Model save
Browse files- README.md +74 -38
- trainer_state.json +873 -297
README.md
CHANGED
|
@@ -16,8 +16,8 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 16 |
|
| 17 |
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on an unknown dataset.
|
| 18 |
It achieves the following results on the evaluation set:
|
| 19 |
-
- Loss: 1.
|
| 20 |
-
- Cer: 0.
|
| 21 |
|
| 22 |
## Model description
|
| 23 |
|
|
@@ -45,48 +45,84 @@ The following hyperparameters were used during training:
|
|
| 45 |
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
|
| 46 |
- lr_scheduler_type: linear
|
| 47 |
- lr_scheduler_warmup_steps: 100
|
| 48 |
-
- num_epochs:
|
| 49 |
- mixed_precision_training: Native AMP
|
| 50 |
|
| 51 |
### Training results
|
| 52 |
|
| 53 |
| Training Loss | Epoch | Step | Validation Loss | Cer |
|
| 54 |
|:-------------:|:-------:|:----:|:---------------:|:------:|
|
| 55 |
-
| 11.
|
| 56 |
-
| 3.
|
| 57 |
-
|
|
| 58 |
-
| 3.
|
| 59 |
-
| 3.
|
| 60 |
-
| 3.
|
| 61 |
-
| 3.
|
| 62 |
-
| 3.
|
| 63 |
-
| 3.
|
| 64 |
-
|
|
| 65 |
-
|
|
| 66 |
-
|
|
| 67 |
-
|
|
| 68 |
-
|
|
| 69 |
-
|
|
| 70 |
-
|
|
| 71 |
-
|
|
| 72 |
-
|
|
| 73 |
-
|
|
| 74 |
-
|
|
| 75 |
-
|
|
| 76 |
-
|
|
| 77 |
-
|
|
| 78 |
-
|
|
| 79 |
-
|
|
| 80 |
-
|
|
| 81 |
-
|
|
| 82 |
-
|
|
| 83 |
-
|
|
| 84 |
-
|
|
| 85 |
-
|
|
| 86 |
-
| 0.
|
| 87 |
-
| 0.
|
| 88 |
-
| 0.
|
| 89 |
-
| 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 90 |
|
| 91 |
|
| 92 |
### Framework versions
|
|
|
|
| 16 |
|
| 17 |
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on an unknown dataset.
|
| 18 |
It achieves the following results on the evaluation set:
|
| 19 |
+
- Loss: 1.6104
|
| 20 |
+
- Cer: 0.3521
|
| 21 |
|
| 22 |
## Model description
|
| 23 |
|
|
|
|
| 45 |
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
|
| 46 |
- lr_scheduler_type: linear
|
| 47 |
- lr_scheduler_warmup_steps: 100
|
| 48 |
+
- num_epochs: 100
|
| 49 |
- mixed_precision_training: Native AMP
|
| 50 |
|
| 51 |
### Training results
|
| 52 |
|
| 53 |
| Training Loss | Epoch | Step | Validation Loss | Cer |
|
| 54 |
|:-------------:|:-------:|:----:|:---------------:|:------:|
|
| 55 |
+
| 11.1682 | 1.3916 | 100 | 3.8553 | 0.9903 |
|
| 56 |
+
| 3.9311 | 2.7832 | 200 | 3.8241 | 0.9903 |
|
| 57 |
+
| 3.8623 | 4.1678 | 300 | 3.7760 | 0.9903 |
|
| 58 |
+
| 3.7693 | 5.5594 | 400 | 3.6686 | 0.9903 |
|
| 59 |
+
| 3.671 | 6.9510 | 500 | 3.5900 | 0.9893 |
|
| 60 |
+
| 3.5618 | 8.3357 | 600 | 3.5169 | 0.9713 |
|
| 61 |
+
| 3.4994 | 9.7273 | 700 | 3.3552 | 0.9699 |
|
| 62 |
+
| 3.3323 | 11.1119 | 800 | 3.1385 | 0.9540 |
|
| 63 |
+
| 3.163 | 12.5035 | 900 | 2.9224 | 0.9186 |
|
| 64 |
+
| 2.7901 | 13.8951 | 1000 | 2.1802 | 0.7828 |
|
| 65 |
+
| 2.3425 | 15.2797 | 1100 | 1.8406 | 0.6529 |
|
| 66 |
+
| 2.0608 | 16.6713 | 1200 | 1.6505 | 0.6329 |
|
| 67 |
+
| 1.8813 | 18.0559 | 1300 | 1.4769 | 0.5715 |
|
| 68 |
+
| 1.6705 | 19.4476 | 1400 | 1.4793 | 0.5581 |
|
| 69 |
+
| 1.558 | 20.8392 | 1500 | 1.3079 | 0.4970 |
|
| 70 |
+
| 1.4213 | 22.2238 | 1600 | 1.3552 | 0.4947 |
|
| 71 |
+
| 1.3122 | 23.6154 | 1700 | 1.2368 | 0.4355 |
|
| 72 |
+
| 1.2303 | 25.0 | 1800 | 1.2108 | 0.4347 |
|
| 73 |
+
| 1.1152 | 26.3916 | 1900 | 1.2177 | 0.4307 |
|
| 74 |
+
| 1.0441 | 27.7832 | 2000 | 1.3236 | 0.4291 |
|
| 75 |
+
| 0.9626 | 29.1678 | 2100 | 1.2738 | 0.4157 |
|
| 76 |
+
| 0.8987 | 30.5594 | 2200 | 1.2683 | 0.4190 |
|
| 77 |
+
| 0.8367 | 31.9510 | 2300 | 1.2570 | 0.4144 |
|
| 78 |
+
| 0.7617 | 33.3357 | 2400 | 1.2331 | 0.3876 |
|
| 79 |
+
| 0.7069 | 34.7273 | 2500 | 1.3284 | 0.4037 |
|
| 80 |
+
| 0.6874 | 36.1119 | 2600 | 1.2948 | 0.3818 |
|
| 81 |
+
| 0.6615 | 37.5035 | 2700 | 1.2998 | 0.3977 |
|
| 82 |
+
| 0.6086 | 38.8951 | 2800 | 1.3369 | 0.3758 |
|
| 83 |
+
| 0.5804 | 40.2797 | 2900 | 1.2815 | 0.3838 |
|
| 84 |
+
| 0.548 | 41.6713 | 3000 | 1.3390 | 0.3766 |
|
| 85 |
+
| 0.5239 | 43.0559 | 3100 | 1.2572 | 0.3673 |
|
| 86 |
+
| 0.4983 | 44.4476 | 3200 | 1.2955 | 0.3671 |
|
| 87 |
+
| 0.4793 | 45.8392 | 3300 | 1.3563 | 0.3729 |
|
| 88 |
+
| 0.438 | 47.2238 | 3400 | 1.4153 | 0.3915 |
|
| 89 |
+
| 0.4274 | 48.6154 | 3500 | 1.3198 | 0.3663 |
|
| 90 |
+
| 0.4064 | 50.0 | 3600 | 1.4351 | 0.3814 |
|
| 91 |
+
| 0.3812 | 51.3916 | 3700 | 1.3514 | 0.3620 |
|
| 92 |
+
| 0.3753 | 52.7832 | 3800 | 1.3715 | 0.3492 |
|
| 93 |
+
| 0.3549 | 54.1678 | 3900 | 1.4133 | 0.3649 |
|
| 94 |
+
| 0.3262 | 55.5594 | 4000 | 1.4260 | 0.3574 |
|
| 95 |
+
| 0.3296 | 56.9510 | 4100 | 1.5134 | 0.3552 |
|
| 96 |
+
| 0.3136 | 58.3357 | 4200 | 1.4696 | 0.3587 |
|
| 97 |
+
| 0.3009 | 59.7273 | 4300 | 1.4326 | 0.3554 |
|
| 98 |
+
| 0.2764 | 61.1119 | 4400 | 1.4486 | 0.3572 |
|
| 99 |
+
| 0.2738 | 62.5035 | 4500 | 1.4463 | 0.3593 |
|
| 100 |
+
| 0.2574 | 63.8951 | 4600 | 1.4303 | 0.3583 |
|
| 101 |
+
| 0.2397 | 65.2797 | 4700 | 1.4538 | 0.3446 |
|
| 102 |
+
| 0.2474 | 66.6713 | 4800 | 1.4416 | 0.3496 |
|
| 103 |
+
| 0.2212 | 68.0559 | 4900 | 1.4766 | 0.3448 |
|
| 104 |
+
| 0.2173 | 69.4476 | 5000 | 1.4785 | 0.3496 |
|
| 105 |
+
| 0.2138 | 70.8392 | 5100 | 1.4859 | 0.3582 |
|
| 106 |
+
| 0.2037 | 72.2238 | 5200 | 1.5022 | 0.3500 |
|
| 107 |
+
| 0.194 | 73.6154 | 5300 | 1.4964 | 0.3490 |
|
| 108 |
+
| 0.1758 | 75.0 | 5400 | 1.5645 | 0.3552 |
|
| 109 |
+
| 0.1693 | 76.3916 | 5500 | 1.5215 | 0.3492 |
|
| 110 |
+
| 0.1682 | 77.7832 | 5600 | 1.5572 | 0.3436 |
|
| 111 |
+
| 0.1616 | 79.1678 | 5700 | 1.4971 | 0.3461 |
|
| 112 |
+
| 0.1625 | 80.5594 | 5800 | 1.5327 | 0.3516 |
|
| 113 |
+
| 0.1432 | 81.9510 | 5900 | 1.5595 | 0.3506 |
|
| 114 |
+
| 0.1348 | 83.3357 | 6000 | 1.5562 | 0.3483 |
|
| 115 |
+
| 0.137 | 84.7273 | 6100 | 1.5902 | 0.3485 |
|
| 116 |
+
| 0.1263 | 86.1119 | 6200 | 1.5853 | 0.3521 |
|
| 117 |
+
| 0.1271 | 87.5035 | 6300 | 1.5977 | 0.3488 |
|
| 118 |
+
| 0.123 | 88.8951 | 6400 | 1.6024 | 0.3498 |
|
| 119 |
+
| 0.117 | 90.2797 | 6500 | 1.6093 | 0.3535 |
|
| 120 |
+
| 0.1077 | 91.6713 | 6600 | 1.5807 | 0.3519 |
|
| 121 |
+
| 0.1072 | 93.0559 | 6700 | 1.5801 | 0.3477 |
|
| 122 |
+
| 0.1063 | 94.4476 | 6800 | 1.5894 | 0.3502 |
|
| 123 |
+
| 0.103 | 95.8392 | 6900 | 1.6027 | 0.3498 |
|
| 124 |
+
| 0.1032 | 97.2238 | 7000 | 1.6034 | 0.3485 |
|
| 125 |
+
| 0.0971 | 98.6154 | 7100 | 1.6104 | 0.3481 |
|
| 126 |
|
| 127 |
|
| 128 |
### Framework versions
|
trainer_state.json
CHANGED
|
@@ -2,596 +2,1172 @@
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
-
"epoch":
|
| 6 |
"eval_steps": 100,
|
| 7 |
-
"global_step":
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
| 11 |
"log_history": [
|
| 12 |
{
|
| 13 |
"epoch": 1.3916083916083917,
|
| 14 |
-
"grad_norm": 0.
|
| 15 |
"learning_rate": 0.000285,
|
| 16 |
-
"loss": 11.
|
| 17 |
"step": 100
|
| 18 |
},
|
| 19 |
{
|
| 20 |
"epoch": 1.3916083916083917,
|
| 21 |
-
"eval_cer": 0.
|
| 22 |
-
"eval_loss": 3.
|
| 23 |
-
"eval_runtime": 6.
|
| 24 |
-
"eval_samples_per_second":
|
| 25 |
-
"eval_steps_per_second": 5.
|
| 26 |
"step": 100
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 2.7832167832167833,
|
| 30 |
-
"grad_norm": 1.
|
| 31 |
-
"learning_rate": 0.
|
| 32 |
-
"loss": 3.
|
| 33 |
"step": 200
|
| 34 |
},
|
| 35 |
{
|
| 36 |
"epoch": 2.7832167832167833,
|
| 37 |
-
"eval_cer": 0.
|
| 38 |
-
"eval_loss": 3.
|
| 39 |
-
"eval_runtime": 6.
|
| 40 |
-
"eval_samples_per_second": 46.
|
| 41 |
-
"eval_steps_per_second": 5.
|
| 42 |
"step": 200
|
| 43 |
},
|
| 44 |
{
|
| 45 |
"epoch": 4.1678321678321675,
|
| 46 |
-
"grad_norm": 1.
|
| 47 |
-
"learning_rate": 0.
|
| 48 |
-
"loss":
|
| 49 |
"step": 300
|
| 50 |
},
|
| 51 |
{
|
| 52 |
"epoch": 4.1678321678321675,
|
| 53 |
-
"eval_cer": 0.
|
| 54 |
-
"eval_loss": 3.
|
| 55 |
-
"eval_runtime": 6.
|
| 56 |
-
"eval_samples_per_second": 46.
|
| 57 |
-
"eval_steps_per_second": 5.
|
| 58 |
"step": 300
|
| 59 |
},
|
| 60 |
{
|
| 61 |
"epoch": 5.559440559440559,
|
| 62 |
-
"grad_norm":
|
| 63 |
-
"learning_rate": 0.
|
| 64 |
-
"loss": 3.
|
| 65 |
"step": 400
|
| 66 |
},
|
| 67 |
{
|
| 68 |
"epoch": 5.559440559440559,
|
| 69 |
"eval_cer": 0.9903044405662207,
|
| 70 |
-
"eval_loss": 3.
|
| 71 |
-
"eval_runtime": 6.
|
| 72 |
-
"eval_samples_per_second": 46.
|
| 73 |
-
"eval_steps_per_second": 5.
|
| 74 |
"step": 400
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 6.951048951048951,
|
| 78 |
-
"grad_norm": 1.
|
| 79 |
-
"learning_rate": 0.
|
| 80 |
-
"loss": 3.
|
| 81 |
"step": 500
|
| 82 |
},
|
| 83 |
{
|
| 84 |
"epoch": 6.951048951048951,
|
| 85 |
-
"eval_cer": 0.
|
| 86 |
-
"eval_loss": 3.
|
| 87 |
-
"eval_runtime": 6.
|
| 88 |
-
"eval_samples_per_second": 46.
|
| 89 |
-
"eval_steps_per_second": 5.
|
| 90 |
"step": 500
|
| 91 |
},
|
| 92 |
{
|
| 93 |
"epoch": 8.335664335664335,
|
| 94 |
-
"grad_norm":
|
| 95 |
-
"learning_rate": 0.
|
| 96 |
-
"loss": 3.
|
| 97 |
"step": 600
|
| 98 |
},
|
| 99 |
{
|
| 100 |
"epoch": 8.335664335664335,
|
| 101 |
-
"eval_cer": 0.
|
| 102 |
-
"eval_loss": 3.
|
| 103 |
-
"eval_runtime": 6.
|
| 104 |
-
"eval_samples_per_second": 46.
|
| 105 |
-
"eval_steps_per_second": 5.
|
| 106 |
"step": 600
|
| 107 |
},
|
| 108 |
{
|
| 109 |
"epoch": 9.727272727272727,
|
| 110 |
-
"grad_norm":
|
| 111 |
-
"learning_rate": 0.
|
| 112 |
-
"loss": 3.
|
| 113 |
"step": 700
|
| 114 |
},
|
| 115 |
{
|
| 116 |
"epoch": 9.727272727272727,
|
| 117 |
-
"eval_cer": 0.
|
| 118 |
-
"eval_loss": 3.
|
| 119 |
-
"eval_runtime": 6.
|
| 120 |
-
"eval_samples_per_second": 46.
|
| 121 |
-
"eval_steps_per_second": 5.
|
| 122 |
"step": 700
|
| 123 |
},
|
| 124 |
{
|
| 125 |
"epoch": 11.111888111888112,
|
| 126 |
-
"grad_norm": 0.
|
| 127 |
-
"learning_rate": 0.
|
| 128 |
-
"loss": 3.
|
| 129 |
"step": 800
|
| 130 |
},
|
| 131 |
{
|
| 132 |
"epoch": 11.111888111888112,
|
| 133 |
-
"eval_cer": 0.
|
| 134 |
-
"eval_loss": 3.
|
| 135 |
-
"eval_runtime": 6.
|
| 136 |
-
"eval_samples_per_second":
|
| 137 |
-
"eval_steps_per_second": 5.
|
| 138 |
"step": 800
|
| 139 |
},
|
| 140 |
{
|
| 141 |
"epoch": 12.503496503496503,
|
| 142 |
-
"grad_norm":
|
| 143 |
-
"learning_rate": 0.
|
| 144 |
-
"loss": 3.
|
| 145 |
"step": 900
|
| 146 |
},
|
| 147 |
{
|
| 148 |
"epoch": 12.503496503496503,
|
| 149 |
-
"eval_cer": 0.
|
| 150 |
-
"eval_loss":
|
| 151 |
-
"eval_runtime": 6.
|
| 152 |
-
"eval_samples_per_second": 46.
|
| 153 |
-
"eval_steps_per_second": 5.
|
| 154 |
"step": 900
|
| 155 |
},
|
| 156 |
{
|
| 157 |
"epoch": 13.895104895104895,
|
| 158 |
-
"grad_norm":
|
| 159 |
-
"learning_rate": 0.
|
| 160 |
-
"loss":
|
| 161 |
"step": 1000
|
| 162 |
},
|
| 163 |
{
|
| 164 |
"epoch": 13.895104895104895,
|
| 165 |
-
"eval_cer": 0.
|
| 166 |
-
"eval_loss":
|
| 167 |
-
"eval_runtime": 6.
|
| 168 |
-
"eval_samples_per_second": 47.
|
| 169 |
-
"eval_steps_per_second": 5.
|
| 170 |
"step": 1000
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 15.27972027972028,
|
| 174 |
-
"grad_norm":
|
| 175 |
-
"learning_rate": 0.
|
| 176 |
-
"loss":
|
| 177 |
"step": 1100
|
| 178 |
},
|
| 179 |
{
|
| 180 |
"epoch": 15.27972027972028,
|
| 181 |
-
"eval_cer": 0.
|
| 182 |
-
"eval_loss":
|
| 183 |
-
"eval_runtime": 6.
|
| 184 |
-
"eval_samples_per_second": 46.
|
| 185 |
-
"eval_steps_per_second": 5.
|
| 186 |
"step": 1100
|
| 187 |
},
|
| 188 |
{
|
| 189 |
"epoch": 16.67132867132867,
|
| 190 |
-
"grad_norm":
|
| 191 |
-
"learning_rate": 0.
|
| 192 |
-
"loss":
|
| 193 |
"step": 1200
|
| 194 |
},
|
| 195 |
{
|
| 196 |
"epoch": 16.67132867132867,
|
| 197 |
-
"eval_cer": 0.
|
| 198 |
-
"eval_loss":
|
| 199 |
-
"eval_runtime": 6.
|
| 200 |
-
"eval_samples_per_second":
|
| 201 |
-
"eval_steps_per_second": 5.
|
| 202 |
"step": 1200
|
| 203 |
},
|
| 204 |
{
|
| 205 |
"epoch": 18.055944055944057,
|
| 206 |
-
"grad_norm":
|
| 207 |
-
"learning_rate": 0.
|
| 208 |
-
"loss":
|
| 209 |
"step": 1300
|
| 210 |
},
|
| 211 |
{
|
| 212 |
"epoch": 18.055944055944057,
|
| 213 |
-
"eval_cer": 0.
|
| 214 |
-
"eval_loss":
|
| 215 |
-
"eval_runtime": 6.
|
| 216 |
-
"eval_samples_per_second": 46.
|
| 217 |
-
"eval_steps_per_second": 5.
|
| 218 |
"step": 1300
|
| 219 |
},
|
| 220 |
{
|
| 221 |
"epoch": 19.447552447552447,
|
| 222 |
-
"grad_norm":
|
| 223 |
-
"learning_rate": 0.
|
| 224 |
-
"loss":
|
| 225 |
"step": 1400
|
| 226 |
},
|
| 227 |
{
|
| 228 |
"epoch": 19.447552447552447,
|
| 229 |
-
"eval_cer": 0.
|
| 230 |
-
"eval_loss":
|
| 231 |
-
"eval_runtime": 6.
|
| 232 |
-
"eval_samples_per_second": 46.
|
| 233 |
-
"eval_steps_per_second": 5.
|
| 234 |
"step": 1400
|
| 235 |
},
|
| 236 |
{
|
| 237 |
"epoch": 20.83916083916084,
|
| 238 |
-
"grad_norm":
|
| 239 |
-
"learning_rate": 0.
|
| 240 |
-
"loss":
|
| 241 |
"step": 1500
|
| 242 |
},
|
| 243 |
{
|
| 244 |
"epoch": 20.83916083916084,
|
| 245 |
-
"eval_cer": 0.
|
| 246 |
-
"eval_loss":
|
| 247 |
-
"eval_runtime": 6.
|
| 248 |
-
"eval_samples_per_second":
|
| 249 |
-
"eval_steps_per_second": 5.
|
| 250 |
"step": 1500
|
| 251 |
},
|
| 252 |
{
|
| 253 |
"epoch": 22.223776223776223,
|
| 254 |
-
"grad_norm":
|
| 255 |
-
"learning_rate": 0.
|
| 256 |
-
"loss":
|
| 257 |
"step": 1600
|
| 258 |
},
|
| 259 |
{
|
| 260 |
"epoch": 22.223776223776223,
|
| 261 |
-
"eval_cer": 0.
|
| 262 |
-
"eval_loss":
|
| 263 |
-
"eval_runtime": 6.
|
| 264 |
-
"eval_samples_per_second":
|
| 265 |
-
"eval_steps_per_second": 5.
|
| 266 |
"step": 1600
|
| 267 |
},
|
| 268 |
{
|
| 269 |
"epoch": 23.615384615384617,
|
| 270 |
-
"grad_norm":
|
| 271 |
-
"learning_rate": 0.
|
| 272 |
-
"loss":
|
| 273 |
"step": 1700
|
| 274 |
},
|
| 275 |
{
|
| 276 |
"epoch": 23.615384615384617,
|
| 277 |
-
"eval_cer": 0.
|
| 278 |
-
"eval_loss":
|
| 279 |
-
"eval_runtime": 6.
|
| 280 |
-
"eval_samples_per_second": 46.
|
| 281 |
-
"eval_steps_per_second": 5.
|
| 282 |
"step": 1700
|
| 283 |
},
|
| 284 |
{
|
| 285 |
"epoch": 25.0,
|
| 286 |
-
"grad_norm":
|
| 287 |
-
"learning_rate": 0.
|
| 288 |
-
"loss":
|
| 289 |
"step": 1800
|
| 290 |
},
|
| 291 |
{
|
| 292 |
"epoch": 25.0,
|
| 293 |
-
"eval_cer": 0.
|
| 294 |
-
"eval_loss":
|
| 295 |
-
"eval_runtime": 6.
|
| 296 |
-
"eval_samples_per_second":
|
| 297 |
-
"eval_steps_per_second": 5.
|
| 298 |
"step": 1800
|
| 299 |
},
|
| 300 |
{
|
| 301 |
"epoch": 26.39160839160839,
|
| 302 |
-
"grad_norm":
|
| 303 |
-
"learning_rate": 0.
|
| 304 |
-
"loss":
|
| 305 |
"step": 1900
|
| 306 |
},
|
| 307 |
{
|
| 308 |
"epoch": 26.39160839160839,
|
| 309 |
-
"eval_cer": 0.
|
| 310 |
-
"eval_loss":
|
| 311 |
-
"eval_runtime": 6.
|
| 312 |
-
"eval_samples_per_second":
|
| 313 |
-
"eval_steps_per_second": 5.
|
| 314 |
"step": 1900
|
| 315 |
},
|
| 316 |
{
|
| 317 |
"epoch": 27.783216783216783,
|
| 318 |
-
"grad_norm":
|
| 319 |
-
"learning_rate": 0.
|
| 320 |
-
"loss":
|
| 321 |
"step": 2000
|
| 322 |
},
|
| 323 |
{
|
| 324 |
"epoch": 27.783216783216783,
|
| 325 |
-
"eval_cer": 0.
|
| 326 |
-
"eval_loss":
|
| 327 |
-
"eval_runtime": 6.
|
| 328 |
-
"eval_samples_per_second": 46.
|
| 329 |
-
"eval_steps_per_second": 5.
|
| 330 |
"step": 2000
|
| 331 |
},
|
| 332 |
{
|
| 333 |
"epoch": 29.167832167832167,
|
| 334 |
-
"grad_norm":
|
| 335 |
-
"learning_rate": 0.
|
| 336 |
-
"loss":
|
| 337 |
"step": 2100
|
| 338 |
},
|
| 339 |
{
|
| 340 |
"epoch": 29.167832167832167,
|
| 341 |
-
"eval_cer": 0.
|
| 342 |
-
"eval_loss": 1.
|
| 343 |
-
"eval_runtime": 6.
|
| 344 |
-
"eval_samples_per_second": 46.
|
| 345 |
-
"eval_steps_per_second": 5.
|
| 346 |
"step": 2100
|
| 347 |
},
|
| 348 |
{
|
| 349 |
"epoch": 30.55944055944056,
|
| 350 |
-
"grad_norm": 2.
|
| 351 |
-
"learning_rate": 0.
|
| 352 |
-
"loss":
|
| 353 |
"step": 2200
|
| 354 |
},
|
| 355 |
{
|
| 356 |
"epoch": 30.55944055944056,
|
| 357 |
-
"eval_cer": 0.
|
| 358 |
-
"eval_loss": 1.
|
| 359 |
-
"eval_runtime": 6.
|
| 360 |
-
"eval_samples_per_second":
|
| 361 |
-
"eval_steps_per_second": 5.
|
| 362 |
"step": 2200
|
| 363 |
},
|
| 364 |
{
|
| 365 |
"epoch": 31.95104895104895,
|
| 366 |
-
"grad_norm": 5.
|
| 367 |
-
"learning_rate": 0.
|
| 368 |
-
"loss":
|
| 369 |
"step": 2300
|
| 370 |
},
|
| 371 |
{
|
| 372 |
"epoch": 31.95104895104895,
|
| 373 |
-
"eval_cer": 0.
|
| 374 |
-
"eval_loss": 1.
|
| 375 |
-
"eval_runtime": 6.
|
| 376 |
-
"eval_samples_per_second":
|
| 377 |
-
"eval_steps_per_second": 5.
|
| 378 |
"step": 2300
|
| 379 |
},
|
| 380 |
{
|
| 381 |
"epoch": 33.33566433566433,
|
| 382 |
-
"grad_norm":
|
| 383 |
-
"learning_rate": 0.
|
| 384 |
-
"loss":
|
| 385 |
"step": 2400
|
| 386 |
},
|
| 387 |
{
|
| 388 |
"epoch": 33.33566433566433,
|
| 389 |
-
"eval_cer": 0.
|
| 390 |
-
"eval_loss": 1.
|
| 391 |
-
"eval_runtime": 6.
|
| 392 |
-
"eval_samples_per_second":
|
| 393 |
-
"eval_steps_per_second": 5.
|
| 394 |
"step": 2400
|
| 395 |
},
|
| 396 |
{
|
| 397 |
"epoch": 34.72727272727273,
|
| 398 |
-
"grad_norm":
|
| 399 |
-
"learning_rate":
|
| 400 |
-
"loss":
|
| 401 |
"step": 2500
|
| 402 |
},
|
| 403 |
{
|
| 404 |
"epoch": 34.72727272727273,
|
| 405 |
-
"eval_cer": 0.
|
| 406 |
-
"eval_loss": 1.
|
| 407 |
-
"eval_runtime": 6.
|
| 408 |
-
"eval_samples_per_second":
|
| 409 |
-
"eval_steps_per_second": 5.
|
| 410 |
"step": 2500
|
| 411 |
},
|
| 412 |
{
|
| 413 |
"epoch": 36.11188811188811,
|
| 414 |
-
"grad_norm":
|
| 415 |
-
"learning_rate":
|
| 416 |
-
"loss":
|
| 417 |
"step": 2600
|
| 418 |
},
|
| 419 |
{
|
| 420 |
"epoch": 36.11188811188811,
|
| 421 |
-
"eval_cer": 0.
|
| 422 |
-
"eval_loss": 1.
|
| 423 |
-
"eval_runtime": 6.
|
| 424 |
-
"eval_samples_per_second": 47.
|
| 425 |
-
"eval_steps_per_second": 5.
|
| 426 |
"step": 2600
|
| 427 |
},
|
| 428 |
{
|
| 429 |
"epoch": 37.50349650349651,
|
| 430 |
-
"grad_norm":
|
| 431 |
-
"learning_rate":
|
| 432 |
-
"loss":
|
| 433 |
"step": 2700
|
| 434 |
},
|
| 435 |
{
|
| 436 |
"epoch": 37.50349650349651,
|
| 437 |
-
"eval_cer": 0.
|
| 438 |
-
"eval_loss": 1.
|
| 439 |
-
"eval_runtime": 6.
|
| 440 |
-
"eval_samples_per_second": 47.
|
| 441 |
-
"eval_steps_per_second": 5.
|
| 442 |
"step": 2700
|
| 443 |
},
|
| 444 |
{
|
| 445 |
"epoch": 38.89510489510489,
|
| 446 |
-
"grad_norm":
|
| 447 |
-
"learning_rate":
|
| 448 |
-
"loss":
|
| 449 |
"step": 2800
|
| 450 |
},
|
| 451 |
{
|
| 452 |
"epoch": 38.89510489510489,
|
| 453 |
-
"eval_cer": 0.
|
| 454 |
-
"eval_loss": 1.
|
| 455 |
-
"eval_runtime": 6.
|
| 456 |
-
"eval_samples_per_second":
|
| 457 |
-
"eval_steps_per_second": 5.
|
| 458 |
"step": 2800
|
| 459 |
},
|
| 460 |
{
|
| 461 |
"epoch": 40.27972027972028,
|
| 462 |
-
"grad_norm": 2.
|
| 463 |
-
"learning_rate":
|
| 464 |
-
"loss":
|
| 465 |
"step": 2900
|
| 466 |
},
|
| 467 |
{
|
| 468 |
"epoch": 40.27972027972028,
|
| 469 |
-
"eval_cer": 0.
|
| 470 |
-
"eval_loss": 1.
|
| 471 |
-
"eval_runtime": 6.
|
| 472 |
-
"eval_samples_per_second":
|
| 473 |
-
"eval_steps_per_second": 5.
|
| 474 |
"step": 2900
|
| 475 |
},
|
| 476 |
{
|
| 477 |
"epoch": 41.67132867132867,
|
| 478 |
-
"grad_norm":
|
| 479 |
-
"learning_rate":
|
| 480 |
-
"loss":
|
| 481 |
"step": 3000
|
| 482 |
},
|
| 483 |
{
|
| 484 |
"epoch": 41.67132867132867,
|
| 485 |
-
"eval_cer": 0.
|
| 486 |
-
"eval_loss": 1.
|
| 487 |
-
"eval_runtime": 6.
|
| 488 |
-
"eval_samples_per_second": 46.
|
| 489 |
-
"eval_steps_per_second": 5.
|
| 490 |
"step": 3000
|
| 491 |
},
|
| 492 |
{
|
| 493 |
"epoch": 43.05594405594405,
|
| 494 |
-
"grad_norm": 1.
|
| 495 |
-
"learning_rate":
|
| 496 |
-
"loss":
|
| 497 |
"step": 3100
|
| 498 |
},
|
| 499 |
{
|
| 500 |
"epoch": 43.05594405594405,
|
| 501 |
-
"eval_cer": 0.
|
| 502 |
-
"eval_loss": 1.
|
| 503 |
-
"eval_runtime": 6.
|
| 504 |
-
"eval_samples_per_second": 46.
|
| 505 |
-
"eval_steps_per_second": 5.
|
| 506 |
"step": 3100
|
| 507 |
},
|
| 508 |
{
|
| 509 |
"epoch": 44.44755244755245,
|
| 510 |
-
"grad_norm": 4.
|
| 511 |
-
"learning_rate":
|
| 512 |
-
"loss": 0.
|
| 513 |
"step": 3200
|
| 514 |
},
|
| 515 |
{
|
| 516 |
"epoch": 44.44755244755245,
|
| 517 |
-
"eval_cer": 0.
|
| 518 |
-
"eval_loss": 1.
|
| 519 |
-
"eval_runtime": 6.
|
| 520 |
-
"eval_samples_per_second": 47.
|
| 521 |
-
"eval_steps_per_second": 5.
|
| 522 |
"step": 3200
|
| 523 |
},
|
| 524 |
{
|
| 525 |
"epoch": 45.83916083916084,
|
| 526 |
-
"grad_norm": 2.
|
| 527 |
-
"learning_rate":
|
| 528 |
-
"loss": 0.
|
| 529 |
"step": 3300
|
| 530 |
},
|
| 531 |
{
|
| 532 |
"epoch": 45.83916083916084,
|
| 533 |
-
"eval_cer": 0.
|
| 534 |
-
"eval_loss": 1.
|
| 535 |
-
"eval_runtime": 6.
|
| 536 |
-
"eval_samples_per_second": 46.
|
| 537 |
-
"eval_steps_per_second": 5.
|
| 538 |
"step": 3300
|
| 539 |
},
|
| 540 |
{
|
| 541 |
"epoch": 47.22377622377623,
|
| 542 |
-
"grad_norm":
|
| 543 |
-
"learning_rate":
|
| 544 |
-
"loss": 0.
|
| 545 |
"step": 3400
|
| 546 |
},
|
| 547 |
{
|
| 548 |
"epoch": 47.22377622377623,
|
| 549 |
-
"eval_cer": 0.
|
| 550 |
-
"eval_loss": 1.
|
| 551 |
-
"eval_runtime": 6.
|
| 552 |
-
"eval_samples_per_second": 47.
|
| 553 |
-
"eval_steps_per_second": 5.
|
| 554 |
"step": 3400
|
| 555 |
},
|
| 556 |
{
|
| 557 |
"epoch": 48.61538461538461,
|
| 558 |
-
"grad_norm": 2.
|
| 559 |
-
"learning_rate":
|
| 560 |
-
"loss": 0.
|
| 561 |
"step": 3500
|
| 562 |
},
|
| 563 |
{
|
| 564 |
"epoch": 48.61538461538461,
|
| 565 |
-
"eval_cer": 0.
|
| 566 |
-
"eval_loss": 1.
|
| 567 |
-
"eval_runtime": 6.
|
| 568 |
-
"eval_samples_per_second": 46.
|
| 569 |
-
"eval_steps_per_second": 5.
|
| 570 |
"step": 3500
|
| 571 |
},
|
| 572 |
{
|
| 573 |
-
"epoch":
|
| 574 |
-
"
|
| 575 |
-
"
|
| 576 |
-
"
|
| 577 |
-
"
|
| 578 |
-
"train_samples_per_second": 23.214,
|
| 579 |
-
"train_steps_per_second": 0.723
|
| 580 |
},
|
| 581 |
{
|
| 582 |
-
"epoch":
|
| 583 |
-
"eval_cer": 0.
|
| 584 |
-
"eval_loss": 1.
|
| 585 |
-
"eval_runtime": 6.
|
| 586 |
-
"eval_samples_per_second":
|
| 587 |
-
"eval_steps_per_second": 5.
|
| 588 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 589 |
}
|
| 590 |
],
|
| 591 |
"logging_steps": 100,
|
| 592 |
-
"max_steps":
|
| 593 |
"num_input_tokens_seen": 0,
|
| 594 |
-
"num_train_epochs":
|
| 595 |
"save_steps": 100,
|
| 596 |
"stateful_callbacks": {
|
| 597 |
"TrainerControl": {
|
|
@@ -605,7 +1181,7 @@
|
|
| 605 |
"attributes": {}
|
| 606 |
}
|
| 607 |
},
|
| 608 |
-
"total_flos":
|
| 609 |
"train_batch_size": 16,
|
| 610 |
"trial_name": null,
|
| 611 |
"trial_params": null
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 98.61538461538461,
|
| 6 |
"eval_steps": 100,
|
| 7 |
+
"global_step": 7100,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
| 11 |
"log_history": [
|
| 12 |
{
|
| 13 |
"epoch": 1.3916083916083917,
|
| 14 |
+
"grad_norm": 0.7105618119239807,
|
| 15 |
"learning_rate": 0.000285,
|
| 16 |
+
"loss": 11.1682,
|
| 17 |
"step": 100
|
| 18 |
},
|
| 19 |
{
|
| 20 |
"epoch": 1.3916083916083917,
|
| 21 |
+
"eval_cer": 0.9903044405662207,
|
| 22 |
+
"eval_loss": 3.855257987976074,
|
| 23 |
+
"eval_runtime": 6.2271,
|
| 24 |
+
"eval_samples_per_second": 45.768,
|
| 25 |
+
"eval_steps_per_second": 5.781,
|
| 26 |
"step": 100
|
| 27 |
},
|
| 28 |
{
|
| 29 |
"epoch": 2.7832167832167833,
|
| 30 |
+
"grad_norm": 1.5593620538711548,
|
| 31 |
+
"learning_rate": 0.0002959285714285714,
|
| 32 |
+
"loss": 3.9311,
|
| 33 |
"step": 200
|
| 34 |
},
|
| 35 |
{
|
| 36 |
"epoch": 2.7832167832167833,
|
| 37 |
+
"eval_cer": 0.9903044405662207,
|
| 38 |
+
"eval_loss": 3.8240556716918945,
|
| 39 |
+
"eval_runtime": 6.1912,
|
| 40 |
+
"eval_samples_per_second": 46.033,
|
| 41 |
+
"eval_steps_per_second": 5.815,
|
| 42 |
"step": 200
|
| 43 |
},
|
| 44 |
{
|
| 45 |
"epoch": 4.1678321678321675,
|
| 46 |
+
"grad_norm": 1.5898423194885254,
|
| 47 |
+
"learning_rate": 0.0002916428571428571,
|
| 48 |
+
"loss": 3.8623,
|
| 49 |
"step": 300
|
| 50 |
},
|
| 51 |
{
|
| 52 |
"epoch": 4.1678321678321675,
|
| 53 |
+
"eval_cer": 0.9903044405662207,
|
| 54 |
+
"eval_loss": 3.7760262489318848,
|
| 55 |
+
"eval_runtime": 6.1779,
|
| 56 |
+
"eval_samples_per_second": 46.132,
|
| 57 |
+
"eval_steps_per_second": 5.827,
|
| 58 |
"step": 300
|
| 59 |
},
|
| 60 |
{
|
| 61 |
"epoch": 5.559440559440559,
|
| 62 |
+
"grad_norm": 2.1614158153533936,
|
| 63 |
+
"learning_rate": 0.00028735714285714286,
|
| 64 |
+
"loss": 3.7693,
|
| 65 |
"step": 400
|
| 66 |
},
|
| 67 |
{
|
| 68 |
"epoch": 5.559440559440559,
|
| 69 |
"eval_cer": 0.9903044405662207,
|
| 70 |
+
"eval_loss": 3.668625831604004,
|
| 71 |
+
"eval_runtime": 6.1599,
|
| 72 |
+
"eval_samples_per_second": 46.267,
|
| 73 |
+
"eval_steps_per_second": 5.844,
|
| 74 |
"step": 400
|
| 75 |
},
|
| 76 |
{
|
| 77 |
"epoch": 6.951048951048951,
|
| 78 |
+
"grad_norm": 1.3560220003128052,
|
| 79 |
+
"learning_rate": 0.00028307142857142854,
|
| 80 |
+
"loss": 3.671,
|
| 81 |
"step": 500
|
| 82 |
},
|
| 83 |
{
|
| 84 |
"epoch": 6.951048951048951,
|
| 85 |
+
"eval_cer": 0.9893348846228427,
|
| 86 |
+
"eval_loss": 3.590017080307007,
|
| 87 |
+
"eval_runtime": 6.1396,
|
| 88 |
+
"eval_samples_per_second": 46.42,
|
| 89 |
+
"eval_steps_per_second": 5.864,
|
| 90 |
"step": 500
|
| 91 |
},
|
| 92 |
{
|
| 93 |
"epoch": 8.335664335664335,
|
| 94 |
+
"grad_norm": 1.1032301187515259,
|
| 95 |
+
"learning_rate": 0.00027878571428571427,
|
| 96 |
+
"loss": 3.5618,
|
| 97 |
"step": 600
|
| 98 |
},
|
| 99 |
{
|
| 100 |
"epoch": 8.335664335664335,
|
| 101 |
+
"eval_cer": 0.9713011440760132,
|
| 102 |
+
"eval_loss": 3.5168509483337402,
|
| 103 |
+
"eval_runtime": 6.166,
|
| 104 |
+
"eval_samples_per_second": 46.221,
|
| 105 |
+
"eval_steps_per_second": 5.838,
|
| 106 |
"step": 600
|
| 107 |
},
|
| 108 |
{
|
| 109 |
"epoch": 9.727272727272727,
|
| 110 |
+
"grad_norm": 1.4139310121536255,
|
| 111 |
+
"learning_rate": 0.0002745,
|
| 112 |
+
"loss": 3.4994,
|
| 113 |
"step": 700
|
| 114 |
},
|
| 115 |
{
|
| 116 |
"epoch": 9.727272727272727,
|
| 117 |
+
"eval_cer": 0.9699437657552841,
|
| 118 |
+
"eval_loss": 3.3551578521728516,
|
| 119 |
+
"eval_runtime": 6.1552,
|
| 120 |
+
"eval_samples_per_second": 46.302,
|
| 121 |
+
"eval_steps_per_second": 5.849,
|
| 122 |
"step": 700
|
| 123 |
},
|
| 124 |
{
|
| 125 |
"epoch": 11.111888111888112,
|
| 126 |
+
"grad_norm": 0.8385369777679443,
|
| 127 |
+
"learning_rate": 0.0002702142857142857,
|
| 128 |
+
"loss": 3.3323,
|
| 129 |
"step": 800
|
| 130 |
},
|
| 131 |
{
|
| 132 |
"epoch": 11.111888111888112,
|
| 133 |
+
"eval_cer": 0.954043048283886,
|
| 134 |
+
"eval_loss": 3.138484001159668,
|
| 135 |
+
"eval_runtime": 6.0631,
|
| 136 |
+
"eval_samples_per_second": 47.005,
|
| 137 |
+
"eval_steps_per_second": 5.938,
|
| 138 |
"step": 800
|
| 139 |
},
|
| 140 |
{
|
| 141 |
"epoch": 12.503496503496503,
|
| 142 |
+
"grad_norm": 3.5932364463806152,
|
| 143 |
+
"learning_rate": 0.0002659285714285714,
|
| 144 |
+
"loss": 3.163,
|
| 145 |
"step": 900
|
| 146 |
},
|
| 147 |
{
|
| 148 |
"epoch": 12.503496503496503,
|
| 149 |
+
"eval_cer": 0.9185573007562536,
|
| 150 |
+
"eval_loss": 2.9224491119384766,
|
| 151 |
+
"eval_runtime": 6.0879,
|
| 152 |
+
"eval_samples_per_second": 46.814,
|
| 153 |
+
"eval_steps_per_second": 5.913,
|
| 154 |
"step": 900
|
| 155 |
},
|
| 156 |
{
|
| 157 |
"epoch": 13.895104895104895,
|
| 158 |
+
"grad_norm": 1.8399698734283447,
|
| 159 |
+
"learning_rate": 0.0002616428571428571,
|
| 160 |
+
"loss": 2.7901,
|
| 161 |
"step": 1000
|
| 162 |
},
|
| 163 |
{
|
| 164 |
"epoch": 13.895104895104895,
|
| 165 |
+
"eval_cer": 0.7828194686833431,
|
| 166 |
+
"eval_loss": 2.180238723754883,
|
| 167 |
+
"eval_runtime": 6.0259,
|
| 168 |
+
"eval_samples_per_second": 47.296,
|
| 169 |
+
"eval_steps_per_second": 5.974,
|
| 170 |
"step": 1000
|
| 171 |
},
|
| 172 |
{
|
| 173 |
"epoch": 15.27972027972028,
|
| 174 |
+
"grad_norm": 1.625369668006897,
|
| 175 |
+
"learning_rate": 0.00025735714285714283,
|
| 176 |
+
"loss": 2.3425,
|
| 177 |
"step": 1100
|
| 178 |
},
|
| 179 |
{
|
| 180 |
"epoch": 15.27972027972028,
|
| 181 |
+
"eval_cer": 0.6528989722707,
|
| 182 |
+
"eval_loss": 1.8405641317367554,
|
| 183 |
+
"eval_runtime": 6.0991,
|
| 184 |
+
"eval_samples_per_second": 46.728,
|
| 185 |
+
"eval_steps_per_second": 5.903,
|
| 186 |
"step": 1100
|
| 187 |
},
|
| 188 |
{
|
| 189 |
"epoch": 16.67132867132867,
|
| 190 |
+
"grad_norm": 2.625293016433716,
|
| 191 |
+
"learning_rate": 0.0002530714285714285,
|
| 192 |
+
"loss": 2.0608,
|
| 193 |
"step": 1200
|
| 194 |
},
|
| 195 |
{
|
| 196 |
"epoch": 16.67132867132867,
|
| 197 |
+
"eval_cer": 0.6329261198371146,
|
| 198 |
+
"eval_loss": 1.6505399942398071,
|
| 199 |
+
"eval_runtime": 6.1263,
|
| 200 |
+
"eval_samples_per_second": 46.52,
|
| 201 |
+
"eval_steps_per_second": 5.876,
|
| 202 |
"step": 1200
|
| 203 |
},
|
| 204 |
{
|
| 205 |
"epoch": 18.055944055944057,
|
| 206 |
+
"grad_norm": 1.92220139503479,
|
| 207 |
+
"learning_rate": 0.00024878571428571425,
|
| 208 |
+
"loss": 1.8813,
|
| 209 |
"step": 1300
|
| 210 |
},
|
| 211 |
{
|
| 212 |
"epoch": 18.055944055944057,
|
| 213 |
+
"eval_cer": 0.5714562730269537,
|
| 214 |
+
"eval_loss": 1.4768792390823364,
|
| 215 |
+
"eval_runtime": 6.1089,
|
| 216 |
+
"eval_samples_per_second": 46.653,
|
| 217 |
+
"eval_steps_per_second": 5.893,
|
| 218 |
"step": 1300
|
| 219 |
},
|
| 220 |
{
|
| 221 |
"epoch": 19.447552447552447,
|
| 222 |
+
"grad_norm": 3.1366982460021973,
|
| 223 |
+
"learning_rate": 0.0002445,
|
| 224 |
+
"loss": 1.6705,
|
| 225 |
"step": 1400
|
| 226 |
},
|
| 227 |
{
|
| 228 |
"epoch": 19.447552447552447,
|
| 229 |
+
"eval_cer": 0.5580764010083382,
|
| 230 |
+
"eval_loss": 1.479285478591919,
|
| 231 |
+
"eval_runtime": 6.1149,
|
| 232 |
+
"eval_samples_per_second": 46.608,
|
| 233 |
+
"eval_steps_per_second": 5.887,
|
| 234 |
"step": 1400
|
| 235 |
},
|
| 236 |
{
|
| 237 |
"epoch": 20.83916083916084,
|
| 238 |
+
"grad_norm": 2.116931200027466,
|
| 239 |
+
"learning_rate": 0.0002402142857142857,
|
| 240 |
+
"loss": 1.558,
|
| 241 |
"step": 1500
|
| 242 |
},
|
| 243 |
{
|
| 244 |
"epoch": 20.83916083916084,
|
| 245 |
+
"eval_cer": 0.4969943765755284,
|
| 246 |
+
"eval_loss": 1.3079291582107544,
|
| 247 |
+
"eval_runtime": 6.0774,
|
| 248 |
+
"eval_samples_per_second": 46.895,
|
| 249 |
+
"eval_steps_per_second": 5.924,
|
| 250 |
"step": 1500
|
| 251 |
},
|
| 252 |
{
|
| 253 |
"epoch": 22.223776223776223,
|
| 254 |
+
"grad_norm": 4.369190692901611,
|
| 255 |
+
"learning_rate": 0.00023592857142857142,
|
| 256 |
+
"loss": 1.4213,
|
| 257 |
"step": 1600
|
| 258 |
},
|
| 259 |
{
|
| 260 |
"epoch": 22.223776223776223,
|
| 261 |
+
"eval_cer": 0.49466744231142135,
|
| 262 |
+
"eval_loss": 1.3551599979400635,
|
| 263 |
+
"eval_runtime": 6.1085,
|
| 264 |
+
"eval_samples_per_second": 46.657,
|
| 265 |
+
"eval_steps_per_second": 5.893,
|
| 266 |
"step": 1600
|
| 267 |
},
|
| 268 |
{
|
| 269 |
"epoch": 23.615384615384617,
|
| 270 |
+
"grad_norm": 2.938127279281616,
|
| 271 |
+
"learning_rate": 0.00023164285714285713,
|
| 272 |
+
"loss": 1.3122,
|
| 273 |
"step": 1700
|
| 274 |
},
|
| 275 |
{
|
| 276 |
"epoch": 23.615384615384617,
|
| 277 |
+
"eval_cer": 0.4355245297653675,
|
| 278 |
+
"eval_loss": 1.236782431602478,
|
| 279 |
+
"eval_runtime": 6.1138,
|
| 280 |
+
"eval_samples_per_second": 46.616,
|
| 281 |
+
"eval_steps_per_second": 5.888,
|
| 282 |
"step": 1700
|
| 283 |
},
|
| 284 |
{
|
| 285 |
"epoch": 25.0,
|
| 286 |
+
"grad_norm": 4.156201362609863,
|
| 287 |
+
"learning_rate": 0.00022735714285714286,
|
| 288 |
+
"loss": 1.2303,
|
| 289 |
"step": 1800
|
| 290 |
},
|
| 291 |
{
|
| 292 |
"epoch": 25.0,
|
| 293 |
+
"eval_cer": 0.4347488850106651,
|
| 294 |
+
"eval_loss": 1.210758924484253,
|
| 295 |
+
"eval_runtime": 6.0871,
|
| 296 |
+
"eval_samples_per_second": 46.82,
|
| 297 |
+
"eval_steps_per_second": 5.914,
|
| 298 |
"step": 1800
|
| 299 |
},
|
| 300 |
{
|
| 301 |
"epoch": 26.39160839160839,
|
| 302 |
+
"grad_norm": 3.738050937652588,
|
| 303 |
+
"learning_rate": 0.00022307142857142854,
|
| 304 |
+
"loss": 1.1152,
|
| 305 |
"step": 1900
|
| 306 |
},
|
| 307 |
{
|
| 308 |
"epoch": 26.39160839160839,
|
| 309 |
+
"eval_cer": 0.4306767500484778,
|
| 310 |
+
"eval_loss": 1.2177391052246094,
|
| 311 |
+
"eval_runtime": 6.1061,
|
| 312 |
+
"eval_samples_per_second": 46.674,
|
| 313 |
+
"eval_steps_per_second": 5.896,
|
| 314 |
"step": 1900
|
| 315 |
},
|
| 316 |
{
|
| 317 |
"epoch": 27.783216783216783,
|
| 318 |
+
"grad_norm": 3.721745014190674,
|
| 319 |
+
"learning_rate": 0.00021878571428571428,
|
| 320 |
+
"loss": 1.0441,
|
| 321 |
"step": 2000
|
| 322 |
},
|
| 323 |
{
|
| 324 |
"epoch": 27.783216783216783,
|
| 325 |
+
"eval_cer": 0.4291254605390731,
|
| 326 |
+
"eval_loss": 1.3235960006713867,
|
| 327 |
+
"eval_runtime": 6.0708,
|
| 328 |
+
"eval_samples_per_second": 46.946,
|
| 329 |
+
"eval_steps_per_second": 5.93,
|
| 330 |
"step": 2000
|
| 331 |
},
|
| 332 |
{
|
| 333 |
"epoch": 29.167832167832167,
|
| 334 |
+
"grad_norm": 2.746555805206299,
|
| 335 |
+
"learning_rate": 0.00021449999999999998,
|
| 336 |
+
"loss": 0.9626,
|
| 337 |
"step": 2100
|
| 338 |
},
|
| 339 |
{
|
| 340 |
"epoch": 29.167832167832167,
|
| 341 |
+
"eval_cer": 0.4157455885204576,
|
| 342 |
+
"eval_loss": 1.2737609148025513,
|
| 343 |
+
"eval_runtime": 6.1093,
|
| 344 |
+
"eval_samples_per_second": 46.65,
|
| 345 |
+
"eval_steps_per_second": 5.893,
|
| 346 |
"step": 2100
|
| 347 |
},
|
| 348 |
{
|
| 349 |
"epoch": 30.55944055944056,
|
| 350 |
+
"grad_norm": 2.8237345218658447,
|
| 351 |
+
"learning_rate": 0.0002102142857142857,
|
| 352 |
+
"loss": 0.8987,
|
| 353 |
"step": 2200
|
| 354 |
},
|
| 355 |
{
|
| 356 |
"epoch": 30.55944055944056,
|
| 357 |
+
"eval_cer": 0.4190420787279426,
|
| 358 |
+
"eval_loss": 1.2683167457580566,
|
| 359 |
+
"eval_runtime": 6.1368,
|
| 360 |
+
"eval_samples_per_second": 46.441,
|
| 361 |
+
"eval_steps_per_second": 5.866,
|
| 362 |
"step": 2200
|
| 363 |
},
|
| 364 |
{
|
| 365 |
"epoch": 31.95104895104895,
|
| 366 |
+
"grad_norm": 5.561631679534912,
|
| 367 |
+
"learning_rate": 0.0002059285714285714,
|
| 368 |
+
"loss": 0.8367,
|
| 369 |
"step": 2300
|
| 370 |
},
|
| 371 |
{
|
| 372 |
"epoch": 31.95104895104895,
|
| 373 |
+
"eval_cer": 0.41438821019972855,
|
| 374 |
+
"eval_loss": 1.2570189237594604,
|
| 375 |
+
"eval_runtime": 6.095,
|
| 376 |
+
"eval_samples_per_second": 46.76,
|
| 377 |
+
"eval_steps_per_second": 5.906,
|
| 378 |
"step": 2300
|
| 379 |
},
|
| 380 |
{
|
| 381 |
"epoch": 33.33566433566433,
|
| 382 |
+
"grad_norm": 1.7682024240493774,
|
| 383 |
+
"learning_rate": 0.00020164285714285713,
|
| 384 |
+
"loss": 0.7617,
|
| 385 |
"step": 2400
|
| 386 |
},
|
| 387 |
{
|
| 388 |
"epoch": 33.33566433566433,
|
| 389 |
+
"eval_cer": 0.3876284661624976,
|
| 390 |
+
"eval_loss": 1.233074426651001,
|
| 391 |
+
"eval_runtime": 6.0871,
|
| 392 |
+
"eval_samples_per_second": 46.821,
|
| 393 |
+
"eval_steps_per_second": 5.914,
|
| 394 |
"step": 2400
|
| 395 |
},
|
| 396 |
{
|
| 397 |
"epoch": 34.72727272727273,
|
| 398 |
+
"grad_norm": 3.293351888656616,
|
| 399 |
+
"learning_rate": 0.00019735714285714284,
|
| 400 |
+
"loss": 0.7069,
|
| 401 |
"step": 2500
|
| 402 |
},
|
| 403 |
{
|
| 404 |
"epoch": 34.72727272727273,
|
| 405 |
+
"eval_cer": 0.40372309482257124,
|
| 406 |
+
"eval_loss": 1.328414797782898,
|
| 407 |
+
"eval_runtime": 6.0987,
|
| 408 |
+
"eval_samples_per_second": 46.731,
|
| 409 |
+
"eval_steps_per_second": 5.903,
|
| 410 |
"step": 2500
|
| 411 |
},
|
| 412 |
{
|
| 413 |
"epoch": 36.11188811188811,
|
| 414 |
+
"grad_norm": 1.8948358297348022,
|
| 415 |
+
"learning_rate": 0.00019307142857142854,
|
| 416 |
+
"loss": 0.6874,
|
| 417 |
"step": 2600
|
| 418 |
},
|
| 419 |
{
|
| 420 |
"epoch": 36.11188811188811,
|
| 421 |
+
"eval_cer": 0.38181113050223,
|
| 422 |
+
"eval_loss": 1.2947708368301392,
|
| 423 |
+
"eval_runtime": 6.0589,
|
| 424 |
+
"eval_samples_per_second": 47.038,
|
| 425 |
+
"eval_steps_per_second": 5.942,
|
| 426 |
"step": 2600
|
| 427 |
},
|
| 428 |
{
|
| 429 |
"epoch": 37.50349650349651,
|
| 430 |
+
"grad_norm": 2.3135173320770264,
|
| 431 |
+
"learning_rate": 0.00018878571428571428,
|
| 432 |
+
"loss": 0.6615,
|
| 433 |
"step": 2700
|
| 434 |
},
|
| 435 |
{
|
| 436 |
"epoch": 37.50349650349651,
|
| 437 |
+
"eval_cer": 0.3977118479736281,
|
| 438 |
+
"eval_loss": 1.299822211265564,
|
| 439 |
+
"eval_runtime": 6.0553,
|
| 440 |
+
"eval_samples_per_second": 47.066,
|
| 441 |
+
"eval_steps_per_second": 5.945,
|
| 442 |
"step": 2700
|
| 443 |
},
|
| 444 |
{
|
| 445 |
"epoch": 38.89510489510489,
|
| 446 |
+
"grad_norm": 8.07669448852539,
|
| 447 |
+
"learning_rate": 0.00018449999999999999,
|
| 448 |
+
"loss": 0.6086,
|
| 449 |
"step": 2800
|
| 450 |
},
|
| 451 |
{
|
| 452 |
"epoch": 38.89510489510489,
|
| 453 |
+
"eval_cer": 0.3757998836532868,
|
| 454 |
+
"eval_loss": 1.3369208574295044,
|
| 455 |
+
"eval_runtime": 6.0593,
|
| 456 |
+
"eval_samples_per_second": 47.035,
|
| 457 |
+
"eval_steps_per_second": 5.941,
|
| 458 |
"step": 2800
|
| 459 |
},
|
| 460 |
{
|
| 461 |
"epoch": 40.27972027972028,
|
| 462 |
+
"grad_norm": 2.3282470703125,
|
| 463 |
+
"learning_rate": 0.00018021428571428572,
|
| 464 |
+
"loss": 0.5804,
|
| 465 |
"step": 2900
|
| 466 |
},
|
| 467 |
{
|
| 468 |
"epoch": 40.27972027972028,
|
| 469 |
+
"eval_cer": 0.38375024238898586,
|
| 470 |
+
"eval_loss": 1.2814927101135254,
|
| 471 |
+
"eval_runtime": 6.0834,
|
| 472 |
+
"eval_samples_per_second": 46.849,
|
| 473 |
+
"eval_steps_per_second": 5.918,
|
| 474 |
"step": 2900
|
| 475 |
},
|
| 476 |
{
|
| 477 |
"epoch": 41.67132867132867,
|
| 478 |
+
"grad_norm": 5.158154010772705,
|
| 479 |
+
"learning_rate": 0.0001759285714285714,
|
| 480 |
+
"loss": 0.548,
|
| 481 |
"step": 3000
|
| 482 |
},
|
| 483 |
{
|
| 484 |
"epoch": 41.67132867132867,
|
| 485 |
+
"eval_cer": 0.37657552840798914,
|
| 486 |
+
"eval_loss": 1.3390411138534546,
|
| 487 |
+
"eval_runtime": 6.0871,
|
| 488 |
+
"eval_samples_per_second": 46.82,
|
| 489 |
+
"eval_steps_per_second": 5.914,
|
| 490 |
"step": 3000
|
| 491 |
},
|
| 492 |
{
|
| 493 |
"epoch": 43.05594405594405,
|
| 494 |
+
"grad_norm": 1.2800214290618896,
|
| 495 |
+
"learning_rate": 0.00017164285714285713,
|
| 496 |
+
"loss": 0.5239,
|
| 497 |
"step": 3100
|
| 498 |
},
|
| 499 |
{
|
| 500 |
"epoch": 43.05594405594405,
|
| 501 |
+
"eval_cer": 0.367267791351561,
|
| 502 |
+
"eval_loss": 1.257192850112915,
|
| 503 |
+
"eval_runtime": 6.0964,
|
| 504 |
+
"eval_samples_per_second": 46.749,
|
| 505 |
+
"eval_steps_per_second": 5.905,
|
| 506 |
"step": 3100
|
| 507 |
},
|
| 508 |
{
|
| 509 |
"epoch": 44.44755244755245,
|
| 510 |
+
"grad_norm": 4.9716010093688965,
|
| 511 |
+
"learning_rate": 0.00016735714285714284,
|
| 512 |
+
"loss": 0.4983,
|
| 513 |
"step": 3200
|
| 514 |
},
|
| 515 |
{
|
| 516 |
"epoch": 44.44755244755245,
|
| 517 |
+
"eval_cer": 0.3670738801628854,
|
| 518 |
+
"eval_loss": 1.295488715171814,
|
| 519 |
+
"eval_runtime": 6.034,
|
| 520 |
+
"eval_samples_per_second": 47.232,
|
| 521 |
+
"eval_steps_per_second": 5.966,
|
| 522 |
"step": 3200
|
| 523 |
},
|
| 524 |
{
|
| 525 |
"epoch": 45.83916083916084,
|
| 526 |
+
"grad_norm": 2.536774158477783,
|
| 527 |
+
"learning_rate": 0.0001631142857142857,
|
| 528 |
+
"loss": 0.4793,
|
| 529 |
"step": 3300
|
| 530 |
},
|
| 531 |
{
|
| 532 |
"epoch": 45.83916083916084,
|
| 533 |
+
"eval_cer": 0.372891215823153,
|
| 534 |
+
"eval_loss": 1.3562514781951904,
|
| 535 |
+
"eval_runtime": 6.0912,
|
| 536 |
+
"eval_samples_per_second": 46.789,
|
| 537 |
+
"eval_steps_per_second": 5.91,
|
| 538 |
"step": 3300
|
| 539 |
},
|
| 540 |
{
|
| 541 |
"epoch": 47.22377622377623,
|
| 542 |
+
"grad_norm": 3.6583845615386963,
|
| 543 |
+
"learning_rate": 0.00015882857142857142,
|
| 544 |
+
"loss": 0.438,
|
| 545 |
"step": 3400
|
| 546 |
},
|
| 547 |
{
|
| 548 |
"epoch": 47.22377622377623,
|
| 549 |
+
"eval_cer": 0.3915066899360093,
|
| 550 |
+
"eval_loss": 1.415280818939209,
|
| 551 |
+
"eval_runtime": 6.0618,
|
| 552 |
+
"eval_samples_per_second": 47.016,
|
| 553 |
+
"eval_steps_per_second": 5.939,
|
| 554 |
"step": 3400
|
| 555 |
},
|
| 556 |
{
|
| 557 |
"epoch": 48.61538461538461,
|
| 558 |
+
"grad_norm": 2.472052574157715,
|
| 559 |
+
"learning_rate": 0.00015454285714285712,
|
| 560 |
+
"loss": 0.4274,
|
| 561 |
"step": 3500
|
| 562 |
},
|
| 563 |
{
|
| 564 |
"epoch": 48.61538461538461,
|
| 565 |
+
"eval_cer": 0.36629823540818307,
|
| 566 |
+
"eval_loss": 1.319765567779541,
|
| 567 |
+
"eval_runtime": 6.1106,
|
| 568 |
+
"eval_samples_per_second": 46.64,
|
| 569 |
+
"eval_steps_per_second": 5.891,
|
| 570 |
"step": 3500
|
| 571 |
},
|
| 572 |
{
|
| 573 |
+
"epoch": 50.0,
|
| 574 |
+
"grad_norm": 7.336581707000732,
|
| 575 |
+
"learning_rate": 0.00015025714285714286,
|
| 576 |
+
"loss": 0.4064,
|
| 577 |
+
"step": 3600
|
|
|
|
|
|
|
| 578 |
},
|
| 579 |
{
|
| 580 |
+
"epoch": 50.0,
|
| 581 |
+
"eval_cer": 0.3814233081248788,
|
| 582 |
+
"eval_loss": 1.4350632429122925,
|
| 583 |
+
"eval_runtime": 6.098,
|
| 584 |
+
"eval_samples_per_second": 46.736,
|
| 585 |
+
"eval_steps_per_second": 5.904,
|
| 586 |
+
"step": 3600
|
| 587 |
+
},
|
| 588 |
+
{
|
| 589 |
+
"epoch": 51.39160839160839,
|
| 590 |
+
"grad_norm": 2.7026124000549316,
|
| 591 |
+
"learning_rate": 0.00014597142857142856,
|
| 592 |
+
"loss": 0.3812,
|
| 593 |
+
"step": 3700
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 51.39160839160839,
|
| 597 |
+
"eval_cer": 0.36203218925732017,
|
| 598 |
+
"eval_loss": 1.351439356803894,
|
| 599 |
+
"eval_runtime": 6.0572,
|
| 600 |
+
"eval_samples_per_second": 47.051,
|
| 601 |
+
"eval_steps_per_second": 5.943,
|
| 602 |
+
"step": 3700
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 52.78321678321678,
|
| 606 |
+
"grad_norm": 1.425048589706421,
|
| 607 |
+
"learning_rate": 0.00014168571428571427,
|
| 608 |
+
"loss": 0.3753,
|
| 609 |
+
"step": 3800
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 52.78321678321678,
|
| 613 |
+
"eval_cer": 0.3492340508047314,
|
| 614 |
+
"eval_loss": 1.3715204000473022,
|
| 615 |
+
"eval_runtime": 6.0504,
|
| 616 |
+
"eval_samples_per_second": 47.104,
|
| 617 |
+
"eval_steps_per_second": 5.95,
|
| 618 |
+
"step": 3800
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 54.16783216783217,
|
| 622 |
+
"grad_norm": 2.945066452026367,
|
| 623 |
+
"learning_rate": 0.0001374,
|
| 624 |
+
"loss": 0.3549,
|
| 625 |
+
"step": 3900
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 54.16783216783217,
|
| 629 |
+
"eval_cer": 0.36494085708745394,
|
| 630 |
+
"eval_loss": 1.4132966995239258,
|
| 631 |
+
"eval_runtime": 6.0971,
|
| 632 |
+
"eval_samples_per_second": 46.743,
|
| 633 |
+
"eval_steps_per_second": 5.904,
|
| 634 |
+
"step": 3900
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 55.55944055944056,
|
| 638 |
+
"grad_norm": 1.2087554931640625,
|
| 639 |
+
"learning_rate": 0.0001331142857142857,
|
| 640 |
+
"loss": 0.3262,
|
| 641 |
+
"step": 4000
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 55.55944055944056,
|
| 645 |
+
"eval_cer": 0.3573783207291061,
|
| 646 |
+
"eval_loss": 1.4259963035583496,
|
| 647 |
+
"eval_runtime": 6.0615,
|
| 648 |
+
"eval_samples_per_second": 47.018,
|
| 649 |
+
"eval_steps_per_second": 5.939,
|
| 650 |
+
"step": 4000
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 56.95104895104895,
|
| 654 |
+
"grad_norm": 6.813267230987549,
|
| 655 |
+
"learning_rate": 0.00012882857142857142,
|
| 656 |
+
"loss": 0.3296,
|
| 657 |
+
"step": 4100
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 56.95104895104895,
|
| 661 |
+
"eval_cer": 0.35524529765367463,
|
| 662 |
+
"eval_loss": 1.5134129524230957,
|
| 663 |
+
"eval_runtime": 6.0512,
|
| 664 |
+
"eval_samples_per_second": 47.098,
|
| 665 |
+
"eval_steps_per_second": 5.949,
|
| 666 |
+
"step": 4100
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 58.33566433566433,
|
| 670 |
+
"grad_norm": 1.582381248474121,
|
| 671 |
+
"learning_rate": 0.00012454285714285713,
|
| 672 |
+
"loss": 0.3136,
|
| 673 |
+
"step": 4200
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 58.33566433566433,
|
| 677 |
+
"eval_cer": 0.35873569904983515,
|
| 678 |
+
"eval_loss": 1.4695625305175781,
|
| 679 |
+
"eval_runtime": 6.0851,
|
| 680 |
+
"eval_samples_per_second": 46.836,
|
| 681 |
+
"eval_steps_per_second": 5.916,
|
| 682 |
+
"step": 4200
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 59.72727272727273,
|
| 686 |
+
"grad_norm": 0.9694690108299255,
|
| 687 |
+
"learning_rate": 0.00012025714285714285,
|
| 688 |
+
"loss": 0.3009,
|
| 689 |
+
"step": 4300
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 59.72727272727273,
|
| 693 |
+
"eval_cer": 0.3554392088423502,
|
| 694 |
+
"eval_loss": 1.432596206665039,
|
| 695 |
+
"eval_runtime": 6.0669,
|
| 696 |
+
"eval_samples_per_second": 46.976,
|
| 697 |
+
"eval_steps_per_second": 5.934,
|
| 698 |
+
"step": 4300
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 61.11188811188811,
|
| 702 |
+
"grad_norm": 1.6826478242874146,
|
| 703 |
+
"learning_rate": 0.00011597142857142855,
|
| 704 |
+
"loss": 0.2764,
|
| 705 |
+
"step": 4400
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 61.11188811188811,
|
| 709 |
+
"eval_cer": 0.35718440954043046,
|
| 710 |
+
"eval_loss": 1.4485613107681274,
|
| 711 |
+
"eval_runtime": 6.0638,
|
| 712 |
+
"eval_samples_per_second": 47.0,
|
| 713 |
+
"eval_steps_per_second": 5.937,
|
| 714 |
+
"step": 4400
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 62.50349650349651,
|
| 718 |
+
"grad_norm": 1.2600972652435303,
|
| 719 |
+
"learning_rate": 0.00011168571428571427,
|
| 720 |
+
"loss": 0.2738,
|
| 721 |
+
"step": 4500
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 62.50349650349651,
|
| 725 |
+
"eval_cer": 0.3593174326158619,
|
| 726 |
+
"eval_loss": 1.446284294128418,
|
| 727 |
+
"eval_runtime": 6.0798,
|
| 728 |
+
"eval_samples_per_second": 46.876,
|
| 729 |
+
"eval_steps_per_second": 5.921,
|
| 730 |
+
"step": 4500
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 63.89510489510489,
|
| 734 |
+
"grad_norm": 2.8400301933288574,
|
| 735 |
+
"learning_rate": 0.00010739999999999998,
|
| 736 |
+
"loss": 0.2574,
|
| 737 |
+
"step": 4600
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 63.89510489510489,
|
| 741 |
+
"eval_cer": 0.358347876672484,
|
| 742 |
+
"eval_loss": 1.4303468465805054,
|
| 743 |
+
"eval_runtime": 6.1117,
|
| 744 |
+
"eval_samples_per_second": 46.632,
|
| 745 |
+
"eval_steps_per_second": 5.89,
|
| 746 |
+
"step": 4600
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 65.27972027972028,
|
| 750 |
+
"grad_norm": 1.22808039188385,
|
| 751 |
+
"learning_rate": 0.0001031142857142857,
|
| 752 |
+
"loss": 0.2397,
|
| 753 |
+
"step": 4700
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 65.27972027972028,
|
| 757 |
+
"eval_cer": 0.3445801822765174,
|
| 758 |
+
"eval_loss": 1.4538311958312988,
|
| 759 |
+
"eval_runtime": 6.0831,
|
| 760 |
+
"eval_samples_per_second": 46.851,
|
| 761 |
+
"eval_steps_per_second": 5.918,
|
| 762 |
+
"step": 4700
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 66.67132867132867,
|
| 766 |
+
"grad_norm": 1.9458190202713013,
|
| 767 |
+
"learning_rate": 9.882857142857141e-05,
|
| 768 |
+
"loss": 0.2474,
|
| 769 |
+
"step": 4800
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 66.67132867132867,
|
| 773 |
+
"eval_cer": 0.3496218731820826,
|
| 774 |
+
"eval_loss": 1.4416499137878418,
|
| 775 |
+
"eval_runtime": 6.1066,
|
| 776 |
+
"eval_samples_per_second": 46.671,
|
| 777 |
+
"eval_steps_per_second": 5.895,
|
| 778 |
+
"step": 4800
|
| 779 |
+
},
|
| 780 |
+
{
|
| 781 |
+
"epoch": 68.05594405594405,
|
| 782 |
+
"grad_norm": 1.5397316217422485,
|
| 783 |
+
"learning_rate": 9.454285714285714e-05,
|
| 784 |
+
"loss": 0.2212,
|
| 785 |
+
"step": 4900
|
| 786 |
+
},
|
| 787 |
+
{
|
| 788 |
+
"epoch": 68.05594405594405,
|
| 789 |
+
"eval_cer": 0.34477409346519294,
|
| 790 |
+
"eval_loss": 1.476562738418579,
|
| 791 |
+
"eval_runtime": 6.125,
|
| 792 |
+
"eval_samples_per_second": 46.53,
|
| 793 |
+
"eval_steps_per_second": 5.878,
|
| 794 |
+
"step": 4900
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 69.44755244755245,
|
| 798 |
+
"grad_norm": 4.152817249298096,
|
| 799 |
+
"learning_rate": 9.025714285714285e-05,
|
| 800 |
+
"loss": 0.2173,
|
| 801 |
+
"step": 5000
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 69.44755244755245,
|
| 805 |
+
"eval_cer": 0.3496218731820826,
|
| 806 |
+
"eval_loss": 1.4784653186798096,
|
| 807 |
+
"eval_runtime": 6.0989,
|
| 808 |
+
"eval_samples_per_second": 46.73,
|
| 809 |
+
"eval_steps_per_second": 5.903,
|
| 810 |
+
"step": 5000
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 70.83916083916084,
|
| 814 |
+
"grad_norm": 1.8647962808609009,
|
| 815 |
+
"learning_rate": 8.597142857142857e-05,
|
| 816 |
+
"loss": 0.2138,
|
| 817 |
+
"step": 5100
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"epoch": 70.83916083916084,
|
| 821 |
+
"eval_cer": 0.3581539654838084,
|
| 822 |
+
"eval_loss": 1.4859139919281006,
|
| 823 |
+
"eval_runtime": 6.0911,
|
| 824 |
+
"eval_samples_per_second": 46.79,
|
| 825 |
+
"eval_steps_per_second": 5.91,
|
| 826 |
+
"step": 5100
|
| 827 |
+
},
|
| 828 |
+
{
|
| 829 |
+
"epoch": 72.22377622377623,
|
| 830 |
+
"grad_norm": 6.688198089599609,
|
| 831 |
+
"learning_rate": 8.168571428571428e-05,
|
| 832 |
+
"loss": 0.2037,
|
| 833 |
+
"step": 5200
|
| 834 |
+
},
|
| 835 |
+
{
|
| 836 |
+
"epoch": 72.22377622377623,
|
| 837 |
+
"eval_cer": 0.3500096955594338,
|
| 838 |
+
"eval_loss": 1.5022231340408325,
|
| 839 |
+
"eval_runtime": 6.1063,
|
| 840 |
+
"eval_samples_per_second": 46.673,
|
| 841 |
+
"eval_steps_per_second": 5.896,
|
| 842 |
+
"step": 5200
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 73.61538461538461,
|
| 846 |
+
"grad_norm": 2.6784660816192627,
|
| 847 |
+
"learning_rate": 7.74e-05,
|
| 848 |
+
"loss": 0.194,
|
| 849 |
+
"step": 5300
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 73.61538461538461,
|
| 853 |
+
"eval_cer": 0.34904013961605584,
|
| 854 |
+
"eval_loss": 1.4964337348937988,
|
| 855 |
+
"eval_runtime": 6.0868,
|
| 856 |
+
"eval_samples_per_second": 46.823,
|
| 857 |
+
"eval_steps_per_second": 5.914,
|
| 858 |
+
"step": 5300
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 75.0,
|
| 862 |
+
"grad_norm": 2.1717565059661865,
|
| 863 |
+
"learning_rate": 7.31142857142857e-05,
|
| 864 |
+
"loss": 0.1758,
|
| 865 |
+
"step": 5400
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 75.0,
|
| 869 |
+
"eval_cer": 0.35524529765367463,
|
| 870 |
+
"eval_loss": 1.5645275115966797,
|
| 871 |
+
"eval_runtime": 6.0591,
|
| 872 |
+
"eval_samples_per_second": 47.037,
|
| 873 |
+
"eval_steps_per_second": 5.941,
|
| 874 |
+
"step": 5400
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 76.39160839160839,
|
| 878 |
+
"grad_norm": 3.3418147563934326,
|
| 879 |
+
"learning_rate": 6.882857142857142e-05,
|
| 880 |
+
"loss": 0.1693,
|
| 881 |
+
"step": 5500
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 76.39160839160839,
|
| 885 |
+
"eval_cer": 0.3492340508047314,
|
| 886 |
+
"eval_loss": 1.5214943885803223,
|
| 887 |
+
"eval_runtime": 6.0965,
|
| 888 |
+
"eval_samples_per_second": 46.748,
|
| 889 |
+
"eval_steps_per_second": 5.905,
|
| 890 |
+
"step": 5500
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 77.78321678321679,
|
| 894 |
+
"grad_norm": 1.1721317768096924,
|
| 895 |
+
"learning_rate": 6.454285714285713e-05,
|
| 896 |
+
"loss": 0.1682,
|
| 897 |
+
"step": 5600
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 77.78321678321679,
|
| 901 |
+
"eval_cer": 0.34361062633313944,
|
| 902 |
+
"eval_loss": 1.557164192199707,
|
| 903 |
+
"eval_runtime": 6.0786,
|
| 904 |
+
"eval_samples_per_second": 46.886,
|
| 905 |
+
"eval_steps_per_second": 5.922,
|
| 906 |
+
"step": 5600
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 79.16783216783217,
|
| 910 |
+
"grad_norm": 1.2348577976226807,
|
| 911 |
+
"learning_rate": 6.0257142857142846e-05,
|
| 912 |
+
"loss": 0.1616,
|
| 913 |
+
"step": 5700
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 79.16783216783217,
|
| 917 |
+
"eval_cer": 0.3461314717859221,
|
| 918 |
+
"eval_loss": 1.497072696685791,
|
| 919 |
+
"eval_runtime": 6.0713,
|
| 920 |
+
"eval_samples_per_second": 46.942,
|
| 921 |
+
"eval_steps_per_second": 5.93,
|
| 922 |
+
"step": 5700
|
| 923 |
+
},
|
| 924 |
+
{
|
| 925 |
+
"epoch": 80.55944055944056,
|
| 926 |
+
"grad_norm": 1.101247787475586,
|
| 927 |
+
"learning_rate": 5.597142857142857e-05,
|
| 928 |
+
"loss": 0.1625,
|
| 929 |
+
"step": 5800
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 80.55944055944056,
|
| 933 |
+
"eval_cer": 0.3515609850688385,
|
| 934 |
+
"eval_loss": 1.5326788425445557,
|
| 935 |
+
"eval_runtime": 6.0886,
|
| 936 |
+
"eval_samples_per_second": 46.809,
|
| 937 |
+
"eval_steps_per_second": 5.913,
|
| 938 |
+
"step": 5800
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 81.95104895104895,
|
| 942 |
+
"grad_norm": 4.953105449676514,
|
| 943 |
+
"learning_rate": 5.168571428571428e-05,
|
| 944 |
+
"loss": 0.1432,
|
| 945 |
+
"step": 5900
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 81.95104895104895,
|
| 949 |
+
"eval_cer": 0.35059142912546054,
|
| 950 |
+
"eval_loss": 1.5595433712005615,
|
| 951 |
+
"eval_runtime": 6.078,
|
| 952 |
+
"eval_samples_per_second": 46.89,
|
| 953 |
+
"eval_steps_per_second": 5.923,
|
| 954 |
+
"step": 5900
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 83.33566433566433,
|
| 958 |
+
"grad_norm": 1.2514948844909668,
|
| 959 |
+
"learning_rate": 4.7399999999999993e-05,
|
| 960 |
+
"loss": 0.1348,
|
| 961 |
+
"step": 6000
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 83.33566433566433,
|
| 965 |
+
"eval_cer": 0.3482644948613535,
|
| 966 |
+
"eval_loss": 1.556195616722107,
|
| 967 |
+
"eval_runtime": 6.0756,
|
| 968 |
+
"eval_samples_per_second": 46.909,
|
| 969 |
+
"eval_steps_per_second": 5.925,
|
| 970 |
+
"step": 6000
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 84.72727272727273,
|
| 974 |
+
"grad_norm": 0.5957698225975037,
|
| 975 |
+
"learning_rate": 4.311428571428571e-05,
|
| 976 |
+
"loss": 0.137,
|
| 977 |
+
"step": 6100
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 84.72727272727273,
|
| 981 |
+
"eval_cer": 0.3484584060500291,
|
| 982 |
+
"eval_loss": 1.590181827545166,
|
| 983 |
+
"eval_runtime": 6.0761,
|
| 984 |
+
"eval_samples_per_second": 46.905,
|
| 985 |
+
"eval_steps_per_second": 5.925,
|
| 986 |
+
"step": 6100
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 86.1118881118881,
|
| 990 |
+
"grad_norm": 2.8553950786590576,
|
| 991 |
+
"learning_rate": 3.882857142857142e-05,
|
| 992 |
+
"loss": 0.1263,
|
| 993 |
+
"step": 6200
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 86.1118881118881,
|
| 997 |
+
"eval_cer": 0.35214271863486524,
|
| 998 |
+
"eval_loss": 1.5852645635604858,
|
| 999 |
+
"eval_runtime": 6.1244,
|
| 1000 |
+
"eval_samples_per_second": 46.535,
|
| 1001 |
+
"eval_steps_per_second": 5.878,
|
| 1002 |
+
"step": 6200
|
| 1003 |
+
},
|
| 1004 |
+
{
|
| 1005 |
+
"epoch": 87.5034965034965,
|
| 1006 |
+
"grad_norm": 0.7932090759277344,
|
| 1007 |
+
"learning_rate": 3.454285714285714e-05,
|
| 1008 |
+
"loss": 0.1271,
|
| 1009 |
+
"step": 6300
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"epoch": 87.5034965034965,
|
| 1013 |
+
"eval_cer": 0.3488462284273803,
|
| 1014 |
+
"eval_loss": 1.597654938697815,
|
| 1015 |
+
"eval_runtime": 6.1436,
|
| 1016 |
+
"eval_samples_per_second": 46.389,
|
| 1017 |
+
"eval_steps_per_second": 5.86,
|
| 1018 |
+
"step": 6300
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 88.8951048951049,
|
| 1022 |
+
"grad_norm": 2.1244022846221924,
|
| 1023 |
+
"learning_rate": 3.0257142857142855e-05,
|
| 1024 |
+
"loss": 0.123,
|
| 1025 |
+
"step": 6400
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 88.8951048951049,
|
| 1029 |
+
"eval_cer": 0.3498157843707582,
|
| 1030 |
+
"eval_loss": 1.6023805141448975,
|
| 1031 |
+
"eval_runtime": 6.0927,
|
| 1032 |
+
"eval_samples_per_second": 46.777,
|
| 1033 |
+
"eval_steps_per_second": 5.909,
|
| 1034 |
+
"step": 6400
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 90.27972027972028,
|
| 1038 |
+
"grad_norm": 1.3532764911651611,
|
| 1039 |
+
"learning_rate": 2.5971428571428572e-05,
|
| 1040 |
+
"loss": 0.117,
|
| 1041 |
+
"step": 6500
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 90.27972027972028,
|
| 1045 |
+
"eval_cer": 0.3535000969555943,
|
| 1046 |
+
"eval_loss": 1.6092747449874878,
|
| 1047 |
+
"eval_runtime": 6.1301,
|
| 1048 |
+
"eval_samples_per_second": 46.492,
|
| 1049 |
+
"eval_steps_per_second": 5.873,
|
| 1050 |
+
"step": 6500
|
| 1051 |
+
},
|
| 1052 |
+
{
|
| 1053 |
+
"epoch": 91.67132867132867,
|
| 1054 |
+
"grad_norm": 3.1814770698547363,
|
| 1055 |
+
"learning_rate": 2.1685714285714286e-05,
|
| 1056 |
+
"loss": 0.1077,
|
| 1057 |
+
"step": 6600
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"epoch": 91.67132867132867,
|
| 1061 |
+
"eval_cer": 0.35194880744618967,
|
| 1062 |
+
"eval_loss": 1.5807358026504517,
|
| 1063 |
+
"eval_runtime": 6.0891,
|
| 1064 |
+
"eval_samples_per_second": 46.805,
|
| 1065 |
+
"eval_steps_per_second": 5.912,
|
| 1066 |
+
"step": 6600
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 93.05594405594405,
|
| 1070 |
+
"grad_norm": 1.0375070571899414,
|
| 1071 |
+
"learning_rate": 1.74e-05,
|
| 1072 |
+
"loss": 0.1072,
|
| 1073 |
+
"step": 6700
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 93.05594405594405,
|
| 1077 |
+
"eval_cer": 0.3476827612953267,
|
| 1078 |
+
"eval_loss": 1.580068588256836,
|
| 1079 |
+
"eval_runtime": 6.1035,
|
| 1080 |
+
"eval_samples_per_second": 46.695,
|
| 1081 |
+
"eval_steps_per_second": 5.898,
|
| 1082 |
+
"step": 6700
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 94.44755244755245,
|
| 1086 |
+
"grad_norm": 1.3363580703735352,
|
| 1087 |
+
"learning_rate": 1.3114285714285713e-05,
|
| 1088 |
+
"loss": 0.1063,
|
| 1089 |
+
"step": 6800
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 94.44755244755245,
|
| 1093 |
+
"eval_cer": 0.35020360674810935,
|
| 1094 |
+
"eval_loss": 1.5893759727478027,
|
| 1095 |
+
"eval_runtime": 6.0883,
|
| 1096 |
+
"eval_samples_per_second": 46.811,
|
| 1097 |
+
"eval_steps_per_second": 5.913,
|
| 1098 |
+
"step": 6800
|
| 1099 |
+
},
|
| 1100 |
+
{
|
| 1101 |
+
"epoch": 95.83916083916084,
|
| 1102 |
+
"grad_norm": 2.259174108505249,
|
| 1103 |
+
"learning_rate": 8.828571428571429e-06,
|
| 1104 |
+
"loss": 0.103,
|
| 1105 |
+
"step": 6900
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 95.83916083916084,
|
| 1109 |
+
"eval_cer": 0.3498157843707582,
|
| 1110 |
+
"eval_loss": 1.602723240852356,
|
| 1111 |
+
"eval_runtime": 6.0922,
|
| 1112 |
+
"eval_samples_per_second": 46.781,
|
| 1113 |
+
"eval_steps_per_second": 5.909,
|
| 1114 |
+
"step": 6900
|
| 1115 |
+
},
|
| 1116 |
+
{
|
| 1117 |
+
"epoch": 97.22377622377623,
|
| 1118 |
+
"grad_norm": 1.8573235273361206,
|
| 1119 |
+
"learning_rate": 4.585714285714285e-06,
|
| 1120 |
+
"loss": 0.1032,
|
| 1121 |
+
"step": 7000
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 97.22377622377623,
|
| 1125 |
+
"eval_cer": 0.3484584060500291,
|
| 1126 |
+
"eval_loss": 1.6034408807754517,
|
| 1127 |
+
"eval_runtime": 6.1003,
|
| 1128 |
+
"eval_samples_per_second": 46.719,
|
| 1129 |
+
"eval_steps_per_second": 5.901,
|
| 1130 |
+
"step": 7000
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 98.61538461538461,
|
| 1134 |
+
"grad_norm": 1.2302757501602173,
|
| 1135 |
+
"learning_rate": 3e-07,
|
| 1136 |
+
"loss": 0.0971,
|
| 1137 |
+
"step": 7100
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 98.61538461538461,
|
| 1141 |
+
"eval_cer": 0.3480705836726779,
|
| 1142 |
+
"eval_loss": 1.6103968620300293,
|
| 1143 |
+
"eval_runtime": 6.1158,
|
| 1144 |
+
"eval_samples_per_second": 46.6,
|
| 1145 |
+
"eval_steps_per_second": 5.886,
|
| 1146 |
+
"step": 7100
|
| 1147 |
+
},
|
| 1148 |
+
{
|
| 1149 |
+
"epoch": 98.61538461538461,
|
| 1150 |
+
"step": 7100,
|
| 1151 |
+
"total_flos": 1.5637471182402683e+19,
|
| 1152 |
+
"train_loss": 1.063481583259475,
|
| 1153 |
+
"train_runtime": 9766.9628,
|
| 1154 |
+
"train_samples_per_second": 23.354,
|
| 1155 |
+
"train_steps_per_second": 0.727
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 98.61538461538461,
|
| 1159 |
+
"eval_cer": 0.35214271863486524,
|
| 1160 |
+
"eval_loss": 1.6103789806365967,
|
| 1161 |
+
"eval_runtime": 6.0538,
|
| 1162 |
+
"eval_samples_per_second": 47.078,
|
| 1163 |
+
"eval_steps_per_second": 5.947,
|
| 1164 |
+
"step": 7100
|
| 1165 |
}
|
| 1166 |
],
|
| 1167 |
"logging_steps": 100,
|
| 1168 |
+
"max_steps": 7100,
|
| 1169 |
"num_input_tokens_seen": 0,
|
| 1170 |
+
"num_train_epochs": 100,
|
| 1171 |
"save_steps": 100,
|
| 1172 |
"stateful_callbacks": {
|
| 1173 |
"TrainerControl": {
|
|
|
|
| 1181 |
"attributes": {}
|
| 1182 |
}
|
| 1183 |
},
|
| 1184 |
+
"total_flos": 1.5637471182402683e+19,
|
| 1185 |
"train_batch_size": 16,
|
| 1186 |
"trial_name": null,
|
| 1187 |
"trial_params": null
|