Training in progress, step 300
Browse files- checkpoint-300/optimizer.pt +1 -1
- checkpoint-300/pytorch_model.bin +1 -1
- checkpoint-300/rng_state.pth +2 -2
- checkpoint-300/scaler.pt +1 -1
- checkpoint-300/scheduler.pt +1 -1
- checkpoint-300/trainer_state.json +116 -116
- checkpoint-300/training_args.bin +2 -2
- pytorch_model.bin +1 -1
- runs/Dec20_20-28-50_129-146-32-172/events.out.tfevents.1671568135.129-146-32-172.151517.0 +2 -2
- whisper_small_ps_augmented.py +1 -0
checkpoint-300/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1934161093
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:22e62120437d4a5d43ed19f559839cbc39250df020c1097fa01f356bcbf4bba8
|
| 3 |
size 1934161093
|
checkpoint-300/pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 967102601
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:18928fdec4a6f78c294cc9f60f25b8acac276650309efc21ceb0ecf42f321cb0
|
| 3 |
size 967102601
|
checkpoint-300/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1ab614b521b578676b3402f6dd6f0907952e16f2d208b34af9d84266a360fc37
|
| 3 |
+
size 14639
|
checkpoint-300/scaler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 557
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ce7f2249620974c2bcc6e52f5d48799672ec47f32baf09849b58c0ab5addde0f
|
| 3 |
size 557
|
checkpoint-300/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 627
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cbb349d7867b639670f1b127be8892c37cb67d6956f365801519932598abf657
|
| 3 |
size 627
|
checkpoint-300/trainer_state.json
CHANGED
|
@@ -1,226 +1,226 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric":
|
| 3 |
"best_model_checkpoint": "./checkpoint-300",
|
| 4 |
-
"epoch":
|
| 5 |
"global_step": 300,
|
| 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.
|
| 12 |
-
"learning_rate":
|
| 13 |
-
"loss": 4.
|
| 14 |
"step": 10
|
| 15 |
},
|
| 16 |
{
|
| 17 |
-
"epoch": 0.
|
| 18 |
-
"learning_rate":
|
| 19 |
-
"loss": 2.
|
| 20 |
"step": 20
|
| 21 |
},
|
| 22 |
{
|
| 23 |
-
"epoch": 0.
|
| 24 |
-
"learning_rate": 9.
|
| 25 |
-
"loss": 2.
|
| 26 |
"step": 30
|
| 27 |
},
|
| 28 |
{
|
| 29 |
-
"epoch": 0.
|
| 30 |
-
"learning_rate": 9.
|
| 31 |
-
"loss": 1.
|
| 32 |
"step": 40
|
| 33 |
},
|
| 34 |
{
|
| 35 |
-
"epoch":
|
| 36 |
-
"learning_rate": 9.
|
| 37 |
-
"loss": 1.
|
| 38 |
"step": 50
|
| 39 |
},
|
| 40 |
{
|
| 41 |
-
"epoch":
|
| 42 |
-
"learning_rate":
|
| 43 |
-
"loss": 1.
|
| 44 |
"step": 60
|
| 45 |
},
|
| 46 |
{
|
| 47 |
-
"epoch":
|
| 48 |
-
"learning_rate":
|
| 49 |
-
"loss": 1.
|
| 50 |
"step": 70
|
| 51 |
},
|
| 52 |
{
|
| 53 |
-
"epoch":
|
| 54 |
-
"learning_rate": 8.
|
| 55 |
-
"loss":
|
| 56 |
"step": 80
|
| 57 |
},
|
| 58 |
{
|
| 59 |
-
"epoch":
|
| 60 |
-
"learning_rate":
|
| 61 |
-
"loss":
|
| 62 |
"step": 90
|
| 63 |
},
|
| 64 |
{
|
| 65 |
-
"epoch":
|
| 66 |
-
"learning_rate":
|
| 67 |
-
"loss": 0.
|
| 68 |
"step": 100
|
| 69 |
},
|
| 70 |
{
|
| 71 |
-
"epoch":
|
| 72 |
-
"eval_cer":
|
| 73 |
-
"eval_loss": 0.
|
| 74 |
-
"eval_runtime":
|
| 75 |
-
"eval_samples_per_second":
|
| 76 |
-
"eval_steps_per_second": 0.
|
| 77 |
-
"eval_wer":
|
| 78 |
"step": 100
|
| 79 |
},
|
| 80 |
{
|
| 81 |
-
"epoch":
|
| 82 |
-
"learning_rate":
|
| 83 |
-
"loss": 0.
|
| 84 |
"step": 110
|
| 85 |
},
|
| 86 |
{
|
| 87 |
-
"epoch":
|
| 88 |
-
"learning_rate":
|
| 89 |
-
"loss": 0.
|
| 90 |
"step": 120
|
| 91 |
},
|
| 92 |
{
|
| 93 |
-
"epoch":
|
| 94 |
-
"learning_rate":
|
| 95 |
-
"loss": 0.
|
| 96 |
"step": 130
|
| 97 |
},
|
| 98 |
{
|
| 99 |
-
"epoch":
|
| 100 |
-
"learning_rate":
|
| 101 |
-
"loss": 0.
|
| 102 |
"step": 140
|
| 103 |
},
|
| 104 |
{
|
| 105 |
-
"epoch":
|
| 106 |
-
"learning_rate":
|
| 107 |
-
"loss": 0.
|
| 108 |
"step": 150
|
| 109 |
},
|
| 110 |
{
|
| 111 |
-
"epoch":
|
| 112 |
-
"learning_rate":
|
| 113 |
-
"loss": 0.
|
| 114 |
"step": 160
|
| 115 |
},
|
| 116 |
{
|
| 117 |
-
"epoch":
|
| 118 |
-
"learning_rate":
|
| 119 |
-
"loss": 0.
|
| 120 |
"step": 170
|
| 121 |
},
|
| 122 |
{
|
| 123 |
-
"epoch":
|
| 124 |
-
"learning_rate":
|
| 125 |
-
"loss": 0.
|
| 126 |
"step": 180
|
| 127 |
},
|
| 128 |
{
|
| 129 |
-
"epoch":
|
| 130 |
-
"learning_rate":
|
| 131 |
-
"loss": 0.
|
| 132 |
"step": 190
|
| 133 |
},
|
| 134 |
{
|
| 135 |
-
"epoch":
|
| 136 |
-
"learning_rate":
|
| 137 |
-
"loss": 0.
|
| 138 |
"step": 200
|
| 139 |
},
|
| 140 |
{
|
| 141 |
-
"epoch":
|
| 142 |
-
"eval_cer":
|
| 143 |
-
"eval_loss": 0.
|
| 144 |
-
"eval_runtime":
|
| 145 |
-
"eval_samples_per_second": 0.
|
| 146 |
-
"eval_steps_per_second": 0.
|
| 147 |
-
"eval_wer":
|
| 148 |
"step": 200
|
| 149 |
},
|
| 150 |
{
|
| 151 |
-
"epoch":
|
| 152 |
-
"learning_rate":
|
| 153 |
-
"loss": 0.
|
| 154 |
"step": 210
|
| 155 |
},
|
| 156 |
{
|
| 157 |
-
"epoch":
|
| 158 |
-
"learning_rate":
|
| 159 |
-
"loss": 0.
|
| 160 |
"step": 220
|
| 161 |
},
|
| 162 |
{
|
| 163 |
-
"epoch":
|
| 164 |
-
"learning_rate":
|
| 165 |
-
"loss": 0.
|
| 166 |
"step": 230
|
| 167 |
},
|
| 168 |
{
|
| 169 |
-
"epoch":
|
| 170 |
-
"learning_rate":
|
| 171 |
-
"loss": 0.
|
| 172 |
"step": 240
|
| 173 |
},
|
| 174 |
{
|
| 175 |
-
"epoch":
|
| 176 |
-
"learning_rate":
|
| 177 |
-
"loss": 0.
|
| 178 |
"step": 250
|
| 179 |
},
|
| 180 |
{
|
| 181 |
-
"epoch":
|
| 182 |
-
"learning_rate":
|
| 183 |
-
"loss": 0.
|
| 184 |
"step": 260
|
| 185 |
},
|
| 186 |
{
|
| 187 |
-
"epoch":
|
| 188 |
-
"learning_rate":
|
| 189 |
-
"loss": 0.
|
| 190 |
"step": 270
|
| 191 |
},
|
| 192 |
{
|
| 193 |
-
"epoch":
|
| 194 |
-
"learning_rate":
|
| 195 |
-
"loss": 0.
|
| 196 |
"step": 280
|
| 197 |
},
|
| 198 |
{
|
| 199 |
-
"epoch":
|
| 200 |
-
"learning_rate": 4.
|
| 201 |
-
"loss": 0.
|
| 202 |
"step": 290
|
| 203 |
},
|
| 204 |
{
|
| 205 |
-
"epoch":
|
| 206 |
-
"learning_rate":
|
| 207 |
-
"loss": 0.
|
| 208 |
"step": 300
|
| 209 |
},
|
| 210 |
{
|
| 211 |
-
"epoch":
|
| 212 |
-
"eval_cer":
|
| 213 |
-
"eval_loss": 0.
|
| 214 |
-
"eval_runtime":
|
| 215 |
-
"eval_samples_per_second": 0.
|
| 216 |
-
"eval_steps_per_second": 0.
|
| 217 |
-
"eval_wer":
|
| 218 |
"step": 300
|
| 219 |
}
|
| 220 |
],
|
| 221 |
-
"max_steps":
|
| 222 |
-
"num_train_epochs":
|
| 223 |
-
"total_flos":
|
| 224 |
"trial_name": null,
|
| 225 |
"trial_params": null
|
| 226 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 50.13619854721549,
|
| 3 |
"best_model_checkpoint": "./checkpoint-300",
|
| 4 |
+
"epoch": 7.141176470588236,
|
| 5 |
"global_step": 300,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
+
"epoch": 0.24,
|
| 12 |
+
"learning_rate": 2.666666666666667e-06,
|
| 13 |
+
"loss": 4.3134,
|
| 14 |
"step": 10
|
| 15 |
},
|
| 16 |
{
|
| 17 |
+
"epoch": 0.47,
|
| 18 |
+
"learning_rate": 6e-06,
|
| 19 |
+
"loss": 2.9329,
|
| 20 |
"step": 20
|
| 21 |
},
|
| 22 |
{
|
| 23 |
+
"epoch": 0.71,
|
| 24 |
+
"learning_rate": 9.333333333333334e-06,
|
| 25 |
+
"loss": 2.0584,
|
| 26 |
"step": 30
|
| 27 |
},
|
| 28 |
{
|
| 29 |
+
"epoch": 0.94,
|
| 30 |
+
"learning_rate": 9.829787234042554e-06,
|
| 31 |
+
"loss": 1.566,
|
| 32 |
"step": 40
|
| 33 |
},
|
| 34 |
{
|
| 35 |
+
"epoch": 1.19,
|
| 36 |
+
"learning_rate": 9.617021276595745e-06,
|
| 37 |
+
"loss": 1.3777,
|
| 38 |
"step": 50
|
| 39 |
},
|
| 40 |
{
|
| 41 |
+
"epoch": 1.42,
|
| 42 |
+
"learning_rate": 9.404255319148937e-06,
|
| 43 |
+
"loss": 1.1469,
|
| 44 |
"step": 60
|
| 45 |
},
|
| 46 |
{
|
| 47 |
+
"epoch": 1.66,
|
| 48 |
+
"learning_rate": 9.191489361702128e-06,
|
| 49 |
+
"loss": 1.0638,
|
| 50 |
"step": 70
|
| 51 |
},
|
| 52 |
{
|
| 53 |
+
"epoch": 1.89,
|
| 54 |
+
"learning_rate": 8.97872340425532e-06,
|
| 55 |
+
"loss": 0.9974,
|
| 56 |
"step": 80
|
| 57 |
},
|
| 58 |
{
|
| 59 |
+
"epoch": 2.14,
|
| 60 |
+
"learning_rate": 8.765957446808512e-06,
|
| 61 |
+
"loss": 0.9615,
|
| 62 |
"step": 90
|
| 63 |
},
|
| 64 |
{
|
| 65 |
+
"epoch": 2.38,
|
| 66 |
+
"learning_rate": 8.553191489361703e-06,
|
| 67 |
+
"loss": 0.8262,
|
| 68 |
"step": 100
|
| 69 |
},
|
| 70 |
{
|
| 71 |
+
"epoch": 2.38,
|
| 72 |
+
"eval_cer": 149.33157314260887,
|
| 73 |
+
"eval_loss": 0.8188337683677673,
|
| 74 |
+
"eval_runtime": 459.9395,
|
| 75 |
+
"eval_samples_per_second": 1.113,
|
| 76 |
+
"eval_steps_per_second": 0.139,
|
| 77 |
+
"eval_wer": 146.85230024213075,
|
| 78 |
"step": 100
|
| 79 |
},
|
| 80 |
{
|
| 81 |
+
"epoch": 2.61,
|
| 82 |
+
"learning_rate": 8.340425531914894e-06,
|
| 83 |
+
"loss": 0.7986,
|
| 84 |
"step": 110
|
| 85 |
},
|
| 86 |
{
|
| 87 |
+
"epoch": 2.85,
|
| 88 |
+
"learning_rate": 8.127659574468085e-06,
|
| 89 |
+
"loss": 0.7707,
|
| 90 |
"step": 120
|
| 91 |
},
|
| 92 |
{
|
| 93 |
+
"epoch": 3.09,
|
| 94 |
+
"learning_rate": 7.914893617021278e-06,
|
| 95 |
+
"loss": 0.7623,
|
| 96 |
"step": 130
|
| 97 |
},
|
| 98 |
{
|
| 99 |
+
"epoch": 3.33,
|
| 100 |
+
"learning_rate": 7.702127659574469e-06,
|
| 101 |
+
"loss": 0.6451,
|
| 102 |
"step": 140
|
| 103 |
},
|
| 104 |
{
|
| 105 |
+
"epoch": 3.56,
|
| 106 |
+
"learning_rate": 7.48936170212766e-06,
|
| 107 |
+
"loss": 0.6205,
|
| 108 |
"step": 150
|
| 109 |
},
|
| 110 |
{
|
| 111 |
+
"epoch": 3.8,
|
| 112 |
+
"learning_rate": 7.2765957446808524e-06,
|
| 113 |
+
"loss": 0.6097,
|
| 114 |
"step": 160
|
| 115 |
},
|
| 116 |
{
|
| 117 |
+
"epoch": 4.05,
|
| 118 |
+
"learning_rate": 7.0638297872340434e-06,
|
| 119 |
+
"loss": 0.6217,
|
| 120 |
"step": 170
|
| 121 |
},
|
| 122 |
{
|
| 123 |
+
"epoch": 4.28,
|
| 124 |
+
"learning_rate": 6.8510638297872344e-06,
|
| 125 |
+
"loss": 0.4584,
|
| 126 |
"step": 180
|
| 127 |
},
|
| 128 |
{
|
| 129 |
+
"epoch": 4.52,
|
| 130 |
+
"learning_rate": 6.6382978723404254e-06,
|
| 131 |
+
"loss": 0.4929,
|
| 132 |
"step": 190
|
| 133 |
},
|
| 134 |
{
|
| 135 |
+
"epoch": 4.75,
|
| 136 |
+
"learning_rate": 6.425531914893618e-06,
|
| 137 |
+
"loss": 0.4843,
|
| 138 |
"step": 200
|
| 139 |
},
|
| 140 |
{
|
| 141 |
+
"epoch": 4.75,
|
| 142 |
+
"eval_cer": 22.099862972494236,
|
| 143 |
+
"eval_loss": 0.6699215173721313,
|
| 144 |
+
"eval_runtime": 1062.3991,
|
| 145 |
+
"eval_samples_per_second": 0.482,
|
| 146 |
+
"eval_steps_per_second": 0.06,
|
| 147 |
+
"eval_wer": 52.66343825665859,
|
| 148 |
"step": 200
|
| 149 |
},
|
| 150 |
{
|
| 151 |
+
"epoch": 4.99,
|
| 152 |
+
"learning_rate": 6.212765957446809e-06,
|
| 153 |
+
"loss": 0.4678,
|
| 154 |
"step": 210
|
| 155 |
},
|
| 156 |
{
|
| 157 |
+
"epoch": 5.24,
|
| 158 |
+
"learning_rate": 6e-06,
|
| 159 |
+
"loss": 0.4014,
|
| 160 |
"step": 220
|
| 161 |
},
|
| 162 |
{
|
| 163 |
+
"epoch": 5.47,
|
| 164 |
+
"learning_rate": 5.787234042553191e-06,
|
| 165 |
+
"loss": 0.3926,
|
| 166 |
"step": 230
|
| 167 |
},
|
| 168 |
{
|
| 169 |
+
"epoch": 5.71,
|
| 170 |
+
"learning_rate": 5.574468085106384e-06,
|
| 171 |
+
"loss": 0.3806,
|
| 172 |
"step": 240
|
| 173 |
},
|
| 174 |
{
|
| 175 |
+
"epoch": 5.94,
|
| 176 |
+
"learning_rate": 5.361702127659575e-06,
|
| 177 |
+
"loss": 0.3855,
|
| 178 |
"step": 250
|
| 179 |
},
|
| 180 |
{
|
| 181 |
+
"epoch": 6.19,
|
| 182 |
+
"learning_rate": 5.148936170212766e-06,
|
| 183 |
+
"loss": 0.3478,
|
| 184 |
"step": 260
|
| 185 |
},
|
| 186 |
{
|
| 187 |
+
"epoch": 6.42,
|
| 188 |
+
"learning_rate": 4.936170212765958e-06,
|
| 189 |
+
"loss": 0.3078,
|
| 190 |
"step": 270
|
| 191 |
},
|
| 192 |
{
|
| 193 |
+
"epoch": 6.66,
|
| 194 |
+
"learning_rate": 4.7234042553191496e-06,
|
| 195 |
+
"loss": 0.3147,
|
| 196 |
"step": 280
|
| 197 |
},
|
| 198 |
{
|
| 199 |
+
"epoch": 6.89,
|
| 200 |
+
"learning_rate": 4.5106382978723406e-06,
|
| 201 |
+
"loss": 0.3213,
|
| 202 |
"step": 290
|
| 203 |
},
|
| 204 |
{
|
| 205 |
+
"epoch": 7.14,
|
| 206 |
+
"learning_rate": 4.297872340425532e-06,
|
| 207 |
+
"loss": 0.287,
|
| 208 |
"step": 300
|
| 209 |
},
|
| 210 |
{
|
| 211 |
+
"epoch": 7.14,
|
| 212 |
+
"eval_cer": 20.677784833394604,
|
| 213 |
+
"eval_loss": 0.6913720369338989,
|
| 214 |
+
"eval_runtime": 1003.603,
|
| 215 |
+
"eval_samples_per_second": 0.51,
|
| 216 |
+
"eval_steps_per_second": 0.064,
|
| 217 |
+
"eval_wer": 50.13619854721549,
|
| 218 |
"step": 300
|
| 219 |
}
|
| 220 |
],
|
| 221 |
+
"max_steps": 500,
|
| 222 |
+
"num_train_epochs": 12,
|
| 223 |
+
"total_flos": 5.58730193559552e+18,
|
| 224 |
"trial_name": null,
|
| 225 |
"trial_params": null
|
| 226 |
}
|
checkpoint-300/training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5518cd575512706c66c89476ae9c64328c1bcac98ec06151bdc9afebf6c503fb
|
| 3 |
+
size 3643
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 967102601
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:18928fdec4a6f78c294cc9f60f25b8acac276650309efc21ceb0ecf42f321cb0
|
| 3 |
size 967102601
|
runs/Dec20_20-28-50_129-146-32-172/events.out.tfevents.1671568135.129-146-32-172.151517.0
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c9f730aa019c3fdfb3bb29b6c3bdc8df372194567bace669c0490525c9864f8b
|
| 3 |
+
size 10049
|
whisper_small_ps_augmented.py
CHANGED
|
@@ -287,6 +287,7 @@ trainer = Seq2SeqTrainer(
|
|
| 287 |
tokenizer=processor.feature_extractor
|
| 288 |
|
| 289 |
)
|
|
|
|
| 290 |
|
| 291 |
"""We'll save the processor object once before starting training. Since the processor is not trainable, it won't change over the course of training:"""
|
| 292 |
|
|
|
|
| 287 |
tokenizer=processor.feature_extractor
|
| 288 |
|
| 289 |
)
|
| 290 |
+
trainer.train(resume_from_checkpoint = True)
|
| 291 |
|
| 292 |
"""We'll save the processor object once before starting training. Since the processor is not trainable, it won't change over the course of training:"""
|
| 293 |
|