |
+
"learning_rate": 3.2285714285714284e-07,
|
| 715 |
+
"loss": 1.8534,
|
| 716 |
+
"step": 2875
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 2.77,
|
| 720 |
+
"learning_rate": 3.157142857142857e-07,
|
| 721 |
+
"loss": 1.9212,
|
| 722 |
+
"step": 2900
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 2.79,
|
| 726 |
+
"learning_rate": 3.0857142857142855e-07,
|
| 727 |
+
"loss": 1.9611,
|
| 728 |
+
"step": 2925
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 2.81,
|
| 732 |
+
"learning_rate": 3.0142857142857143e-07,
|
| 733 |
+
"loss": 1.9114,
|
| 734 |
+
"step": 2950
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 2.84,
|
| 738 |
+
"learning_rate": 2.9428571428571426e-07,
|
| 739 |
+
"loss": 1.8668,
|
| 740 |
+
"step": 2975
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 2.86,
|
| 744 |
+
"learning_rate": 2.8714285714285714e-07,
|
| 745 |
+
"loss": 1.8798,
|
| 746 |
+
"step": 3000
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 2.86,
|
| 750 |
+
"eval_loss": 2.237186908721924,
|
| 751 |
+
"eval_runtime": 535.0618,
|
| 752 |
+
"eval_samples_per_second": 2.803,
|
| 753 |
+
"eval_steps_per_second": 0.351,
|
| 754 |
+
"eval_wer": 103.72296173044926,
|
| 755 |
+
"step": 3000
|
| 756 |
+
}
|
| 757 |
+
],
|
| 758 |
+
"max_steps": 4000,
|
| 759 |
+
"num_train_epochs": 4,
|
| 760 |
+
"total_flos": 1.18136182136832e+18,
|
| 761 |
+
"trial_name": null,
|
| 762 |
+
"trial_params": null
|
| 763 |
+
}
|
checkpoint-3000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bef1dc6124409479b93eb5f081a655debb2a0dea8030612856467bf2b8c0655b
|
| 3 |
+
size 3503
|
checkpoint-4000/config.json
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "whisper-tiny-it-7/checkpoint-3000",
|
| 3 |
+
"activation_dropout": 0.0,
|
| 4 |
+
"activation_function": "gelu",
|
| 5 |
+
"architectures": [
|
| 6 |
+
"WhisperForConditionalGeneration"
|
| 7 |
+
],
|
| 8 |
+
"attention_dropout": 0.0,
|
| 9 |
+
"begin_suppress_tokens": [
|
| 10 |
+
220,
|
| 11 |
+
50257
|
| 12 |
+
],
|
| 13 |
+
"bos_token_id": 50257,
|
| 14 |
+
"d_model": 384,
|
| 15 |
+
"decoder_attention_heads": 8,
|
| 16 |
+
"decoder_ffn_dim": 1536,
|
| 17 |
+
"decoder_layerdrop": 0.0,
|
| 18 |
+
"decoder_layers": 4,
|
| 19 |
+
"decoder_start_token_id": 50258,
|
| 20 |
+
"dropout": 0.0,
|
| 21 |
+
"encoder_attention_heads": 8,
|
| 22 |
+
"encoder_ffn_dim": 1536,
|
| 23 |
+
"encoder_layerdrop": 0.0,
|
| 24 |
+
"encoder_layers": 4,
|
| 25 |
+
"eos_token_id": 50257,
|
| 26 |
+
"forced_decoder_ids": null,
|
| 27 |
+
"init_std": 0.02,
|
| 28 |
+
"is_encoder_decoder": true,
|
| 29 |
+
"max_length": 448,
|
| 30 |
+
"max_source_positions": 1500,
|
| 31 |
+
"max_target_positions": 448,
|
| 32 |
+
"model_type": "whisper",
|
| 33 |
+
"num_hidden_layers": 4,
|
| 34 |
+
"num_mel_bins": 80,
|
| 35 |
+
"pad_token_id": 50257,
|
| 36 |
+
"scale_embedding": false,
|
| 37 |
+
"suppress_tokens": [],
|
| 38 |
+
"torch_dtype": "float32",
|
| 39 |
+
"transformers_version": "4.26.0.dev0",
|
| 40 |
+
"use_cache": false,
|
| 41 |
+
"vocab_size": 51865
|
| 42 |
+
}
|
checkpoint-4000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:77b2a19baff235e768806af31d34444fee263daa4423f777225bebce5e97a1a0
|
| 3 |
+
size 302183173
|
checkpoint-4000/preprocessor_config.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-4000/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c9be5e9df9f00bfdd5d138f78f9f789612269c98b5bb8da962b555529f3bd8ca
|
| 3 |
+
size 151098921
|
checkpoint-4000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:28b59d44c59de9feec2281dbd6b65684c9ea536796d865ea82d0fe435a511610
|
| 3 |
+
size 14511
|
checkpoint-4000/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:544e7b5a8fc630539d2540fd05210195724f8aea448af1dfbc846116394ed633
|
| 3 |
+
size 557
|
checkpoint-4000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8b84f8d892c12e39431236e737b4a7b41e29b6f5528f8aff1d79f4858fbbee65
|
| 3 |
+
size 627
|
checkpoint-4000/trainer_state.json
ADDED
|
@@ -0,0 +1,1012 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 97.56655574043262,
|
| 3 |
+
"best_model_checkpoint": "./whisper-tiny-it-7/checkpoint-4000",
|
| 4 |
+
"epoch": 3.816793893129771,
|
| 5 |
+
"global_step": 4000,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.02,
|
| 12 |
+
"learning_rate": 4e-08,
|
| 13 |
+
"loss": 8.8525,
|
| 14 |
+
"step": 25
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.05,
|
| 18 |
+
"learning_rate": 9e-08,
|
| 19 |
+
"loss": 8.857,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.07,
|
| 24 |
+
"learning_rate": 1.4e-07,
|
| 25 |
+
"loss": 8.6107,
|
| 26 |
+
"step": 75
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.1,
|
| 30 |
+
"learning_rate": 1.8999999999999998e-07,
|
| 31 |
+
"loss": 8.1659,
|
| 32 |
+
"step": 100
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.12,
|
| 36 |
+
"learning_rate": 2.4e-07,
|
| 37 |
+
"loss": 7.7359,
|
| 38 |
+
"step": 125
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.14,
|
| 42 |
+
"learning_rate": 2.9e-07,
|
| 43 |
+
"loss": 7.2816,
|
| 44 |
+
"step": 150
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.17,
|
| 48 |
+
"learning_rate": 3.4000000000000003e-07,
|
| 49 |
+
"loss": 6.9829,
|
| 50 |
+
"step": 175
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.19,
|
| 54 |
+
"learning_rate": 3.8999999999999997e-07,
|
| 55 |
+
"loss": 6.5257,
|
| 56 |
+
"step": 200
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.21,
|
| 60 |
+
"learning_rate": 4.3999999999999997e-07,
|
| 61 |
+
"loss": 6.1594,
|
| 62 |
+
"step": 225
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.24,
|
| 66 |
+
"learning_rate": 4.9e-07,
|
| 67 |
+
"loss": 5.9187,
|
| 68 |
+
"step": 250
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.26,
|
| 72 |
+
"learning_rate": 5.4e-07,
|
| 73 |
+
"loss": 5.7123,
|
| 74 |
+
"step": 275
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.29,
|
| 78 |
+
"learning_rate": 5.9e-07,
|
| 79 |
+
"loss": 5.4899,
|
| 80 |
+
"step": 300
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.31,
|
| 84 |
+
"learning_rate": 6.4e-07,
|
| 85 |
+
"loss": 5.1635,
|
| 86 |
+
"step": 325
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.33,
|
| 90 |
+
"learning_rate": 6.9e-07,
|
| 91 |
+
"loss": 4.9477,
|
| 92 |
+
"step": 350
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.36,
|
| 96 |
+
"learning_rate": 7.4e-07,
|
| 97 |
+
"loss": 4.8882,
|
| 98 |
+
"step": 375
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.38,
|
| 102 |
+
"learning_rate": 7.9e-07,
|
| 103 |
+
"loss": 4.7571,
|
| 104 |
+
"step": 400
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.41,
|
| 108 |
+
"learning_rate": 8.399999999999999e-07,
|
| 109 |
+
"loss": 4.7402,
|
| 110 |
+
"step": 425
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.43,
|
| 114 |
+
"learning_rate": 8.9e-07,
|
| 115 |
+
"loss": 4.6719,
|
| 116 |
+
"step": 450
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.45,
|
| 120 |
+
"learning_rate": 9.399999999999999e-07,
|
| 121 |
+
"loss": 4.5627,
|
| 122 |
+
"step": 475
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.48,
|
| 126 |
+
"learning_rate": 9.9e-07,
|
| 127 |
+
"loss": 4.6078,
|
| 128 |
+
"step": 500
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.5,
|
| 132 |
+
"learning_rate": 9.942857142857144e-07,
|
| 133 |
+
"loss": 4.5315,
|
| 134 |
+
"step": 525
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.52,
|
| 138 |
+
"learning_rate": 9.871428571428572e-07,
|
| 139 |
+
"loss": 4.4441,
|
| 140 |
+
"step": 550
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.55,
|
| 144 |
+
"learning_rate": 9.8e-07,
|
| 145 |
+
"loss": 4.4877,
|
| 146 |
+
"step": 575
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.57,
|
| 150 |
+
"learning_rate": 9.72857142857143e-07,
|
| 151 |
+
"loss": 4.4972,
|
| 152 |
+
"step": 600
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.6,
|
| 156 |
+
"learning_rate": 9.657142857142857e-07,
|
| 157 |
+
"loss": 4.4206,
|
| 158 |
+
"step": 625
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.62,
|
| 162 |
+
"learning_rate": 9.585714285714285e-07,
|
| 163 |
+
"loss": 4.4121,
|
| 164 |
+
"step": 650
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.64,
|
| 168 |
+
"learning_rate": 9.514285714285713e-07,
|
| 169 |
+
"loss": 4.3059,
|
| 170 |
+
"step": 675
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.67,
|
| 174 |
+
"learning_rate": 9.442857142857142e-07,
|
| 175 |
+
"loss": 4.3611,
|
| 176 |
+
"step": 700
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.69,
|
| 180 |
+
"learning_rate": 9.371428571428571e-07,
|
| 181 |
+
"loss": 4.343,
|
| 182 |
+
"step": 725
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.72,
|
| 186 |
+
"learning_rate": 9.3e-07,
|
| 187 |
+
"loss": 4.3366,
|
| 188 |
+
"step": 750
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.74,
|
| 192 |
+
"learning_rate": 9.228571428571428e-07,
|
| 193 |
+
"loss": 4.3313,
|
| 194 |
+
"step": 775
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.76,
|
| 198 |
+
"learning_rate": 9.157142857142856e-07,
|
| 199 |
+
"loss": 4.317,
|
| 200 |
+
"step": 800
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.79,
|
| 204 |
+
"learning_rate": 9.085714285714285e-07,
|
| 205 |
+
"loss": 4.2746,
|
| 206 |
+
"step": 825
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.81,
|
| 210 |
+
"learning_rate": 9.014285714285714e-07,
|
| 211 |
+
"loss": 4.2817,
|
| 212 |
+
"step": 850
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.83,
|
| 216 |
+
"learning_rate": 8.942857142857142e-07,
|
| 217 |
+
"loss": 4.2529,
|
| 218 |
+
"step": 875
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.86,
|
| 222 |
+
"learning_rate": 8.871428571428571e-07,
|
| 223 |
+
"loss": 4.248,
|
| 224 |
+
"step": 900
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.88,
|
| 228 |
+
"learning_rate": 8.799999999999999e-07,
|
| 229 |
+
"loss": 4.2731,
|
| 230 |
+
"step": 925
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.91,
|
| 234 |
+
"learning_rate": 8.728571428571428e-07,
|
| 235 |
+
"loss": 4.2709,
|
| 236 |
+
"step": 950
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.93,
|
| 240 |
+
"learning_rate": 8.657142857142857e-07,
|
| 241 |
+
"loss": 4.2254,
|
| 242 |
+
"step": 975
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.95,
|
| 246 |
+
"learning_rate": 8.585714285714285e-07,
|
| 247 |
+
"loss": 4.204,
|
| 248 |
+
"step": 1000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.95,
|
| 252 |
+
"eval_loss": 4.17011833190918,
|
| 253 |
+
"eval_runtime": 1604.0964,
|
| 254 |
+
"eval_samples_per_second": 0.935,
|
| 255 |
+
"eval_steps_per_second": 0.117,
|
| 256 |
+
"eval_wer": 644.20410427066,
|
| 257 |
+
"step": 1000
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 0.98,
|
| 261 |
+
"learning_rate": 8.514285714285714e-07,
|
| 262 |
+
"loss": 4.2357,
|
| 263 |
+
"step": 1025
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 1.0,
|
| 267 |
+
"learning_rate": 8.442857142857142e-07,
|
| 268 |
+
"loss": 4.1855,
|
| 269 |
+
"step": 1050
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 1.03,
|
| 273 |
+
"learning_rate": 8.371428571428571e-07,
|
| 274 |
+
"loss": 4.1227,
|
| 275 |
+
"step": 1075
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 1.05,
|
| 279 |
+
"learning_rate": 8.299999999999999e-07,
|
| 280 |
+
"loss": 4.1665,
|
| 281 |
+
"step": 1100
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 1.07,
|
| 285 |
+
"learning_rate": 8.228571428571428e-07,
|
| 286 |
+
"loss": 4.1772,
|
| 287 |
+
"step": 1125
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 1.1,
|
| 291 |
+
"learning_rate": 8.157142857142856e-07,
|
| 292 |
+
"loss": 4.1451,
|
| 293 |
+
"step": 1150
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 1.12,
|
| 297 |
+
"learning_rate": 8.085714285714285e-07,
|
| 298 |
+
"loss": 4.0833,
|
| 299 |
+
"step": 1175
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 1.15,
|
| 303 |
+
"learning_rate": 8.014285714285713e-07,
|
| 304 |
+
"loss": 4.1437,
|
| 305 |
+
"step": 1200
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 1.17,
|
| 309 |
+
"learning_rate": 7.942857142857142e-07,
|
| 310 |
+
"loss": 4.1286,
|
| 311 |
+
"step": 1225
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 1.19,
|
| 315 |
+
"learning_rate": 7.871428571428571e-07,
|
| 316 |
+
"loss": 4.1125,
|
| 317 |
+
"step": 1250
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 1.22,
|
| 321 |
+
"learning_rate": 7.799999999999999e-07,
|
| 322 |
+
"loss": 4.0927,
|
| 323 |
+
"step": 1275
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 1.24,
|
| 327 |
+
"learning_rate": 7.728571428571428e-07,
|
| 328 |
+
"loss": 4.1013,
|
| 329 |
+
"step": 1300
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 1.26,
|
| 333 |
+
"learning_rate": 7.657142857142856e-07,
|
| 334 |
+
"loss": 4.0453,
|
| 335 |
+
"step": 1325
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 1.29,
|
| 339 |
+
"learning_rate": 7.585714285714285e-07,
|
| 340 |
+
"loss": 4.1317,
|
| 341 |
+
"step": 1350
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 1.31,
|
| 345 |
+
"learning_rate": 7.514285714285715e-07,
|
| 346 |
+
"loss": 4.0638,
|
| 347 |
+
"step": 1375
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 1.34,
|
| 351 |
+
"learning_rate": 7.442857142857143e-07,
|
| 352 |
+
"loss": 4.1043,
|
| 353 |
+
"step": 1400
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 1.36,
|
| 357 |
+
"learning_rate": 7.371428571428571e-07,
|
| 358 |
+
"loss": 4.0761,
|
| 359 |
+
"step": 1425
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 1.38,
|
| 363 |
+
"learning_rate": 7.3e-07,
|
| 364 |
+
"loss": 3.9902,
|
| 365 |
+
"step": 1450
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 1.41,
|
| 369 |
+
"learning_rate": 7.228571428571429e-07,
|
| 370 |
+
"loss": 4.0019,
|
| 371 |
+
"step": 1475
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 1.43,
|
| 375 |
+
"learning_rate": 7.157142857142858e-07,
|
| 376 |
+
"loss": 4.0683,
|
| 377 |
+
"step": 1500
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 1.46,
|
| 381 |
+
"learning_rate": 7.085714285714285e-07,
|
| 382 |
+
"loss": 3.9523,
|
| 383 |
+
"step": 1525
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 1.48,
|
| 387 |
+
"learning_rate": 7.014285714285714e-07,
|
| 388 |
+
"loss": 3.9421,
|
| 389 |
+
"step": 1550
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 1.5,
|
| 393 |
+
"learning_rate": 6.942857142857143e-07,
|
| 394 |
+
"loss": 3.7698,
|
| 395 |
+
"step": 1575
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 1.53,
|
| 399 |
+
"learning_rate": 6.871428571428572e-07,
|
| 400 |
+
"loss": 3.7565,
|
| 401 |
+
"step": 1600
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 1.55,
|
| 405 |
+
"learning_rate": 6.800000000000001e-07,
|
| 406 |
+
"loss": 3.6593,
|
| 407 |
+
"step": 1625
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 1.57,
|
| 411 |
+
"learning_rate": 6.728571428571428e-07,
|
| 412 |
+
"loss": 3.4785,
|
| 413 |
+
"step": 1650
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 1.6,
|
| 417 |
+
"learning_rate": 6.657142857142857e-07,
|
| 418 |
+
"loss": 3.4676,
|
| 419 |
+
"step": 1675
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 1.62,
|
| 423 |
+
"learning_rate": 6.585714285714286e-07,
|
| 424 |
+
"loss": 3.3216,
|
| 425 |
+
"step": 1700
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 1.65,
|
| 429 |
+
"learning_rate": 6.514285714285715e-07,
|
| 430 |
+
"loss": 3.2336,
|
| 431 |
+
"step": 1725
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 1.67,
|
| 435 |
+
"learning_rate": 6.442857142857142e-07,
|
| 436 |
+
"loss": 3.1206,
|
| 437 |
+
"step": 1750
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 1.69,
|
| 441 |
+
"learning_rate": 6.371428571428571e-07,
|
| 442 |
+
"loss": 3.0662,
|
| 443 |
+
"step": 1775
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 1.72,
|
| 447 |
+
"learning_rate": 6.3e-07,
|
| 448 |
+
"loss": 2.972,
|
| 449 |
+
"step": 1800
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 1.74,
|
| 453 |
+
"learning_rate": 6.228571428571429e-07,
|
| 454 |
+
"loss": 2.8847,
|
| 455 |
+
"step": 1825
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 1.77,
|
| 459 |
+
"learning_rate": 6.157142857142858e-07,
|
| 460 |
+
"loss": 2.8173,
|
| 461 |
+
"step": 1850
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 1.79,
|
| 465 |
+
"learning_rate": 6.085714285714285e-07,
|
| 466 |
+
"loss": 2.7043,
|
| 467 |
+
"step": 1875
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 1.81,
|
| 471 |
+
"learning_rate": 6.014285714285714e-07,
|
| 472 |
+
"loss": 2.7055,
|
| 473 |
+
"step": 1900
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 1.84,
|
| 477 |
+
"learning_rate": 5.942857142857143e-07,
|
| 478 |
+
"loss": 2.7277,
|
| 479 |
+
"step": 1925
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 1.86,
|
| 483 |
+
"learning_rate": 5.871428571428572e-07,
|
| 484 |
+
"loss": 2.6155,
|
| 485 |
+
"step": 1950
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 1.88,
|
| 489 |
+
"learning_rate": 5.8e-07,
|
| 490 |
+
"loss": 2.533,
|
| 491 |
+
"step": 1975
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 1.91,
|
| 495 |
+
"learning_rate": 5.728571428571428e-07,
|
| 496 |
+
"loss": 2.4631,
|
| 497 |
+
"step": 2000
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 1.91,
|
| 501 |
+
"eval_loss": 2.7998619079589844,
|
| 502 |
+
"eval_runtime": 865.3409,
|
| 503 |
+
"eval_samples_per_second": 1.733,
|
| 504 |
+
"eval_steps_per_second": 0.217,
|
| 505 |
+
"eval_wer": 160.23294509151413,
|
| 506 |
+
"step": 2000
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 1.93,
|
| 510 |
+
"learning_rate": 5.657142857142857e-07,
|
| 511 |
+
"loss": 2.4711,
|
| 512 |
+
"step": 2025
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 1.96,
|
| 516 |
+
"learning_rate": 5.585714285714286e-07,
|
| 517 |
+
"loss": 2.3802,
|
| 518 |
+
"step": 2050
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 1.98,
|
| 522 |
+
"learning_rate": 5.514285714285714e-07,
|
| 523 |
+
"loss": 2.4496,
|
| 524 |
+
"step": 2075
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 2.0,
|
| 528 |
+
"learning_rate": 5.442857142857142e-07,
|
| 529 |
+
"loss": 2.3227,
|
| 530 |
+
"step": 2100
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 2.03,
|
| 534 |
+
"learning_rate": 5.371428571428571e-07,
|
| 535 |
+
"loss": 2.2921,
|
| 536 |
+
"step": 2125
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 2.05,
|
| 540 |
+
"learning_rate": 5.3e-07,
|
| 541 |
+
"loss": 2.2983,
|
| 542 |
+
"step": 2150
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 2.08,
|
| 546 |
+
"learning_rate": 5.228571428571429e-07,
|
| 547 |
+
"loss": 2.279,
|
| 548 |
+
"step": 2175
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 2.1,
|
| 552 |
+
"learning_rate": 5.157142857142857e-07,
|
| 553 |
+
"loss": 2.298,
|
| 554 |
+
"step": 2200
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 2.12,
|
| 558 |
+
"learning_rate": 5.085714285714285e-07,
|
| 559 |
+
"loss": 2.2966,
|
| 560 |
+
"step": 2225
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 2.15,
|
| 564 |
+
"learning_rate": 5.014285714285714e-07,
|
| 565 |
+
"loss": 2.2024,
|
| 566 |
+
"step": 2250
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 2.17,
|
| 570 |
+
"learning_rate": 4.942857142857143e-07,
|
| 571 |
+
"loss": 2.2506,
|
| 572 |
+
"step": 2275
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 2.19,
|
| 576 |
+
"learning_rate": 4.871428571428572e-07,
|
| 577 |
+
"loss": 2.1831,
|
| 578 |
+
"step": 2300
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 2.22,
|
| 582 |
+
"learning_rate": 4.8e-07,
|
| 583 |
+
"loss": 2.1786,
|
| 584 |
+
"step": 2325
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 2.24,
|
| 588 |
+
"learning_rate": 4.7285714285714284e-07,
|
| 589 |
+
"loss": 2.1795,
|
| 590 |
+
"step": 2350
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 2.27,
|
| 594 |
+
"learning_rate": 4.6571428571428566e-07,
|
| 595 |
+
"loss": 2.1797,
|
| 596 |
+
"step": 2375
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 2.29,
|
| 600 |
+
"learning_rate": 4.5857142857142854e-07,
|
| 601 |
+
"loss": 2.1381,
|
| 602 |
+
"step": 2400
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 2.31,
|
| 606 |
+
"learning_rate": 4.5142857142857137e-07,
|
| 607 |
+
"loss": 2.0612,
|
| 608 |
+
"step": 2425
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 2.34,
|
| 612 |
+
"learning_rate": 4.4428571428571425e-07,
|
| 613 |
+
"loss": 2.0958,
|
| 614 |
+
"step": 2450
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 2.36,
|
| 618 |
+
"learning_rate": 4.3714285714285713e-07,
|
| 619 |
+
"loss": 2.0547,
|
| 620 |
+
"step": 2475
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 2.39,
|
| 624 |
+
"learning_rate": 4.2999999999999996e-07,
|
| 625 |
+
"loss": 2.0996,
|
| 626 |
+
"step": 2500
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 2.41,
|
| 630 |
+
"learning_rate": 4.2285714285714284e-07,
|
| 631 |
+
"loss": 2.0421,
|
| 632 |
+
"step": 2525
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 2.43,
|
| 636 |
+
"learning_rate": 4.1571428571428567e-07,
|
| 637 |
+
"loss": 1.9812,
|
| 638 |
+
"step": 2550
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 2.46,
|
| 642 |
+
"learning_rate": 4.0857142857142855e-07,
|
| 643 |
+
"loss": 2.0337,
|
| 644 |
+
"step": 2575
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 2.48,
|
| 648 |
+
"learning_rate": 4.014285714285714e-07,
|
| 649 |
+
"loss": 2.0123,
|
| 650 |
+
"step": 2600
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 2.5,
|
| 654 |
+
"learning_rate": 3.9428571428571425e-07,
|
| 655 |
+
"loss": 2.0056,
|
| 656 |
+
"step": 2625
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 2.53,
|
| 660 |
+
"learning_rate": 3.871428571428571e-07,
|
| 661 |
+
"loss": 1.9827,
|
| 662 |
+
"step": 2650
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 2.55,
|
| 666 |
+
"learning_rate": 3.7999999999999996e-07,
|
| 667 |
+
"loss": 2.0326,
|
| 668 |
+
"step": 2675
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 2.58,
|
| 672 |
+
"learning_rate": 3.728571428571428e-07,
|
| 673 |
+
"loss": 1.9789,
|
| 674 |
+
"step": 2700
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 2.6,
|
| 678 |
+
"learning_rate": 3.657142857142857e-07,
|
| 679 |
+
"loss": 1.9974,
|
| 680 |
+
"step": 2725
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 2.62,
|
| 684 |
+
"learning_rate": 3.585714285714286e-07,
|
| 685 |
+
"loss": 1.9701,
|
| 686 |
+
"step": 2750
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 2.65,
|
| 690 |
+
"learning_rate": 3.5142857142857143e-07,
|
| 691 |
+
"loss": 1.9797,
|
| 692 |
+
"step": 2775
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 2.67,
|
| 696 |
+
"learning_rate": 3.442857142857143e-07,
|
| 697 |
+
"loss": 1.93,
|
| 698 |
+
"step": 2800
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 2.7,
|
| 702 |
+
"learning_rate": 3.3714285714285714e-07,
|
| 703 |
+
"loss": 1.9068,
|
| 704 |
+
"step": 2825
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 2.72,
|
| 708 |
+
"learning_rate": 3.3e-07,
|
| 709 |
+
"loss": 1.9442,
|
| 710 |
+
"step": 2850
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 2.74,
|
| 714 |
+
"learning_rate": 3.2285714285714284e-07,
|
| 715 |
+
"loss": 1.8534,
|
| 716 |
+
"step": 2875
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 2.77,
|
| 720 |
+
"learning_rate": 3.157142857142857e-07,
|
| 721 |
+
"loss": 1.9212,
|
| 722 |
+
"step": 2900
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 2.79,
|
| 726 |
+
"learning_rate": 3.0857142857142855e-07,
|
| 727 |
+
"loss": 1.9611,
|
| 728 |
+
"step": 2925
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 2.81,
|
| 732 |
+
"learning_rate": 3.0142857142857143e-07,
|
| 733 |
+
"loss": 1.9114,
|
| 734 |
+
"step": 2950
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 2.84,
|
| 738 |
+
"learning_rate": 2.9428571428571426e-07,
|
| 739 |
+
"loss": 1.8668,
|
| 740 |
+
"step": 2975
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 2.86,
|
| 744 |
+
"learning_rate": 2.8714285714285714e-07,
|
| 745 |
+
"loss": 1.8798,
|
| 746 |
+
"step": 3000
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 2.86,
|
| 750 |
+
"eval_loss": 2.237186908721924,
|
| 751 |
+
"eval_runtime": 535.0618,
|
| 752 |
+
"eval_samples_per_second": 2.803,
|
| 753 |
+
"eval_steps_per_second": 0.351,
|
| 754 |
+
"eval_wer": 103.72296173044926,
|
| 755 |
+
"step": 3000
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 2.89,
|
| 759 |
+
"learning_rate": 2.8e-07,
|
| 760 |
+
"loss": 1.9297,
|
| 761 |
+
"step": 3025
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 2.91,
|
| 765 |
+
"learning_rate": 2.7285714285714285e-07,
|
| 766 |
+
"loss": 1.9183,
|
| 767 |
+
"step": 3050
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 2.93,
|
| 771 |
+
"learning_rate": 2.657142857142857e-07,
|
| 772 |
+
"loss": 1.8872,
|
| 773 |
+
"step": 3075
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 2.96,
|
| 777 |
+
"learning_rate": 2.5857142857142855e-07,
|
| 778 |
+
"loss": 1.8601,
|
| 779 |
+
"step": 3100
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 2.98,
|
| 783 |
+
"learning_rate": 2.5142857142857143e-07,
|
| 784 |
+
"loss": 1.8245,
|
| 785 |
+
"step": 3125
|
| 786 |
+
},
|
| 787 |
+
{
|
| 788 |
+
"epoch": 3.01,
|
| 789 |
+
"learning_rate": 2.4428571428571426e-07,
|
| 790 |
+
"loss": 1.8316,
|
| 791 |
+
"step": 3150
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 3.03,
|
| 795 |
+
"learning_rate": 2.3714285714285714e-07,
|
| 796 |
+
"loss": 1.8584,
|
| 797 |
+
"step": 3175
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 3.05,
|
| 801 |
+
"learning_rate": 2.3e-07,
|
| 802 |
+
"loss": 1.7541,
|
| 803 |
+
"step": 3200
|
| 804 |
+
},
|
| 805 |
+
{
|
| 806 |
+
"epoch": 3.08,
|
| 807 |
+
"learning_rate": 2.2285714285714285e-07,
|
| 808 |
+
"loss": 1.7497,
|
| 809 |
+
"step": 3225
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 3.1,
|
| 813 |
+
"learning_rate": 2.157142857142857e-07,
|
| 814 |
+
"loss": 1.8537,
|
| 815 |
+
"step": 3250
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 3.12,
|
| 819 |
+
"learning_rate": 2.0857142857142856e-07,
|
| 820 |
+
"loss": 1.8072,
|
| 821 |
+
"step": 3275
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 3.15,
|
| 825 |
+
"learning_rate": 2.014285714285714e-07,
|
| 826 |
+
"loss": 1.8278,
|
| 827 |
+
"step": 3300
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"epoch": 3.17,
|
| 831 |
+
"learning_rate": 1.9428571428571426e-07,
|
| 832 |
+
"loss": 1.8196,
|
| 833 |
+
"step": 3325
|
| 834 |
+
},
|
| 835 |
+
{
|
| 836 |
+
"epoch": 3.2,
|
| 837 |
+
"learning_rate": 1.8714285714285712e-07,
|
| 838 |
+
"loss": 1.79,
|
| 839 |
+
"step": 3350
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 3.22,
|
| 843 |
+
"learning_rate": 1.8e-07,
|
| 844 |
+
"loss": 1.8385,
|
| 845 |
+
"step": 3375
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"epoch": 3.24,
|
| 849 |
+
"learning_rate": 1.7285714285714285e-07,
|
| 850 |
+
"loss": 1.7812,
|
| 851 |
+
"step": 3400
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 3.27,
|
| 855 |
+
"learning_rate": 1.657142857142857e-07,
|
| 856 |
+
"loss": 1.7756,
|
| 857 |
+
"step": 3425
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 3.29,
|
| 861 |
+
"learning_rate": 1.5857142857142856e-07,
|
| 862 |
+
"loss": 1.7903,
|
| 863 |
+
"step": 3450
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 3.32,
|
| 867 |
+
"learning_rate": 1.5142857142857144e-07,
|
| 868 |
+
"loss": 1.7772,
|
| 869 |
+
"step": 3475
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 3.34,
|
| 873 |
+
"learning_rate": 1.442857142857143e-07,
|
| 874 |
+
"loss": 1.8031,
|
| 875 |
+
"step": 3500
|
| 876 |
+
},
|
| 877 |
+
{
|
| 878 |
+
"epoch": 3.36,
|
| 879 |
+
"learning_rate": 1.3714285714285715e-07,
|
| 880 |
+
"loss": 1.7402,
|
| 881 |
+
"step": 3525
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 3.39,
|
| 885 |
+
"learning_rate": 1.3e-07,
|
| 886 |
+
"loss": 1.7751,
|
| 887 |
+
"step": 3550
|
| 888 |
+
},
|
| 889 |
+
{
|
| 890 |
+
"epoch": 3.41,
|
| 891 |
+
"learning_rate": 1.2285714285714285e-07,
|
| 892 |
+
"loss": 1.8169,
|
| 893 |
+
"step": 3575
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 3.44,
|
| 897 |
+
"learning_rate": 1.1571428571428571e-07,
|
| 898 |
+
"loss": 1.7689,
|
| 899 |
+
"step": 3600
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 3.46,
|
| 903 |
+
"learning_rate": 1.0857142857142856e-07,
|
| 904 |
+
"loss": 1.8012,
|
| 905 |
+
"step": 3625
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 3.48,
|
| 909 |
+
"learning_rate": 1.0142857142857141e-07,
|
| 910 |
+
"loss": 1.7949,
|
| 911 |
+
"step": 3650
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 3.51,
|
| 915 |
+
"learning_rate": 9.42857142857143e-08,
|
| 916 |
+
"loss": 1.7676,
|
| 917 |
+
"step": 3675
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"epoch": 3.53,
|
| 921 |
+
"learning_rate": 8.714285714285715e-08,
|
| 922 |
+
"loss": 1.7626,
|
| 923 |
+
"step": 3700
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"epoch": 3.55,
|
| 927 |
+
"learning_rate": 8e-08,
|
| 928 |
+
"loss": 1.8011,
|
| 929 |
+
"step": 3725
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 3.58,
|
| 933 |
+
"learning_rate": 7.285714285714286e-08,
|
| 934 |
+
"loss": 1.7573,
|
| 935 |
+
"step": 3750
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 3.6,
|
| 939 |
+
"learning_rate": 6.571428571428571e-08,
|
| 940 |
+
"loss": 1.7832,
|
| 941 |
+
"step": 3775
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 3.63,
|
| 945 |
+
"learning_rate": 5.857142857142857e-08,
|
| 946 |
+
"loss": 1.8049,
|
| 947 |
+
"step": 3800
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 3.65,
|
| 951 |
+
"learning_rate": 5.142857142857142e-08,
|
| 952 |
+
"loss": 1.7823,
|
| 953 |
+
"step": 3825
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 3.67,
|
| 957 |
+
"learning_rate": 4.4285714285714283e-08,
|
| 958 |
+
"loss": 1.7331,
|
| 959 |
+
"step": 3850
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 3.7,
|
| 963 |
+
"learning_rate": 3.7142857142857144e-08,
|
| 964 |
+
"loss": 1.7468,
|
| 965 |
+
"step": 3875
|
| 966 |
+
},
|
| 967 |
+
{
|
| 968 |
+
"epoch": 3.72,
|
| 969 |
+
"learning_rate": 3e-08,
|
| 970 |
+
"loss": 1.8197,
|
| 971 |
+
"step": 3900
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"epoch": 3.75,
|
| 975 |
+
"learning_rate": 2.2857142857142858e-08,
|
| 976 |
+
"loss": 1.8088,
|
| 977 |
+
"step": 3925
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 3.77,
|
| 981 |
+
"learning_rate": 1.5714285714285715e-08,
|
| 982 |
+
"loss": 1.7755,
|
| 983 |
+
"step": 3950
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 3.79,
|
| 987 |
+
"learning_rate": 8.571428571428572e-09,
|
| 988 |
+
"loss": 1.7471,
|
| 989 |
+
"step": 3975
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 3.82,
|
| 993 |
+
"learning_rate": 1.4285714285714286e-09,
|
| 994 |
+
"loss": 1.7353,
|
| 995 |
+
"step": 4000
|
| 996 |
+
},
|
| 997 |
+
{
|
| 998 |
+
"epoch": 3.82,
|
| 999 |
+
"eval_loss": 2.137834072113037,
|
| 1000 |
+
"eval_runtime": 453.4496,
|
| 1001 |
+
"eval_samples_per_second": 3.308,
|
| 1002 |
+
"eval_steps_per_second": 0.415,
|
| 1003 |
+
"eval_wer": 97.56655574043262,
|
| 1004 |
+
"step": 4000
|
| 1005 |
+
}
|
| 1006 |
+
],
|
| 1007 |
+
"max_steps": 4000,
|
| 1008 |
+
"num_train_epochs": 4,
|
| 1009 |
+
"total_flos": 1.57509165109248e+18,
|
| 1010 |
+
"trial_name": null,
|
| 1011 |
+
"trial_params": null
|
| 1012 |
+
}
|
checkpoint-4000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c20c43146ef36d4705861b39f63a0d14edd9a86b44771cc3a2223883763a5532
|
| 3 |
+
size 3515
|