End of training
Browse files- README.md +245 -0
- all_results.json +13 -0
- config.json +52 -0
- eval_results.json +8 -0
- preprocessor_config.json +22 -0
- pytorch_model.bin +3 -0
- train_results.json +7 -0
- trainer_state.json +2593 -0
- training_args.bin +3 -0
README.md
ADDED
|
@@ -0,0 +1,245 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
base_model: google/vit-base-patch16-224-in21k
|
| 4 |
+
tags:
|
| 5 |
+
- generated_from_trainer
|
| 6 |
+
datasets:
|
| 7 |
+
- imagefolder
|
| 8 |
+
metrics:
|
| 9 |
+
- accuracy
|
| 10 |
+
model-index:
|
| 11 |
+
- name: vit-base-GTZAN
|
| 12 |
+
results:
|
| 13 |
+
- task:
|
| 14 |
+
name: Image Classification
|
| 15 |
+
type: image-classification
|
| 16 |
+
dataset:
|
| 17 |
+
name: imagefolder
|
| 18 |
+
type: imagefolder
|
| 19 |
+
config: default
|
| 20 |
+
split: train
|
| 21 |
+
args: default
|
| 22 |
+
metrics:
|
| 23 |
+
- name: Accuracy
|
| 24 |
+
type: accuracy
|
| 25 |
+
value: 0.7566137566137566
|
| 26 |
+
---
|
| 27 |
+
|
| 28 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 29 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 30 |
+
|
| 31 |
+
# vit-base-GTZAN
|
| 32 |
+
|
| 33 |
+
This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the imagefolder dataset.
|
| 34 |
+
It achieves the following results on the evaluation set:
|
| 35 |
+
- Loss: 0.8328
|
| 36 |
+
- Accuracy: 0.7566
|
| 37 |
+
|
| 38 |
+
## Model description
|
| 39 |
+
|
| 40 |
+
More information needed
|
| 41 |
+
|
| 42 |
+
## Intended uses & limitations
|
| 43 |
+
|
| 44 |
+
More information needed
|
| 45 |
+
|
| 46 |
+
## Training and evaluation data
|
| 47 |
+
|
| 48 |
+
More information needed
|
| 49 |
+
|
| 50 |
+
## Training procedure
|
| 51 |
+
|
| 52 |
+
### Training hyperparameters
|
| 53 |
+
|
| 54 |
+
The following hyperparameters were used during training:
|
| 55 |
+
- learning_rate: 0.0002
|
| 56 |
+
- train_batch_size: 16
|
| 57 |
+
- eval_batch_size: 8
|
| 58 |
+
- seed: 42
|
| 59 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 60 |
+
- lr_scheduler_type: linear
|
| 61 |
+
- num_epochs: 16
|
| 62 |
+
|
| 63 |
+
### Training results
|
| 64 |
+
|
| 65 |
+
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|
| 66 |
+
|:-------------:|:-----:|:----:|:---------------:|:--------:|
|
| 67 |
+
| 2.3756 | 0.09 | 10 | 2.2861 | 0.2116 |
|
| 68 |
+
| 2.3051 | 0.19 | 20 | 2.1907 | 0.3439 |
|
| 69 |
+
| 2.1219 | 0.28 | 30 | 2.0214 | 0.3175 |
|
| 70 |
+
| 2.0542 | 0.37 | 40 | 1.9059 | 0.4074 |
|
| 71 |
+
| 1.8132 | 0.47 | 50 | 1.8472 | 0.3862 |
|
| 72 |
+
| 1.8854 | 0.56 | 60 | 1.6832 | 0.4603 |
|
| 73 |
+
| 1.6981 | 0.65 | 70 | 1.6008 | 0.4974 |
|
| 74 |
+
| 1.5251 | 0.75 | 80 | 1.4685 | 0.5026 |
|
| 75 |
+
| 1.4463 | 0.84 | 90 | 1.3713 | 0.6138 |
|
| 76 |
+
| 1.4335 | 0.93 | 100 | 1.4270 | 0.4974 |
|
| 77 |
+
| 1.1147 | 1.03 | 110 | 1.2793 | 0.5926 |
|
| 78 |
+
| 1.3568 | 1.12 | 120 | 1.3360 | 0.5661 |
|
| 79 |
+
| 1.3077 | 1.21 | 130 | 1.4520 | 0.5079 |
|
| 80 |
+
| 1.2801 | 1.31 | 140 | 1.2765 | 0.5661 |
|
| 81 |
+
| 1.2894 | 1.4 | 150 | 1.1949 | 0.6138 |
|
| 82 |
+
| 1.2657 | 1.5 | 160 | 1.1937 | 0.6349 |
|
| 83 |
+
| 0.8784 | 1.59 | 170 | 1.2190 | 0.6032 |
|
| 84 |
+
| 1.1575 | 1.68 | 180 | 1.2268 | 0.6138 |
|
| 85 |
+
| 0.9848 | 1.78 | 190 | 1.0572 | 0.6561 |
|
| 86 |
+
| 0.9409 | 1.87 | 200 | 1.1609 | 0.6349 |
|
| 87 |
+
| 0.9448 | 1.96 | 210 | 1.2327 | 0.6085 |
|
| 88 |
+
| 1.0819 | 2.06 | 220 | 1.1699 | 0.5820 |
|
| 89 |
+
| 0.7485 | 2.15 | 230 | 1.1041 | 0.6508 |
|
| 90 |
+
| 0.8934 | 2.24 | 240 | 1.1672 | 0.5873 |
|
| 91 |
+
| 0.8609 | 2.34 | 250 | 1.1900 | 0.6190 |
|
| 92 |
+
| 0.7935 | 2.43 | 260 | 1.0623 | 0.6402 |
|
| 93 |
+
| 0.8013 | 2.52 | 270 | 0.9873 | 0.6878 |
|
| 94 |
+
| 0.6669 | 2.62 | 280 | 1.0078 | 0.6561 |
|
| 95 |
+
| 0.7847 | 2.71 | 290 | 1.1484 | 0.6085 |
|
| 96 |
+
| 0.7222 | 2.8 | 300 | 1.1295 | 0.6243 |
|
| 97 |
+
| 0.7844 | 2.9 | 310 | 0.9414 | 0.7249 |
|
| 98 |
+
| 0.8057 | 2.99 | 320 | 1.0504 | 0.6667 |
|
| 99 |
+
| 0.4843 | 3.08 | 330 | 0.9874 | 0.6508 |
|
| 100 |
+
| 0.6766 | 3.18 | 340 | 1.1496 | 0.6508 |
|
| 101 |
+
| 0.4818 | 3.27 | 350 | 1.0968 | 0.6878 |
|
| 102 |
+
| 0.5351 | 3.36 | 360 | 1.1394 | 0.6296 |
|
| 103 |
+
| 0.5035 | 3.46 | 370 | 0.9815 | 0.7090 |
|
| 104 |
+
| 0.4032 | 3.55 | 380 | 1.0882 | 0.6402 |
|
| 105 |
+
| 0.639 | 3.64 | 390 | 1.2611 | 0.6085 |
|
| 106 |
+
| 0.5156 | 3.74 | 400 | 1.0376 | 0.6561 |
|
| 107 |
+
| 0.4884 | 3.83 | 410 | 0.9506 | 0.6984 |
|
| 108 |
+
| 0.5875 | 3.93 | 420 | 0.8479 | 0.7513 |
|
| 109 |
+
| 0.6982 | 4.02 | 430 | 1.0895 | 0.6825 |
|
| 110 |
+
| 0.3966 | 4.11 | 440 | 0.9709 | 0.6984 |
|
| 111 |
+
| 0.377 | 4.21 | 450 | 0.9754 | 0.6772 |
|
| 112 |
+
| 0.3417 | 4.3 | 460 | 1.1687 | 0.6508 |
|
| 113 |
+
| 0.336 | 4.39 | 470 | 0.9826 | 0.6984 |
|
| 114 |
+
| 0.5201 | 4.49 | 480 | 1.1770 | 0.6614 |
|
| 115 |
+
| 0.1737 | 4.58 | 490 | 1.0491 | 0.6878 |
|
| 116 |
+
| 0.2545 | 4.67 | 500 | 1.1352 | 0.6984 |
|
| 117 |
+
| 0.3752 | 4.77 | 510 | 1.0300 | 0.6931 |
|
| 118 |
+
| 0.3667 | 4.86 | 520 | 1.0355 | 0.6825 |
|
| 119 |
+
| 0.2797 | 4.95 | 530 | 0.9882 | 0.6984 |
|
| 120 |
+
| 0.1646 | 5.05 | 540 | 1.0728 | 0.6984 |
|
| 121 |
+
| 0.2199 | 5.14 | 550 | 0.8328 | 0.7566 |
|
| 122 |
+
| 0.2191 | 5.23 | 560 | 0.9280 | 0.7460 |
|
| 123 |
+
| 0.12 | 5.33 | 570 | 1.0978 | 0.7037 |
|
| 124 |
+
| 0.2608 | 5.42 | 580 | 1.1158 | 0.6878 |
|
| 125 |
+
| 0.2 | 5.51 | 590 | 1.0873 | 0.7354 |
|
| 126 |
+
| 0.1899 | 5.61 | 600 | 1.0560 | 0.7143 |
|
| 127 |
+
| 0.1113 | 5.7 | 610 | 1.1144 | 0.7037 |
|
| 128 |
+
| 0.2279 | 5.79 | 620 | 1.2535 | 0.6667 |
|
| 129 |
+
| 0.1563 | 5.89 | 630 | 1.0803 | 0.7354 |
|
| 130 |
+
| 0.2182 | 5.98 | 640 | 1.3904 | 0.6349 |
|
| 131 |
+
| 0.1781 | 6.07 | 650 | 1.3461 | 0.6720 |
|
| 132 |
+
| 0.1395 | 6.17 | 660 | 1.2769 | 0.6825 |
|
| 133 |
+
| 0.2308 | 6.26 | 670 | 1.2213 | 0.6931 |
|
| 134 |
+
| 0.1899 | 6.36 | 680 | 1.0948 | 0.7143 |
|
| 135 |
+
| 0.1702 | 6.45 | 690 | 1.2383 | 0.6931 |
|
| 136 |
+
| 0.1055 | 6.54 | 700 | 1.4010 | 0.6349 |
|
| 137 |
+
| 0.1151 | 6.64 | 710 | 1.2607 | 0.6720 |
|
| 138 |
+
| 0.2415 | 6.73 | 720 | 1.0520 | 0.7302 |
|
| 139 |
+
| 0.117 | 6.82 | 730 | 1.0548 | 0.7354 |
|
| 140 |
+
| 0.184 | 6.92 | 740 | 1.1872 | 0.6984 |
|
| 141 |
+
| 0.1997 | 7.01 | 750 | 1.1128 | 0.7249 |
|
| 142 |
+
| 0.0645 | 7.1 | 760 | 1.1514 | 0.6984 |
|
| 143 |
+
| 0.1025 | 7.2 | 770 | 1.2252 | 0.7037 |
|
| 144 |
+
| 0.0407 | 7.29 | 780 | 1.0571 | 0.7513 |
|
| 145 |
+
| 0.1752 | 7.38 | 790 | 1.0812 | 0.7354 |
|
| 146 |
+
| 0.1143 | 7.48 | 800 | 1.2182 | 0.7143 |
|
| 147 |
+
| 0.1542 | 7.57 | 810 | 1.1789 | 0.7143 |
|
| 148 |
+
| 0.0859 | 7.66 | 820 | 1.1392 | 0.7196 |
|
| 149 |
+
| 0.119 | 7.76 | 830 | 1.1568 | 0.7354 |
|
| 150 |
+
| 0.0913 | 7.85 | 840 | 1.1097 | 0.6984 |
|
| 151 |
+
| 0.085 | 7.94 | 850 | 1.1189 | 0.7460 |
|
| 152 |
+
| 0.0201 | 8.04 | 860 | 1.1283 | 0.7143 |
|
| 153 |
+
| 0.0509 | 8.13 | 870 | 1.1005 | 0.7407 |
|
| 154 |
+
| 0.0326 | 8.22 | 880 | 1.0490 | 0.7302 |
|
| 155 |
+
| 0.0728 | 8.32 | 890 | 1.2511 | 0.7196 |
|
| 156 |
+
| 0.0486 | 8.41 | 900 | 1.1833 | 0.7143 |
|
| 157 |
+
| 0.0645 | 8.5 | 910 | 0.9881 | 0.7725 |
|
| 158 |
+
| 0.0194 | 8.6 | 920 | 1.0412 | 0.7566 |
|
| 159 |
+
| 0.0215 | 8.69 | 930 | 1.2485 | 0.7196 |
|
| 160 |
+
| 0.0853 | 8.79 | 940 | 1.0864 | 0.7672 |
|
| 161 |
+
| 0.0412 | 8.88 | 950 | 1.1796 | 0.7249 |
|
| 162 |
+
| 0.0645 | 8.97 | 960 | 1.3152 | 0.6878 |
|
| 163 |
+
| 0.0654 | 9.07 | 970 | 1.2789 | 0.6931 |
|
| 164 |
+
| 0.0352 | 9.16 | 980 | 1.1928 | 0.7196 |
|
| 165 |
+
| 0.0137 | 9.25 | 990 | 1.1643 | 0.7354 |
|
| 166 |
+
| 0.0227 | 9.35 | 1000 | 1.2256 | 0.7143 |
|
| 167 |
+
| 0.0391 | 9.44 | 1010 | 1.2089 | 0.7196 |
|
| 168 |
+
| 0.0163 | 9.53 | 1020 | 1.3880 | 0.6931 |
|
| 169 |
+
| 0.0225 | 9.63 | 1030 | 1.3944 | 0.6931 |
|
| 170 |
+
| 0.0348 | 9.72 | 1040 | 1.3257 | 0.7143 |
|
| 171 |
+
| 0.0354 | 9.81 | 1050 | 1.1538 | 0.7460 |
|
| 172 |
+
| 0.0412 | 9.91 | 1060 | 1.2372 | 0.7249 |
|
| 173 |
+
| 0.055 | 10.0 | 1070 | 1.2266 | 0.7090 |
|
| 174 |
+
| 0.0115 | 10.09 | 1080 | 1.2353 | 0.7249 |
|
| 175 |
+
| 0.011 | 10.19 | 1090 | 1.2655 | 0.7249 |
|
| 176 |
+
| 0.0105 | 10.28 | 1100 | 1.2831 | 0.7354 |
|
| 177 |
+
| 0.0248 | 10.37 | 1110 | 1.3138 | 0.7143 |
|
| 178 |
+
| 0.0287 | 10.47 | 1120 | 1.2472 | 0.7196 |
|
| 179 |
+
| 0.017 | 10.56 | 1130 | 1.1517 | 0.7619 |
|
| 180 |
+
| 0.0326 | 10.65 | 1140 | 1.1729 | 0.7513 |
|
| 181 |
+
| 0.0298 | 10.75 | 1150 | 1.1991 | 0.7460 |
|
| 182 |
+
| 0.0087 | 10.84 | 1160 | 1.1965 | 0.7196 |
|
| 183 |
+
| 0.0104 | 10.93 | 1170 | 1.2006 | 0.7302 |
|
| 184 |
+
| 0.0176 | 11.03 | 1180 | 1.2819 | 0.7196 |
|
| 185 |
+
| 0.0088 | 11.12 | 1190 | 1.2860 | 0.7249 |
|
| 186 |
+
| 0.0218 | 11.21 | 1200 | 1.1996 | 0.7407 |
|
| 187 |
+
| 0.011 | 11.31 | 1210 | 1.1905 | 0.7407 |
|
| 188 |
+
| 0.0195 | 11.4 | 1220 | 1.1777 | 0.7460 |
|
| 189 |
+
| 0.012 | 11.5 | 1230 | 1.1417 | 0.7566 |
|
| 190 |
+
| 0.0075 | 11.59 | 1240 | 1.1429 | 0.7619 |
|
| 191 |
+
| 0.0131 | 11.68 | 1250 | 1.1381 | 0.7672 |
|
| 192 |
+
| 0.0078 | 11.78 | 1260 | 1.1562 | 0.7566 |
|
| 193 |
+
| 0.0071 | 11.87 | 1270 | 1.1708 | 0.7619 |
|
| 194 |
+
| 0.04 | 11.96 | 1280 | 1.1965 | 0.7513 |
|
| 195 |
+
| 0.0066 | 12.06 | 1290 | 1.2295 | 0.7354 |
|
| 196 |
+
| 0.0179 | 12.15 | 1300 | 1.2337 | 0.7354 |
|
| 197 |
+
| 0.0072 | 12.24 | 1310 | 1.2376 | 0.7407 |
|
| 198 |
+
| 0.0189 | 12.34 | 1320 | 1.2402 | 0.7354 |
|
| 199 |
+
| 0.0067 | 12.43 | 1330 | 1.2426 | 0.7407 |
|
| 200 |
+
| 0.014 | 12.52 | 1340 | 1.2199 | 0.7460 |
|
| 201 |
+
| 0.0065 | 12.62 | 1350 | 1.2070 | 0.7513 |
|
| 202 |
+
| 0.0119 | 12.71 | 1360 | 1.2172 | 0.7513 |
|
| 203 |
+
| 0.0065 | 12.8 | 1370 | 1.2299 | 0.7460 |
|
| 204 |
+
| 0.0139 | 12.9 | 1380 | 1.2095 | 0.7513 |
|
| 205 |
+
| 0.0195 | 12.99 | 1390 | 1.1914 | 0.7513 |
|
| 206 |
+
| 0.0102 | 13.08 | 1400 | 1.1972 | 0.7513 |
|
| 207 |
+
| 0.0162 | 13.18 | 1410 | 1.2006 | 0.7566 |
|
| 208 |
+
| 0.0057 | 13.27 | 1420 | 1.2135 | 0.7566 |
|
| 209 |
+
| 0.0099 | 13.36 | 1430 | 1.2060 | 0.7566 |
|
| 210 |
+
| 0.0092 | 13.46 | 1440 | 1.2094 | 0.7513 |
|
| 211 |
+
| 0.0059 | 13.55 | 1450 | 1.2153 | 0.7460 |
|
| 212 |
+
| 0.0132 | 13.64 | 1460 | 1.2271 | 0.7513 |
|
| 213 |
+
| 0.0224 | 13.74 | 1470 | 1.2394 | 0.7460 |
|
| 214 |
+
| 0.0116 | 13.83 | 1480 | 1.2354 | 0.7460 |
|
| 215 |
+
| 0.0096 | 13.93 | 1490 | 1.2316 | 0.7460 |
|
| 216 |
+
| 0.0055 | 14.02 | 1500 | 1.2332 | 0.7460 |
|
| 217 |
+
| 0.009 | 14.11 | 1510 | 1.2355 | 0.7460 |
|
| 218 |
+
| 0.0058 | 14.21 | 1520 | 1.2447 | 0.7460 |
|
| 219 |
+
| 0.01 | 14.3 | 1530 | 1.2437 | 0.7460 |
|
| 220 |
+
| 0.0055 | 14.39 | 1540 | 1.2422 | 0.7460 |
|
| 221 |
+
| 0.0187 | 14.49 | 1550 | 1.2215 | 0.7513 |
|
| 222 |
+
| 0.0103 | 14.58 | 1560 | 1.2178 | 0.7513 |
|
| 223 |
+
| 0.0053 | 14.67 | 1570 | 1.2217 | 0.7460 |
|
| 224 |
+
| 0.01 | 14.77 | 1580 | 1.2267 | 0.7460 |
|
| 225 |
+
| 0.0238 | 14.86 | 1590 | 1.2279 | 0.7460 |
|
| 226 |
+
| 0.0091 | 14.95 | 1600 | 1.2242 | 0.7460 |
|
| 227 |
+
| 0.0053 | 15.05 | 1610 | 1.2232 | 0.7513 |
|
| 228 |
+
| 0.0101 | 15.14 | 1620 | 1.2257 | 0.7460 |
|
| 229 |
+
| 0.0189 | 15.23 | 1630 | 1.2277 | 0.7460 |
|
| 230 |
+
| 0.0056 | 15.33 | 1640 | 1.2336 | 0.7460 |
|
| 231 |
+
| 0.0052 | 15.42 | 1650 | 1.2353 | 0.7460 |
|
| 232 |
+
| 0.0054 | 15.51 | 1660 | 1.2359 | 0.7460 |
|
| 233 |
+
| 0.0054 | 15.61 | 1670 | 1.2362 | 0.7460 |
|
| 234 |
+
| 0.0102 | 15.7 | 1680 | 1.2348 | 0.7513 |
|
| 235 |
+
| 0.0193 | 15.79 | 1690 | 1.2326 | 0.7513 |
|
| 236 |
+
| 0.0104 | 15.89 | 1700 | 1.2315 | 0.7513 |
|
| 237 |
+
| 0.0095 | 15.98 | 1710 | 1.2312 | 0.7513 |
|
| 238 |
+
|
| 239 |
+
|
| 240 |
+
### Framework versions
|
| 241 |
+
|
| 242 |
+
- Transformers 4.33.2
|
| 243 |
+
- Pytorch 2.2.0.dev20230912+cu121
|
| 244 |
+
- Datasets 2.14.5
|
| 245 |
+
- Tokenizers 0.13.3
|
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 16.0,
|
| 3 |
+
"eval_accuracy": 0.7566137566137566,
|
| 4 |
+
"eval_loss": 0.8328016996383667,
|
| 5 |
+
"eval_runtime": 3.3904,
|
| 6 |
+
"eval_samples_per_second": 55.746,
|
| 7 |
+
"eval_steps_per_second": 7.079,
|
| 8 |
+
"total_flos": 6.192053917739827e+17,
|
| 9 |
+
"train_loss": 0.32865437336057146,
|
| 10 |
+
"train_runtime": 1354.021,
|
| 11 |
+
"train_samples_per_second": 20.053,
|
| 12 |
+
"train_steps_per_second": 1.264
|
| 13 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,52 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "google/vit-base-patch16-224-in21k",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"ViTForImageClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.0,
|
| 7 |
+
"encoder_stride": 16,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.0,
|
| 10 |
+
"hidden_size": 768,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "afro",
|
| 13 |
+
"1": "classical",
|
| 14 |
+
"10": "reggae",
|
| 15 |
+
"11": "rock",
|
| 16 |
+
"2": "country",
|
| 17 |
+
"3": "disco",
|
| 18 |
+
"4": "electro",
|
| 19 |
+
"5": "jazz",
|
| 20 |
+
"6": "latin",
|
| 21 |
+
"7": "metal",
|
| 22 |
+
"8": "pop",
|
| 23 |
+
"9": "rap"
|
| 24 |
+
},
|
| 25 |
+
"image_size": 224,
|
| 26 |
+
"initializer_range": 0.02,
|
| 27 |
+
"intermediate_size": 3072,
|
| 28 |
+
"label2id": {
|
| 29 |
+
"afro": "0",
|
| 30 |
+
"classical": "1",
|
| 31 |
+
"country": "2",
|
| 32 |
+
"disco": "3",
|
| 33 |
+
"electro": "4",
|
| 34 |
+
"jazz": "5",
|
| 35 |
+
"latin": "6",
|
| 36 |
+
"metal": "7",
|
| 37 |
+
"pop": "8",
|
| 38 |
+
"rap": "9",
|
| 39 |
+
"reggae": "10",
|
| 40 |
+
"rock": "11"
|
| 41 |
+
},
|
| 42 |
+
"layer_norm_eps": 1e-12,
|
| 43 |
+
"model_type": "vit",
|
| 44 |
+
"num_attention_heads": 12,
|
| 45 |
+
"num_channels": 3,
|
| 46 |
+
"num_hidden_layers": 12,
|
| 47 |
+
"patch_size": 16,
|
| 48 |
+
"problem_type": "single_label_classification",
|
| 49 |
+
"qkv_bias": true,
|
| 50 |
+
"torch_dtype": "float32",
|
| 51 |
+
"transformers_version": "4.33.2"
|
| 52 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 16.0,
|
| 3 |
+
"eval_accuracy": 0.7566137566137566,
|
| 4 |
+
"eval_loss": 0.8328016996383667,
|
| 5 |
+
"eval_runtime": 3.3904,
|
| 6 |
+
"eval_samples_per_second": 55.746,
|
| 7 |
+
"eval_steps_per_second": 7.079
|
| 8 |
+
}
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_normalize": true,
|
| 3 |
+
"do_rescale": true,
|
| 4 |
+
"do_resize": true,
|
| 5 |
+
"image_mean": [
|
| 6 |
+
0.5,
|
| 7 |
+
0.5,
|
| 8 |
+
0.5
|
| 9 |
+
],
|
| 10 |
+
"image_processor_type": "ViTImageProcessor",
|
| 11 |
+
"image_std": [
|
| 12 |
+
0.5,
|
| 13 |
+
0.5,
|
| 14 |
+
0.5
|
| 15 |
+
],
|
| 16 |
+
"resample": 2,
|
| 17 |
+
"rescale_factor": 0.00392156862745098,
|
| 18 |
+
"size": {
|
| 19 |
+
"height": 224,
|
| 20 |
+
"width": 224
|
| 21 |
+
}
|
| 22 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:40990a012b0e96c2eb53e0f5127a5e7b78c0475f0a73d08a084d0bce96591d87
|
| 3 |
+
size 343299882
|
train_results.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 16.0,
|
| 3 |
+
"train_loss": 0.32865437336057146,
|
| 4 |
+
"train_runtime": 1354.021,
|
| 5 |
+
"train_samples_per_second": 20.053,
|
| 6 |
+
"train_steps_per_second": 1.264
|
| 7 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2593 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.8328016996383667,
|
| 3 |
+
"best_model_checkpoint": "./vit-base-GTZAN\\checkpoint-550",
|
| 4 |
+
"epoch": 16.0,
|
| 5 |
+
"eval_steps": 10,
|
| 6 |
+
"global_step": 1712,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.09,
|
| 13 |
+
"learning_rate": 0.0001988317757009346,
|
| 14 |
+
"loss": 2.3756,
|
| 15 |
+
"step": 10
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.09,
|
| 19 |
+
"eval_accuracy": 0.21164021164021163,
|
| 20 |
+
"eval_loss": 2.286055326461792,
|
| 21 |
+
"eval_runtime": 3.2931,
|
| 22 |
+
"eval_samples_per_second": 57.393,
|
| 23 |
+
"eval_steps_per_second": 7.288,
|
| 24 |
+
"step": 10
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.19,
|
| 28 |
+
"learning_rate": 0.00019766355140186917,
|
| 29 |
+
"loss": 2.3051,
|
| 30 |
+
"step": 20
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.19,
|
| 34 |
+
"eval_accuracy": 0.3439153439153439,
|
| 35 |
+
"eval_loss": 2.1907191276550293,
|
| 36 |
+
"eval_runtime": 3.5708,
|
| 37 |
+
"eval_samples_per_second": 52.929,
|
| 38 |
+
"eval_steps_per_second": 6.721,
|
| 39 |
+
"step": 20
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.28,
|
| 43 |
+
"learning_rate": 0.00019649532710280373,
|
| 44 |
+
"loss": 2.1219,
|
| 45 |
+
"step": 30
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.28,
|
| 49 |
+
"eval_accuracy": 0.31746031746031744,
|
| 50 |
+
"eval_loss": 2.021427631378174,
|
| 51 |
+
"eval_runtime": 3.5214,
|
| 52 |
+
"eval_samples_per_second": 53.671,
|
| 53 |
+
"eval_steps_per_second": 6.815,
|
| 54 |
+
"step": 30
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.37,
|
| 58 |
+
"learning_rate": 0.00019532710280373834,
|
| 59 |
+
"loss": 2.0542,
|
| 60 |
+
"step": 40
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 0.37,
|
| 64 |
+
"eval_accuracy": 0.4074074074074074,
|
| 65 |
+
"eval_loss": 1.905888319015503,
|
| 66 |
+
"eval_runtime": 3.4441,
|
| 67 |
+
"eval_samples_per_second": 54.877,
|
| 68 |
+
"eval_steps_per_second": 6.969,
|
| 69 |
+
"step": 40
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.47,
|
| 73 |
+
"learning_rate": 0.00019415887850467292,
|
| 74 |
+
"loss": 1.8132,
|
| 75 |
+
"step": 50
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.47,
|
| 79 |
+
"eval_accuracy": 0.3862433862433862,
|
| 80 |
+
"eval_loss": 1.8471989631652832,
|
| 81 |
+
"eval_runtime": 3.2687,
|
| 82 |
+
"eval_samples_per_second": 57.821,
|
| 83 |
+
"eval_steps_per_second": 7.342,
|
| 84 |
+
"step": 50
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.56,
|
| 88 |
+
"learning_rate": 0.0001929906542056075,
|
| 89 |
+
"loss": 1.8854,
|
| 90 |
+
"step": 60
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 0.56,
|
| 94 |
+
"eval_accuracy": 0.4603174603174603,
|
| 95 |
+
"eval_loss": 1.6832294464111328,
|
| 96 |
+
"eval_runtime": 3.1893,
|
| 97 |
+
"eval_samples_per_second": 59.261,
|
| 98 |
+
"eval_steps_per_second": 7.525,
|
| 99 |
+
"step": 60
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.65,
|
| 103 |
+
"learning_rate": 0.00019182242990654205,
|
| 104 |
+
"loss": 1.6981,
|
| 105 |
+
"step": 70
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.65,
|
| 109 |
+
"eval_accuracy": 0.4973544973544973,
|
| 110 |
+
"eval_loss": 1.600830078125,
|
| 111 |
+
"eval_runtime": 3.1721,
|
| 112 |
+
"eval_samples_per_second": 59.583,
|
| 113 |
+
"eval_steps_per_second": 7.566,
|
| 114 |
+
"step": 70
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.75,
|
| 118 |
+
"learning_rate": 0.00019065420560747664,
|
| 119 |
+
"loss": 1.5251,
|
| 120 |
+
"step": 80
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 0.75,
|
| 124 |
+
"eval_accuracy": 0.5026455026455027,
|
| 125 |
+
"eval_loss": 1.468518614768982,
|
| 126 |
+
"eval_runtime": 3.3583,
|
| 127 |
+
"eval_samples_per_second": 56.278,
|
| 128 |
+
"eval_steps_per_second": 7.146,
|
| 129 |
+
"step": 80
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.84,
|
| 133 |
+
"learning_rate": 0.00018948598130841122,
|
| 134 |
+
"loss": 1.4463,
|
| 135 |
+
"step": 90
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.84,
|
| 139 |
+
"eval_accuracy": 0.6137566137566137,
|
| 140 |
+
"eval_loss": 1.3712642192840576,
|
| 141 |
+
"eval_runtime": 3.2333,
|
| 142 |
+
"eval_samples_per_second": 58.454,
|
| 143 |
+
"eval_steps_per_second": 7.423,
|
| 144 |
+
"step": 90
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.93,
|
| 148 |
+
"learning_rate": 0.0001883177570093458,
|
| 149 |
+
"loss": 1.4335,
|
| 150 |
+
"step": 100
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.93,
|
| 154 |
+
"eval_accuracy": 0.4973544973544973,
|
| 155 |
+
"eval_loss": 1.4269553422927856,
|
| 156 |
+
"eval_runtime": 3.7842,
|
| 157 |
+
"eval_samples_per_second": 49.944,
|
| 158 |
+
"eval_steps_per_second": 6.342,
|
| 159 |
+
"step": 100
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 1.03,
|
| 163 |
+
"learning_rate": 0.00018726635514018693,
|
| 164 |
+
"loss": 1.1147,
|
| 165 |
+
"step": 110
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 1.03,
|
| 169 |
+
"eval_accuracy": 0.5925925925925926,
|
| 170 |
+
"eval_loss": 1.2793240547180176,
|
| 171 |
+
"eval_runtime": 3.3494,
|
| 172 |
+
"eval_samples_per_second": 56.427,
|
| 173 |
+
"eval_steps_per_second": 7.165,
|
| 174 |
+
"step": 110
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 1.12,
|
| 178 |
+
"learning_rate": 0.0001860981308411215,
|
| 179 |
+
"loss": 1.3568,
|
| 180 |
+
"step": 120
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 1.12,
|
| 184 |
+
"eval_accuracy": 0.5661375661375662,
|
| 185 |
+
"eval_loss": 1.3360047340393066,
|
| 186 |
+
"eval_runtime": 3.3593,
|
| 187 |
+
"eval_samples_per_second": 56.261,
|
| 188 |
+
"eval_steps_per_second": 7.144,
|
| 189 |
+
"step": 120
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 1.21,
|
| 193 |
+
"learning_rate": 0.0001849299065420561,
|
| 194 |
+
"loss": 1.3077,
|
| 195 |
+
"step": 130
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 1.21,
|
| 199 |
+
"eval_accuracy": 0.5079365079365079,
|
| 200 |
+
"eval_loss": 1.452016830444336,
|
| 201 |
+
"eval_runtime": 3.1904,
|
| 202 |
+
"eval_samples_per_second": 59.241,
|
| 203 |
+
"eval_steps_per_second": 7.523,
|
| 204 |
+
"step": 130
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 1.31,
|
| 208 |
+
"learning_rate": 0.00018376168224299065,
|
| 209 |
+
"loss": 1.2801,
|
| 210 |
+
"step": 140
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"epoch": 1.31,
|
| 214 |
+
"eval_accuracy": 0.5661375661375662,
|
| 215 |
+
"eval_loss": 1.2764976024627686,
|
| 216 |
+
"eval_runtime": 3.1251,
|
| 217 |
+
"eval_samples_per_second": 60.477,
|
| 218 |
+
"eval_steps_per_second": 7.68,
|
| 219 |
+
"step": 140
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 1.4,
|
| 223 |
+
"learning_rate": 0.00018259345794392523,
|
| 224 |
+
"loss": 1.2894,
|
| 225 |
+
"step": 150
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 1.4,
|
| 229 |
+
"eval_accuracy": 0.6137566137566137,
|
| 230 |
+
"eval_loss": 1.194944977760315,
|
| 231 |
+
"eval_runtime": 3.3657,
|
| 232 |
+
"eval_samples_per_second": 56.154,
|
| 233 |
+
"eval_steps_per_second": 7.131,
|
| 234 |
+
"step": 150
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 1.5,
|
| 238 |
+
"learning_rate": 0.0001814252336448598,
|
| 239 |
+
"loss": 1.2657,
|
| 240 |
+
"step": 160
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 1.5,
|
| 244 |
+
"eval_accuracy": 0.6349206349206349,
|
| 245 |
+
"eval_loss": 1.1937017440795898,
|
| 246 |
+
"eval_runtime": 3.4004,
|
| 247 |
+
"eval_samples_per_second": 55.582,
|
| 248 |
+
"eval_steps_per_second": 7.058,
|
| 249 |
+
"step": 160
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 1.59,
|
| 253 |
+
"learning_rate": 0.0001802570093457944,
|
| 254 |
+
"loss": 0.8784,
|
| 255 |
+
"step": 170
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 1.59,
|
| 259 |
+
"eval_accuracy": 0.6031746031746031,
|
| 260 |
+
"eval_loss": 1.2189592123031616,
|
| 261 |
+
"eval_runtime": 3.069,
|
| 262 |
+
"eval_samples_per_second": 61.583,
|
| 263 |
+
"eval_steps_per_second": 7.82,
|
| 264 |
+
"step": 170
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 1.68,
|
| 268 |
+
"learning_rate": 0.000179088785046729,
|
| 269 |
+
"loss": 1.1575,
|
| 270 |
+
"step": 180
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 1.68,
|
| 274 |
+
"eval_accuracy": 0.6137566137566137,
|
| 275 |
+
"eval_loss": 1.2268363237380981,
|
| 276 |
+
"eval_runtime": 3.1612,
|
| 277 |
+
"eval_samples_per_second": 59.788,
|
| 278 |
+
"eval_steps_per_second": 7.592,
|
| 279 |
+
"step": 180
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 1.78,
|
| 283 |
+
"learning_rate": 0.00017792056074766356,
|
| 284 |
+
"loss": 0.9848,
|
| 285 |
+
"step": 190
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 1.78,
|
| 289 |
+
"eval_accuracy": 0.656084656084656,
|
| 290 |
+
"eval_loss": 1.0571767091751099,
|
| 291 |
+
"eval_runtime": 3.0146,
|
| 292 |
+
"eval_samples_per_second": 62.695,
|
| 293 |
+
"eval_steps_per_second": 7.961,
|
| 294 |
+
"step": 190
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 1.87,
|
| 298 |
+
"learning_rate": 0.00017675233644859814,
|
| 299 |
+
"loss": 0.9409,
|
| 300 |
+
"step": 200
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"epoch": 1.87,
|
| 304 |
+
"eval_accuracy": 0.6349206349206349,
|
| 305 |
+
"eval_loss": 1.160926103591919,
|
| 306 |
+
"eval_runtime": 3.1953,
|
| 307 |
+
"eval_samples_per_second": 59.149,
|
| 308 |
+
"eval_steps_per_second": 7.511,
|
| 309 |
+
"step": 200
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 1.96,
|
| 313 |
+
"learning_rate": 0.00017558411214953272,
|
| 314 |
+
"loss": 0.9448,
|
| 315 |
+
"step": 210
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 1.96,
|
| 319 |
+
"eval_accuracy": 0.6084656084656085,
|
| 320 |
+
"eval_loss": 1.2327442169189453,
|
| 321 |
+
"eval_runtime": 3.0705,
|
| 322 |
+
"eval_samples_per_second": 61.554,
|
| 323 |
+
"eval_steps_per_second": 7.816,
|
| 324 |
+
"step": 210
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 2.06,
|
| 328 |
+
"learning_rate": 0.0001744158878504673,
|
| 329 |
+
"loss": 1.0819,
|
| 330 |
+
"step": 220
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"epoch": 2.06,
|
| 334 |
+
"eval_accuracy": 0.582010582010582,
|
| 335 |
+
"eval_loss": 1.169877290725708,
|
| 336 |
+
"eval_runtime": 3.2052,
|
| 337 |
+
"eval_samples_per_second": 58.967,
|
| 338 |
+
"eval_steps_per_second": 7.488,
|
| 339 |
+
"step": 220
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 2.15,
|
| 343 |
+
"learning_rate": 0.00017324766355140188,
|
| 344 |
+
"loss": 0.7485,
|
| 345 |
+
"step": 230
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 2.15,
|
| 349 |
+
"eval_accuracy": 0.6507936507936508,
|
| 350 |
+
"eval_loss": 1.1040607690811157,
|
| 351 |
+
"eval_runtime": 3.1129,
|
| 352 |
+
"eval_samples_per_second": 60.714,
|
| 353 |
+
"eval_steps_per_second": 7.71,
|
| 354 |
+
"step": 230
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 2.24,
|
| 358 |
+
"learning_rate": 0.00017207943925233644,
|
| 359 |
+
"loss": 0.8934,
|
| 360 |
+
"step": 240
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 2.24,
|
| 364 |
+
"eval_accuracy": 0.5873015873015873,
|
| 365 |
+
"eval_loss": 1.1672364473342896,
|
| 366 |
+
"eval_runtime": 3.2331,
|
| 367 |
+
"eval_samples_per_second": 58.458,
|
| 368 |
+
"eval_steps_per_second": 7.423,
|
| 369 |
+
"step": 240
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 2.34,
|
| 373 |
+
"learning_rate": 0.00017091121495327102,
|
| 374 |
+
"loss": 0.8609,
|
| 375 |
+
"step": 250
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 2.34,
|
| 379 |
+
"eval_accuracy": 0.6190476190476191,
|
| 380 |
+
"eval_loss": 1.1899827718734741,
|
| 381 |
+
"eval_runtime": 3.138,
|
| 382 |
+
"eval_samples_per_second": 60.23,
|
| 383 |
+
"eval_steps_per_second": 7.648,
|
| 384 |
+
"step": 250
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 2.43,
|
| 388 |
+
"learning_rate": 0.00016974299065420563,
|
| 389 |
+
"loss": 0.7935,
|
| 390 |
+
"step": 260
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 2.43,
|
| 394 |
+
"eval_accuracy": 0.6402116402116402,
|
| 395 |
+
"eval_loss": 1.0623040199279785,
|
| 396 |
+
"eval_runtime": 3.1191,
|
| 397 |
+
"eval_samples_per_second": 60.594,
|
| 398 |
+
"eval_steps_per_second": 7.694,
|
| 399 |
+
"step": 260
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 2.52,
|
| 403 |
+
"learning_rate": 0.0001685747663551402,
|
| 404 |
+
"loss": 0.8013,
|
| 405 |
+
"step": 270
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 2.52,
|
| 409 |
+
"eval_accuracy": 0.6878306878306878,
|
| 410 |
+
"eval_loss": 0.9872872233390808,
|
| 411 |
+
"eval_runtime": 3.0196,
|
| 412 |
+
"eval_samples_per_second": 62.59,
|
| 413 |
+
"eval_steps_per_second": 7.948,
|
| 414 |
+
"step": 270
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 2.62,
|
| 418 |
+
"learning_rate": 0.00016740654205607477,
|
| 419 |
+
"loss": 0.6669,
|
| 420 |
+
"step": 280
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 2.62,
|
| 424 |
+
"eval_accuracy": 0.656084656084656,
|
| 425 |
+
"eval_loss": 1.0078377723693848,
|
| 426 |
+
"eval_runtime": 3.0773,
|
| 427 |
+
"eval_samples_per_second": 61.418,
|
| 428 |
+
"eval_steps_per_second": 7.799,
|
| 429 |
+
"step": 280
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 2.71,
|
| 433 |
+
"learning_rate": 0.00016623831775700935,
|
| 434 |
+
"loss": 0.7847,
|
| 435 |
+
"step": 290
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 2.71,
|
| 439 |
+
"eval_accuracy": 0.6084656084656085,
|
| 440 |
+
"eval_loss": 1.148421049118042,
|
| 441 |
+
"eval_runtime": 3.1668,
|
| 442 |
+
"eval_samples_per_second": 59.682,
|
| 443 |
+
"eval_steps_per_second": 7.579,
|
| 444 |
+
"step": 290
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 2.8,
|
| 448 |
+
"learning_rate": 0.00016507009345794393,
|
| 449 |
+
"loss": 0.7222,
|
| 450 |
+
"step": 300
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 2.8,
|
| 454 |
+
"eval_accuracy": 0.6243386243386243,
|
| 455 |
+
"eval_loss": 1.1295346021652222,
|
| 456 |
+
"eval_runtime": 3.1049,
|
| 457 |
+
"eval_samples_per_second": 60.871,
|
| 458 |
+
"eval_steps_per_second": 7.73,
|
| 459 |
+
"step": 300
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 2.9,
|
| 463 |
+
"learning_rate": 0.0001639018691588785,
|
| 464 |
+
"loss": 0.7844,
|
| 465 |
+
"step": 310
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 2.9,
|
| 469 |
+
"eval_accuracy": 0.7248677248677249,
|
| 470 |
+
"eval_loss": 0.9414218068122864,
|
| 471 |
+
"eval_runtime": 3.1524,
|
| 472 |
+
"eval_samples_per_second": 59.954,
|
| 473 |
+
"eval_steps_per_second": 7.613,
|
| 474 |
+
"step": 310
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 2.99,
|
| 478 |
+
"learning_rate": 0.0001627336448598131,
|
| 479 |
+
"loss": 0.8057,
|
| 480 |
+
"step": 320
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 2.99,
|
| 484 |
+
"eval_accuracy": 0.6666666666666666,
|
| 485 |
+
"eval_loss": 1.050380825996399,
|
| 486 |
+
"eval_runtime": 3.1613,
|
| 487 |
+
"eval_samples_per_second": 59.785,
|
| 488 |
+
"eval_steps_per_second": 7.592,
|
| 489 |
+
"step": 320
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 3.08,
|
| 493 |
+
"learning_rate": 0.00016156542056074767,
|
| 494 |
+
"loss": 0.4843,
|
| 495 |
+
"step": 330
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 3.08,
|
| 499 |
+
"eval_accuracy": 0.6507936507936508,
|
| 500 |
+
"eval_loss": 0.9873789548873901,
|
| 501 |
+
"eval_runtime": 3.1677,
|
| 502 |
+
"eval_samples_per_second": 59.664,
|
| 503 |
+
"eval_steps_per_second": 7.576,
|
| 504 |
+
"step": 330
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 3.18,
|
| 508 |
+
"learning_rate": 0.00016039719626168226,
|
| 509 |
+
"loss": 0.6766,
|
| 510 |
+
"step": 340
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 3.18,
|
| 514 |
+
"eval_accuracy": 0.6507936507936508,
|
| 515 |
+
"eval_loss": 1.1496278047561646,
|
| 516 |
+
"eval_runtime": 3.0545,
|
| 517 |
+
"eval_samples_per_second": 61.877,
|
| 518 |
+
"eval_steps_per_second": 7.857,
|
| 519 |
+
"step": 340
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 3.27,
|
| 523 |
+
"learning_rate": 0.00015922897196261684,
|
| 524 |
+
"loss": 0.4818,
|
| 525 |
+
"step": 350
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 3.27,
|
| 529 |
+
"eval_accuracy": 0.6878306878306878,
|
| 530 |
+
"eval_loss": 1.0967848300933838,
|
| 531 |
+
"eval_runtime": 3.1038,
|
| 532 |
+
"eval_samples_per_second": 60.892,
|
| 533 |
+
"eval_steps_per_second": 7.732,
|
| 534 |
+
"step": 350
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 3.36,
|
| 538 |
+
"learning_rate": 0.00015806074766355142,
|
| 539 |
+
"loss": 0.5351,
|
| 540 |
+
"step": 360
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 3.36,
|
| 544 |
+
"eval_accuracy": 0.6296296296296297,
|
| 545 |
+
"eval_loss": 1.1393951177597046,
|
| 546 |
+
"eval_runtime": 3.0544,
|
| 547 |
+
"eval_samples_per_second": 61.877,
|
| 548 |
+
"eval_steps_per_second": 7.857,
|
| 549 |
+
"step": 360
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 3.46,
|
| 553 |
+
"learning_rate": 0.000156892523364486,
|
| 554 |
+
"loss": 0.5035,
|
| 555 |
+
"step": 370
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 3.46,
|
| 559 |
+
"eval_accuracy": 0.708994708994709,
|
| 560 |
+
"eval_loss": 0.9814600944519043,
|
| 561 |
+
"eval_runtime": 3.6264,
|
| 562 |
+
"eval_samples_per_second": 52.118,
|
| 563 |
+
"eval_steps_per_second": 6.618,
|
| 564 |
+
"step": 370
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 3.55,
|
| 568 |
+
"learning_rate": 0.00015572429906542056,
|
| 569 |
+
"loss": 0.4032,
|
| 570 |
+
"step": 380
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 3.55,
|
| 574 |
+
"eval_accuracy": 0.6402116402116402,
|
| 575 |
+
"eval_loss": 1.0881891250610352,
|
| 576 |
+
"eval_runtime": 3.1861,
|
| 577 |
+
"eval_samples_per_second": 59.32,
|
| 578 |
+
"eval_steps_per_second": 7.533,
|
| 579 |
+
"step": 380
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 3.64,
|
| 583 |
+
"learning_rate": 0.00015455607476635514,
|
| 584 |
+
"loss": 0.639,
|
| 585 |
+
"step": 390
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 3.64,
|
| 589 |
+
"eval_accuracy": 0.6084656084656085,
|
| 590 |
+
"eval_loss": 1.2611372470855713,
|
| 591 |
+
"eval_runtime": 3.2278,
|
| 592 |
+
"eval_samples_per_second": 58.554,
|
| 593 |
+
"eval_steps_per_second": 7.435,
|
| 594 |
+
"step": 390
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 3.74,
|
| 598 |
+
"learning_rate": 0.00015338785046728972,
|
| 599 |
+
"loss": 0.5156,
|
| 600 |
+
"step": 400
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 3.74,
|
| 604 |
+
"eval_accuracy": 0.656084656084656,
|
| 605 |
+
"eval_loss": 1.037644624710083,
|
| 606 |
+
"eval_runtime": 3.6337,
|
| 607 |
+
"eval_samples_per_second": 52.013,
|
| 608 |
+
"eval_steps_per_second": 6.605,
|
| 609 |
+
"step": 400
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 3.83,
|
| 613 |
+
"learning_rate": 0.00015221962616822433,
|
| 614 |
+
"loss": 0.4884,
|
| 615 |
+
"step": 410
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 3.83,
|
| 619 |
+
"eval_accuracy": 0.6984126984126984,
|
| 620 |
+
"eval_loss": 0.950617790222168,
|
| 621 |
+
"eval_runtime": 3.4889,
|
| 622 |
+
"eval_samples_per_second": 54.172,
|
| 623 |
+
"eval_steps_per_second": 6.879,
|
| 624 |
+
"step": 410
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 3.93,
|
| 628 |
+
"learning_rate": 0.00015105140186915888,
|
| 629 |
+
"loss": 0.5875,
|
| 630 |
+
"step": 420
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 3.93,
|
| 634 |
+
"eval_accuracy": 0.7513227513227513,
|
| 635 |
+
"eval_loss": 0.8479233980178833,
|
| 636 |
+
"eval_runtime": 3.0768,
|
| 637 |
+
"eval_samples_per_second": 61.428,
|
| 638 |
+
"eval_steps_per_second": 7.8,
|
| 639 |
+
"step": 420
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 4.02,
|
| 643 |
+
"learning_rate": 0.00014988317757009346,
|
| 644 |
+
"loss": 0.6982,
|
| 645 |
+
"step": 430
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 4.02,
|
| 649 |
+
"eval_accuracy": 0.6825396825396826,
|
| 650 |
+
"eval_loss": 1.0894999504089355,
|
| 651 |
+
"eval_runtime": 3.1014,
|
| 652 |
+
"eval_samples_per_second": 60.941,
|
| 653 |
+
"eval_steps_per_second": 7.738,
|
| 654 |
+
"step": 430
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 4.11,
|
| 658 |
+
"learning_rate": 0.00014871495327102805,
|
| 659 |
+
"loss": 0.3966,
|
| 660 |
+
"step": 440
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 4.11,
|
| 664 |
+
"eval_accuracy": 0.6984126984126984,
|
| 665 |
+
"eval_loss": 0.9708942770957947,
|
| 666 |
+
"eval_runtime": 3.0279,
|
| 667 |
+
"eval_samples_per_second": 62.42,
|
| 668 |
+
"eval_steps_per_second": 7.926,
|
| 669 |
+
"step": 440
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 4.21,
|
| 673 |
+
"learning_rate": 0.00014754672897196263,
|
| 674 |
+
"loss": 0.377,
|
| 675 |
+
"step": 450
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 4.21,
|
| 679 |
+
"eval_accuracy": 0.6772486772486772,
|
| 680 |
+
"eval_loss": 0.9754199385643005,
|
| 681 |
+
"eval_runtime": 3.1346,
|
| 682 |
+
"eval_samples_per_second": 60.295,
|
| 683 |
+
"eval_steps_per_second": 7.657,
|
| 684 |
+
"step": 450
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 4.3,
|
| 688 |
+
"learning_rate": 0.0001463785046728972,
|
| 689 |
+
"loss": 0.3417,
|
| 690 |
+
"step": 460
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 4.3,
|
| 694 |
+
"eval_accuracy": 0.6507936507936508,
|
| 695 |
+
"eval_loss": 1.1686595678329468,
|
| 696 |
+
"eval_runtime": 3.0945,
|
| 697 |
+
"eval_samples_per_second": 61.076,
|
| 698 |
+
"eval_steps_per_second": 7.756,
|
| 699 |
+
"step": 460
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 4.39,
|
| 703 |
+
"learning_rate": 0.00014521028037383176,
|
| 704 |
+
"loss": 0.336,
|
| 705 |
+
"step": 470
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 4.39,
|
| 709 |
+
"eval_accuracy": 0.6984126984126984,
|
| 710 |
+
"eval_loss": 0.9825548529624939,
|
| 711 |
+
"eval_runtime": 3.1296,
|
| 712 |
+
"eval_samples_per_second": 60.39,
|
| 713 |
+
"eval_steps_per_second": 7.669,
|
| 714 |
+
"step": 470
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 4.49,
|
| 718 |
+
"learning_rate": 0.00014404205607476635,
|
| 719 |
+
"loss": 0.5201,
|
| 720 |
+
"step": 480
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 4.49,
|
| 724 |
+
"eval_accuracy": 0.6613756613756614,
|
| 725 |
+
"eval_loss": 1.177049160003662,
|
| 726 |
+
"eval_runtime": 3.0811,
|
| 727 |
+
"eval_samples_per_second": 61.341,
|
| 728 |
+
"eval_steps_per_second": 7.789,
|
| 729 |
+
"step": 480
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 4.58,
|
| 733 |
+
"learning_rate": 0.00014287383177570095,
|
| 734 |
+
"loss": 0.1737,
|
| 735 |
+
"step": 490
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 4.58,
|
| 739 |
+
"eval_accuracy": 0.6878306878306878,
|
| 740 |
+
"eval_loss": 1.0491447448730469,
|
| 741 |
+
"eval_runtime": 3.1745,
|
| 742 |
+
"eval_samples_per_second": 59.537,
|
| 743 |
+
"eval_steps_per_second": 7.56,
|
| 744 |
+
"step": 490
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 4.67,
|
| 748 |
+
"learning_rate": 0.00014170560747663554,
|
| 749 |
+
"loss": 0.2545,
|
| 750 |
+
"step": 500
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"epoch": 4.67,
|
| 754 |
+
"eval_accuracy": 0.6984126984126984,
|
| 755 |
+
"eval_loss": 1.1352397203445435,
|
| 756 |
+
"eval_runtime": 3.0543,
|
| 757 |
+
"eval_samples_per_second": 61.879,
|
| 758 |
+
"eval_steps_per_second": 7.858,
|
| 759 |
+
"step": 500
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 4.77,
|
| 763 |
+
"learning_rate": 0.0001405373831775701,
|
| 764 |
+
"loss": 0.3752,
|
| 765 |
+
"step": 510
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 4.77,
|
| 769 |
+
"eval_accuracy": 0.6931216931216931,
|
| 770 |
+
"eval_loss": 1.0300217866897583,
|
| 771 |
+
"eval_runtime": 3.2562,
|
| 772 |
+
"eval_samples_per_second": 58.043,
|
| 773 |
+
"eval_steps_per_second": 7.371,
|
| 774 |
+
"step": 510
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 4.86,
|
| 778 |
+
"learning_rate": 0.00013936915887850467,
|
| 779 |
+
"loss": 0.3667,
|
| 780 |
+
"step": 520
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 4.86,
|
| 784 |
+
"eval_accuracy": 0.6825396825396826,
|
| 785 |
+
"eval_loss": 1.0355250835418701,
|
| 786 |
+
"eval_runtime": 3.0388,
|
| 787 |
+
"eval_samples_per_second": 62.196,
|
| 788 |
+
"eval_steps_per_second": 7.898,
|
| 789 |
+
"step": 520
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 4.95,
|
| 793 |
+
"learning_rate": 0.00013820093457943925,
|
| 794 |
+
"loss": 0.2797,
|
| 795 |
+
"step": 530
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 4.95,
|
| 799 |
+
"eval_accuracy": 0.6984126984126984,
|
| 800 |
+
"eval_loss": 0.9882386326789856,
|
| 801 |
+
"eval_runtime": 3.3521,
|
| 802 |
+
"eval_samples_per_second": 56.383,
|
| 803 |
+
"eval_steps_per_second": 7.16,
|
| 804 |
+
"step": 530
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 5.05,
|
| 808 |
+
"learning_rate": 0.00013703271028037384,
|
| 809 |
+
"loss": 0.1646,
|
| 810 |
+
"step": 540
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 5.05,
|
| 814 |
+
"eval_accuracy": 0.6984126984126984,
|
| 815 |
+
"eval_loss": 1.0728471279144287,
|
| 816 |
+
"eval_runtime": 3.1345,
|
| 817 |
+
"eval_samples_per_second": 60.297,
|
| 818 |
+
"eval_steps_per_second": 7.657,
|
| 819 |
+
"step": 540
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 5.14,
|
| 823 |
+
"learning_rate": 0.00013586448598130842,
|
| 824 |
+
"loss": 0.2199,
|
| 825 |
+
"step": 550
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 5.14,
|
| 829 |
+
"eval_accuracy": 0.7566137566137566,
|
| 830 |
+
"eval_loss": 0.8328016996383667,
|
| 831 |
+
"eval_runtime": 3.3035,
|
| 832 |
+
"eval_samples_per_second": 57.213,
|
| 833 |
+
"eval_steps_per_second": 7.265,
|
| 834 |
+
"step": 550
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 5.23,
|
| 838 |
+
"learning_rate": 0.000134696261682243,
|
| 839 |
+
"loss": 0.2191,
|
| 840 |
+
"step": 560
|
| 841 |
+
},
|
| 842 |
+
{
|
| 843 |
+
"epoch": 5.23,
|
| 844 |
+
"eval_accuracy": 0.746031746031746,
|
| 845 |
+
"eval_loss": 0.9280108213424683,
|
| 846 |
+
"eval_runtime": 3.0545,
|
| 847 |
+
"eval_samples_per_second": 61.876,
|
| 848 |
+
"eval_steps_per_second": 7.857,
|
| 849 |
+
"step": 560
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 5.33,
|
| 853 |
+
"learning_rate": 0.00013352803738317758,
|
| 854 |
+
"loss": 0.12,
|
| 855 |
+
"step": 570
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 5.33,
|
| 859 |
+
"eval_accuracy": 0.7037037037037037,
|
| 860 |
+
"eval_loss": 1.0978211164474487,
|
| 861 |
+
"eval_runtime": 3.1612,
|
| 862 |
+
"eval_samples_per_second": 59.788,
|
| 863 |
+
"eval_steps_per_second": 7.592,
|
| 864 |
+
"step": 570
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 5.42,
|
| 868 |
+
"learning_rate": 0.00013235981308411216,
|
| 869 |
+
"loss": 0.2608,
|
| 870 |
+
"step": 580
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 5.42,
|
| 874 |
+
"eval_accuracy": 0.6878306878306878,
|
| 875 |
+
"eval_loss": 1.1158148050308228,
|
| 876 |
+
"eval_runtime": 3.1212,
|
| 877 |
+
"eval_samples_per_second": 60.553,
|
| 878 |
+
"eval_steps_per_second": 7.689,
|
| 879 |
+
"step": 580
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 5.51,
|
| 883 |
+
"learning_rate": 0.00013119158878504674,
|
| 884 |
+
"loss": 0.2,
|
| 885 |
+
"step": 590
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 5.51,
|
| 889 |
+
"eval_accuracy": 0.7354497354497355,
|
| 890 |
+
"eval_loss": 1.087318778038025,
|
| 891 |
+
"eval_runtime": 3.0411,
|
| 892 |
+
"eval_samples_per_second": 62.148,
|
| 893 |
+
"eval_steps_per_second": 7.892,
|
| 894 |
+
"step": 590
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 5.61,
|
| 898 |
+
"learning_rate": 0.00013002336448598133,
|
| 899 |
+
"loss": 0.1899,
|
| 900 |
+
"step": 600
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 5.61,
|
| 904 |
+
"eval_accuracy": 0.7142857142857143,
|
| 905 |
+
"eval_loss": 1.0559666156768799,
|
| 906 |
+
"eval_runtime": 3.1128,
|
| 907 |
+
"eval_samples_per_second": 60.718,
|
| 908 |
+
"eval_steps_per_second": 7.71,
|
| 909 |
+
"step": 600
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 5.7,
|
| 913 |
+
"learning_rate": 0.00012885514018691588,
|
| 914 |
+
"loss": 0.1113,
|
| 915 |
+
"step": 610
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 5.7,
|
| 919 |
+
"eval_accuracy": 0.7037037037037037,
|
| 920 |
+
"eval_loss": 1.1143887042999268,
|
| 921 |
+
"eval_runtime": 3.038,
|
| 922 |
+
"eval_samples_per_second": 62.211,
|
| 923 |
+
"eval_steps_per_second": 7.9,
|
| 924 |
+
"step": 610
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"epoch": 5.79,
|
| 928 |
+
"learning_rate": 0.00012768691588785046,
|
| 929 |
+
"loss": 0.2279,
|
| 930 |
+
"step": 620
|
| 931 |
+
},
|
| 932 |
+
{
|
| 933 |
+
"epoch": 5.79,
|
| 934 |
+
"eval_accuracy": 0.6666666666666666,
|
| 935 |
+
"eval_loss": 1.2534881830215454,
|
| 936 |
+
"eval_runtime": 3.4146,
|
| 937 |
+
"eval_samples_per_second": 55.351,
|
| 938 |
+
"eval_steps_per_second": 7.029,
|
| 939 |
+
"step": 620
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 5.89,
|
| 943 |
+
"learning_rate": 0.00012651869158878504,
|
| 944 |
+
"loss": 0.1563,
|
| 945 |
+
"step": 630
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 5.89,
|
| 949 |
+
"eval_accuracy": 0.7354497354497355,
|
| 950 |
+
"eval_loss": 1.0802773237228394,
|
| 951 |
+
"eval_runtime": 3.0678,
|
| 952 |
+
"eval_samples_per_second": 61.608,
|
| 953 |
+
"eval_steps_per_second": 7.823,
|
| 954 |
+
"step": 630
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 5.98,
|
| 958 |
+
"learning_rate": 0.00012535046728971965,
|
| 959 |
+
"loss": 0.2182,
|
| 960 |
+
"step": 640
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 5.98,
|
| 964 |
+
"eval_accuracy": 0.6349206349206349,
|
| 965 |
+
"eval_loss": 1.3904473781585693,
|
| 966 |
+
"eval_runtime": 3.1818,
|
| 967 |
+
"eval_samples_per_second": 59.399,
|
| 968 |
+
"eval_steps_per_second": 7.543,
|
| 969 |
+
"step": 640
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 6.07,
|
| 973 |
+
"learning_rate": 0.0001241822429906542,
|
| 974 |
+
"loss": 0.1781,
|
| 975 |
+
"step": 650
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 6.07,
|
| 979 |
+
"eval_accuracy": 0.671957671957672,
|
| 980 |
+
"eval_loss": 1.3461418151855469,
|
| 981 |
+
"eval_runtime": 3.0144,
|
| 982 |
+
"eval_samples_per_second": 62.699,
|
| 983 |
+
"eval_steps_per_second": 7.962,
|
| 984 |
+
"step": 650
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 6.17,
|
| 988 |
+
"learning_rate": 0.0001230140186915888,
|
| 989 |
+
"loss": 0.1395,
|
| 990 |
+
"step": 660
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 6.17,
|
| 994 |
+
"eval_accuracy": 0.6825396825396826,
|
| 995 |
+
"eval_loss": 1.276949405670166,
|
| 996 |
+
"eval_runtime": 3.1879,
|
| 997 |
+
"eval_samples_per_second": 59.287,
|
| 998 |
+
"eval_steps_per_second": 7.529,
|
| 999 |
+
"step": 660
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 6.26,
|
| 1003 |
+
"learning_rate": 0.00012184579439252337,
|
| 1004 |
+
"loss": 0.2308,
|
| 1005 |
+
"step": 670
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 6.26,
|
| 1009 |
+
"eval_accuracy": 0.6931216931216931,
|
| 1010 |
+
"eval_loss": 1.2213280200958252,
|
| 1011 |
+
"eval_runtime": 3.1475,
|
| 1012 |
+
"eval_samples_per_second": 60.047,
|
| 1013 |
+
"eval_steps_per_second": 7.625,
|
| 1014 |
+
"step": 670
|
| 1015 |
+
},
|
| 1016 |
+
{
|
| 1017 |
+
"epoch": 6.36,
|
| 1018 |
+
"learning_rate": 0.00012067757009345795,
|
| 1019 |
+
"loss": 0.1899,
|
| 1020 |
+
"step": 680
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 6.36,
|
| 1024 |
+
"eval_accuracy": 0.7142857142857143,
|
| 1025 |
+
"eval_loss": 1.0947775840759277,
|
| 1026 |
+
"eval_runtime": 3.3614,
|
| 1027 |
+
"eval_samples_per_second": 56.226,
|
| 1028 |
+
"eval_steps_per_second": 7.14,
|
| 1029 |
+
"step": 680
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 6.45,
|
| 1033 |
+
"learning_rate": 0.00011950934579439252,
|
| 1034 |
+
"loss": 0.1702,
|
| 1035 |
+
"step": 690
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 6.45,
|
| 1039 |
+
"eval_accuracy": 0.6931216931216931,
|
| 1040 |
+
"eval_loss": 1.2383404970169067,
|
| 1041 |
+
"eval_runtime": 3.5678,
|
| 1042 |
+
"eval_samples_per_second": 52.974,
|
| 1043 |
+
"eval_steps_per_second": 6.727,
|
| 1044 |
+
"step": 690
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"epoch": 6.54,
|
| 1048 |
+
"learning_rate": 0.0001183411214953271,
|
| 1049 |
+
"loss": 0.1055,
|
| 1050 |
+
"step": 700
|
| 1051 |
+
},
|
| 1052 |
+
{
|
| 1053 |
+
"epoch": 6.54,
|
| 1054 |
+
"eval_accuracy": 0.6349206349206349,
|
| 1055 |
+
"eval_loss": 1.4009695053100586,
|
| 1056 |
+
"eval_runtime": 3.1702,
|
| 1057 |
+
"eval_samples_per_second": 59.618,
|
| 1058 |
+
"eval_steps_per_second": 7.571,
|
| 1059 |
+
"step": 700
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 6.64,
|
| 1063 |
+
"learning_rate": 0.00011717289719626168,
|
| 1064 |
+
"loss": 0.1151,
|
| 1065 |
+
"step": 710
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 6.64,
|
| 1069 |
+
"eval_accuracy": 0.671957671957672,
|
| 1070 |
+
"eval_loss": 1.2606865167617798,
|
| 1071 |
+
"eval_runtime": 3.2176,
|
| 1072 |
+
"eval_samples_per_second": 58.74,
|
| 1073 |
+
"eval_steps_per_second": 7.459,
|
| 1074 |
+
"step": 710
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 6.73,
|
| 1078 |
+
"learning_rate": 0.00011600467289719628,
|
| 1079 |
+
"loss": 0.2415,
|
| 1080 |
+
"step": 720
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 6.73,
|
| 1084 |
+
"eval_accuracy": 0.7301587301587301,
|
| 1085 |
+
"eval_loss": 1.0520366430282593,
|
| 1086 |
+
"eval_runtime": 3.1343,
|
| 1087 |
+
"eval_samples_per_second": 60.3,
|
| 1088 |
+
"eval_steps_per_second": 7.657,
|
| 1089 |
+
"step": 720
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 6.82,
|
| 1093 |
+
"learning_rate": 0.00011483644859813085,
|
| 1094 |
+
"loss": 0.117,
|
| 1095 |
+
"step": 730
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 6.82,
|
| 1099 |
+
"eval_accuracy": 0.7354497354497355,
|
| 1100 |
+
"eval_loss": 1.0547845363616943,
|
| 1101 |
+
"eval_runtime": 3.1212,
|
| 1102 |
+
"eval_samples_per_second": 60.554,
|
| 1103 |
+
"eval_steps_per_second": 7.689,
|
| 1104 |
+
"step": 730
|
| 1105 |
+
},
|
| 1106 |
+
{
|
| 1107 |
+
"epoch": 6.92,
|
| 1108 |
+
"learning_rate": 0.00011366822429906543,
|
| 1109 |
+
"loss": 0.184,
|
| 1110 |
+
"step": 740
|
| 1111 |
+
},
|
| 1112 |
+
{
|
| 1113 |
+
"epoch": 6.92,
|
| 1114 |
+
"eval_accuracy": 0.6984126984126984,
|
| 1115 |
+
"eval_loss": 1.1872239112854004,
|
| 1116 |
+
"eval_runtime": 3.2412,
|
| 1117 |
+
"eval_samples_per_second": 58.311,
|
| 1118 |
+
"eval_steps_per_second": 7.405,
|
| 1119 |
+
"step": 740
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 7.01,
|
| 1123 |
+
"learning_rate": 0.00011250000000000001,
|
| 1124 |
+
"loss": 0.1997,
|
| 1125 |
+
"step": 750
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 7.01,
|
| 1129 |
+
"eval_accuracy": 0.7248677248677249,
|
| 1130 |
+
"eval_loss": 1.1127938032150269,
|
| 1131 |
+
"eval_runtime": 3.2145,
|
| 1132 |
+
"eval_samples_per_second": 58.796,
|
| 1133 |
+
"eval_steps_per_second": 7.466,
|
| 1134 |
+
"step": 750
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 7.1,
|
| 1138 |
+
"learning_rate": 0.00011133177570093458,
|
| 1139 |
+
"loss": 0.0645,
|
| 1140 |
+
"step": 760
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 7.1,
|
| 1144 |
+
"eval_accuracy": 0.6984126984126984,
|
| 1145 |
+
"eval_loss": 1.1514036655426025,
|
| 1146 |
+
"eval_runtime": 4.5617,
|
| 1147 |
+
"eval_samples_per_second": 41.432,
|
| 1148 |
+
"eval_steps_per_second": 5.261,
|
| 1149 |
+
"step": 760
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 7.2,
|
| 1153 |
+
"learning_rate": 0.00011016355140186916,
|
| 1154 |
+
"loss": 0.1025,
|
| 1155 |
+
"step": 770
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 7.2,
|
| 1159 |
+
"eval_accuracy": 0.7037037037037037,
|
| 1160 |
+
"eval_loss": 1.2251726388931274,
|
| 1161 |
+
"eval_runtime": 3.3597,
|
| 1162 |
+
"eval_samples_per_second": 56.255,
|
| 1163 |
+
"eval_steps_per_second": 7.143,
|
| 1164 |
+
"step": 770
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 7.29,
|
| 1168 |
+
"learning_rate": 0.00010899532710280373,
|
| 1169 |
+
"loss": 0.0407,
|
| 1170 |
+
"step": 780
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"epoch": 7.29,
|
| 1174 |
+
"eval_accuracy": 0.7513227513227513,
|
| 1175 |
+
"eval_loss": 1.0571043491363525,
|
| 1176 |
+
"eval_runtime": 3.3442,
|
| 1177 |
+
"eval_samples_per_second": 56.516,
|
| 1178 |
+
"eval_steps_per_second": 7.177,
|
| 1179 |
+
"step": 780
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 7.38,
|
| 1183 |
+
"learning_rate": 0.00010782710280373834,
|
| 1184 |
+
"loss": 0.1752,
|
| 1185 |
+
"step": 790
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 7.38,
|
| 1189 |
+
"eval_accuracy": 0.7354497354497355,
|
| 1190 |
+
"eval_loss": 1.081169843673706,
|
| 1191 |
+
"eval_runtime": 3.2833,
|
| 1192 |
+
"eval_samples_per_second": 57.564,
|
| 1193 |
+
"eval_steps_per_second": 7.31,
|
| 1194 |
+
"step": 790
|
| 1195 |
+
},
|
| 1196 |
+
{
|
| 1197 |
+
"epoch": 7.48,
|
| 1198 |
+
"learning_rate": 0.0001066588785046729,
|
| 1199 |
+
"loss": 0.1143,
|
| 1200 |
+
"step": 800
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 7.48,
|
| 1204 |
+
"eval_accuracy": 0.7142857142857143,
|
| 1205 |
+
"eval_loss": 1.2182053327560425,
|
| 1206 |
+
"eval_runtime": 3.3002,
|
| 1207 |
+
"eval_samples_per_second": 57.269,
|
| 1208 |
+
"eval_steps_per_second": 7.272,
|
| 1209 |
+
"step": 800
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 7.57,
|
| 1213 |
+
"learning_rate": 0.00010549065420560749,
|
| 1214 |
+
"loss": 0.1542,
|
| 1215 |
+
"step": 810
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 7.57,
|
| 1219 |
+
"eval_accuracy": 0.7142857142857143,
|
| 1220 |
+
"eval_loss": 1.1789014339447021,
|
| 1221 |
+
"eval_runtime": 3.2576,
|
| 1222 |
+
"eval_samples_per_second": 58.018,
|
| 1223 |
+
"eval_steps_per_second": 7.367,
|
| 1224 |
+
"step": 810
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 7.66,
|
| 1228 |
+
"learning_rate": 0.00010432242990654207,
|
| 1229 |
+
"loss": 0.0859,
|
| 1230 |
+
"step": 820
|
| 1231 |
+
},
|
| 1232 |
+
{
|
| 1233 |
+
"epoch": 7.66,
|
| 1234 |
+
"eval_accuracy": 0.7195767195767195,
|
| 1235 |
+
"eval_loss": 1.1392204761505127,
|
| 1236 |
+
"eval_runtime": 3.3449,
|
| 1237 |
+
"eval_samples_per_second": 56.505,
|
| 1238 |
+
"eval_steps_per_second": 7.175,
|
| 1239 |
+
"step": 820
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 7.76,
|
| 1243 |
+
"learning_rate": 0.00010315420560747664,
|
| 1244 |
+
"loss": 0.119,
|
| 1245 |
+
"step": 830
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 7.76,
|
| 1249 |
+
"eval_accuracy": 0.7354497354497355,
|
| 1250 |
+
"eval_loss": 1.1567529439926147,
|
| 1251 |
+
"eval_runtime": 3.3747,
|
| 1252 |
+
"eval_samples_per_second": 56.004,
|
| 1253 |
+
"eval_steps_per_second": 7.112,
|
| 1254 |
+
"step": 830
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"epoch": 7.85,
|
| 1258 |
+
"learning_rate": 0.00010198598130841122,
|
| 1259 |
+
"loss": 0.0913,
|
| 1260 |
+
"step": 840
|
| 1261 |
+
},
|
| 1262 |
+
{
|
| 1263 |
+
"epoch": 7.85,
|
| 1264 |
+
"eval_accuracy": 0.6984126984126984,
|
| 1265 |
+
"eval_loss": 1.109655499458313,
|
| 1266 |
+
"eval_runtime": 3.3749,
|
| 1267 |
+
"eval_samples_per_second": 56.001,
|
| 1268 |
+
"eval_steps_per_second": 7.111,
|
| 1269 |
+
"step": 840
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 7.94,
|
| 1273 |
+
"learning_rate": 0.00010081775700934579,
|
| 1274 |
+
"loss": 0.085,
|
| 1275 |
+
"step": 850
|
| 1276 |
+
},
|
| 1277 |
+
{
|
| 1278 |
+
"epoch": 7.94,
|
| 1279 |
+
"eval_accuracy": 0.746031746031746,
|
| 1280 |
+
"eval_loss": 1.118941068649292,
|
| 1281 |
+
"eval_runtime": 3.2544,
|
| 1282 |
+
"eval_samples_per_second": 58.076,
|
| 1283 |
+
"eval_steps_per_second": 7.375,
|
| 1284 |
+
"step": 850
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 8.04,
|
| 1288 |
+
"learning_rate": 9.964953271028038e-05,
|
| 1289 |
+
"loss": 0.0201,
|
| 1290 |
+
"step": 860
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 8.04,
|
| 1294 |
+
"eval_accuracy": 0.7142857142857143,
|
| 1295 |
+
"eval_loss": 1.1282655000686646,
|
| 1296 |
+
"eval_runtime": 3.2809,
|
| 1297 |
+
"eval_samples_per_second": 57.606,
|
| 1298 |
+
"eval_steps_per_second": 7.315,
|
| 1299 |
+
"step": 860
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 8.13,
|
| 1303 |
+
"learning_rate": 9.848130841121495e-05,
|
| 1304 |
+
"loss": 0.0509,
|
| 1305 |
+
"step": 870
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 8.13,
|
| 1309 |
+
"eval_accuracy": 0.7407407407407407,
|
| 1310 |
+
"eval_loss": 1.1005280017852783,
|
| 1311 |
+
"eval_runtime": 3.2115,
|
| 1312 |
+
"eval_samples_per_second": 58.851,
|
| 1313 |
+
"eval_steps_per_second": 7.473,
|
| 1314 |
+
"step": 870
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"epoch": 8.22,
|
| 1318 |
+
"learning_rate": 9.731308411214953e-05,
|
| 1319 |
+
"loss": 0.0326,
|
| 1320 |
+
"step": 880
|
| 1321 |
+
},
|
| 1322 |
+
{
|
| 1323 |
+
"epoch": 8.22,
|
| 1324 |
+
"eval_accuracy": 0.7301587301587301,
|
| 1325 |
+
"eval_loss": 1.0489766597747803,
|
| 1326 |
+
"eval_runtime": 3.6148,
|
| 1327 |
+
"eval_samples_per_second": 52.285,
|
| 1328 |
+
"eval_steps_per_second": 6.639,
|
| 1329 |
+
"step": 880
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 8.32,
|
| 1333 |
+
"learning_rate": 9.614485981308413e-05,
|
| 1334 |
+
"loss": 0.0728,
|
| 1335 |
+
"step": 890
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 8.32,
|
| 1339 |
+
"eval_accuracy": 0.7195767195767195,
|
| 1340 |
+
"eval_loss": 1.2510557174682617,
|
| 1341 |
+
"eval_runtime": 3.2769,
|
| 1342 |
+
"eval_samples_per_second": 57.676,
|
| 1343 |
+
"eval_steps_per_second": 7.324,
|
| 1344 |
+
"step": 890
|
| 1345 |
+
},
|
| 1346 |
+
{
|
| 1347 |
+
"epoch": 8.41,
|
| 1348 |
+
"learning_rate": 9.49766355140187e-05,
|
| 1349 |
+
"loss": 0.0486,
|
| 1350 |
+
"step": 900
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 8.41,
|
| 1354 |
+
"eval_accuracy": 0.7142857142857143,
|
| 1355 |
+
"eval_loss": 1.1832597255706787,
|
| 1356 |
+
"eval_runtime": 3.441,
|
| 1357 |
+
"eval_samples_per_second": 54.925,
|
| 1358 |
+
"eval_steps_per_second": 6.975,
|
| 1359 |
+
"step": 900
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 8.5,
|
| 1363 |
+
"learning_rate": 9.380841121495328e-05,
|
| 1364 |
+
"loss": 0.0645,
|
| 1365 |
+
"step": 910
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 8.5,
|
| 1369 |
+
"eval_accuracy": 0.7724867724867724,
|
| 1370 |
+
"eval_loss": 0.9881472587585449,
|
| 1371 |
+
"eval_runtime": 3.3474,
|
| 1372 |
+
"eval_samples_per_second": 56.463,
|
| 1373 |
+
"eval_steps_per_second": 7.17,
|
| 1374 |
+
"step": 910
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 8.6,
|
| 1378 |
+
"learning_rate": 9.264018691588785e-05,
|
| 1379 |
+
"loss": 0.0194,
|
| 1380 |
+
"step": 920
|
| 1381 |
+
},
|
| 1382 |
+
{
|
| 1383 |
+
"epoch": 8.6,
|
| 1384 |
+
"eval_accuracy": 0.7566137566137566,
|
| 1385 |
+
"eval_loss": 1.0412092208862305,
|
| 1386 |
+
"eval_runtime": 3.305,
|
| 1387 |
+
"eval_samples_per_second": 57.187,
|
| 1388 |
+
"eval_steps_per_second": 7.262,
|
| 1389 |
+
"step": 920
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 8.69,
|
| 1393 |
+
"learning_rate": 9.147196261682244e-05,
|
| 1394 |
+
"loss": 0.0215,
|
| 1395 |
+
"step": 930
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 8.69,
|
| 1399 |
+
"eval_accuracy": 0.7195767195767195,
|
| 1400 |
+
"eval_loss": 1.2484722137451172,
|
| 1401 |
+
"eval_runtime": 3.5183,
|
| 1402 |
+
"eval_samples_per_second": 53.719,
|
| 1403 |
+
"eval_steps_per_second": 6.821,
|
| 1404 |
+
"step": 930
|
| 1405 |
+
},
|
| 1406 |
+
{
|
| 1407 |
+
"epoch": 8.79,
|
| 1408 |
+
"learning_rate": 9.030373831775701e-05,
|
| 1409 |
+
"loss": 0.0853,
|
| 1410 |
+
"step": 940
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 8.79,
|
| 1414 |
+
"eval_accuracy": 0.7671957671957672,
|
| 1415 |
+
"eval_loss": 1.0863738059997559,
|
| 1416 |
+
"eval_runtime": 3.3344,
|
| 1417 |
+
"eval_samples_per_second": 56.681,
|
| 1418 |
+
"eval_steps_per_second": 7.198,
|
| 1419 |
+
"step": 940
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 8.88,
|
| 1423 |
+
"learning_rate": 8.913551401869159e-05,
|
| 1424 |
+
"loss": 0.0412,
|
| 1425 |
+
"step": 950
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 8.88,
|
| 1429 |
+
"eval_accuracy": 0.7248677248677249,
|
| 1430 |
+
"eval_loss": 1.179593801498413,
|
| 1431 |
+
"eval_runtime": 3.4549,
|
| 1432 |
+
"eval_samples_per_second": 54.705,
|
| 1433 |
+
"eval_steps_per_second": 6.947,
|
| 1434 |
+
"step": 950
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 8.97,
|
| 1438 |
+
"learning_rate": 8.796728971962617e-05,
|
| 1439 |
+
"loss": 0.0645,
|
| 1440 |
+
"step": 960
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 8.97,
|
| 1444 |
+
"eval_accuracy": 0.6878306878306878,
|
| 1445 |
+
"eval_loss": 1.3152216672897339,
|
| 1446 |
+
"eval_runtime": 3.4461,
|
| 1447 |
+
"eval_samples_per_second": 54.844,
|
| 1448 |
+
"eval_steps_per_second": 6.964,
|
| 1449 |
+
"step": 960
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 9.07,
|
| 1453 |
+
"learning_rate": 8.679906542056075e-05,
|
| 1454 |
+
"loss": 0.0654,
|
| 1455 |
+
"step": 970
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 9.07,
|
| 1459 |
+
"eval_accuracy": 0.6931216931216931,
|
| 1460 |
+
"eval_loss": 1.2788838148117065,
|
| 1461 |
+
"eval_runtime": 3.1953,
|
| 1462 |
+
"eval_samples_per_second": 59.15,
|
| 1463 |
+
"eval_steps_per_second": 7.511,
|
| 1464 |
+
"step": 970
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"epoch": 9.16,
|
| 1468 |
+
"learning_rate": 8.563084112149534e-05,
|
| 1469 |
+
"loss": 0.0352,
|
| 1470 |
+
"step": 980
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"epoch": 9.16,
|
| 1474 |
+
"eval_accuracy": 0.7195767195767195,
|
| 1475 |
+
"eval_loss": 1.1927844285964966,
|
| 1476 |
+
"eval_runtime": 3.4374,
|
| 1477 |
+
"eval_samples_per_second": 54.983,
|
| 1478 |
+
"eval_steps_per_second": 6.982,
|
| 1479 |
+
"step": 980
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 9.25,
|
| 1483 |
+
"learning_rate": 8.44626168224299e-05,
|
| 1484 |
+
"loss": 0.0137,
|
| 1485 |
+
"step": 990
|
| 1486 |
+
},
|
| 1487 |
+
{
|
| 1488 |
+
"epoch": 9.25,
|
| 1489 |
+
"eval_accuracy": 0.7354497354497355,
|
| 1490 |
+
"eval_loss": 1.1643190383911133,
|
| 1491 |
+
"eval_runtime": 3.1879,
|
| 1492 |
+
"eval_samples_per_second": 59.287,
|
| 1493 |
+
"eval_steps_per_second": 7.529,
|
| 1494 |
+
"step": 990
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 9.35,
|
| 1498 |
+
"learning_rate": 8.32943925233645e-05,
|
| 1499 |
+
"loss": 0.0227,
|
| 1500 |
+
"step": 1000
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 9.35,
|
| 1504 |
+
"eval_accuracy": 0.7142857142857143,
|
| 1505 |
+
"eval_loss": 1.2255942821502686,
|
| 1506 |
+
"eval_runtime": 3.4426,
|
| 1507 |
+
"eval_samples_per_second": 54.901,
|
| 1508 |
+
"eval_steps_per_second": 6.972,
|
| 1509 |
+
"step": 1000
|
| 1510 |
+
},
|
| 1511 |
+
{
|
| 1512 |
+
"epoch": 9.44,
|
| 1513 |
+
"learning_rate": 8.212616822429907e-05,
|
| 1514 |
+
"loss": 0.0391,
|
| 1515 |
+
"step": 1010
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 9.44,
|
| 1519 |
+
"eval_accuracy": 0.7195767195767195,
|
| 1520 |
+
"eval_loss": 1.2088606357574463,
|
| 1521 |
+
"eval_runtime": 3.1411,
|
| 1522 |
+
"eval_samples_per_second": 60.17,
|
| 1523 |
+
"eval_steps_per_second": 7.641,
|
| 1524 |
+
"step": 1010
|
| 1525 |
+
},
|
| 1526 |
+
{
|
| 1527 |
+
"epoch": 9.53,
|
| 1528 |
+
"learning_rate": 8.095794392523365e-05,
|
| 1529 |
+
"loss": 0.0163,
|
| 1530 |
+
"step": 1020
|
| 1531 |
+
},
|
| 1532 |
+
{
|
| 1533 |
+
"epoch": 9.53,
|
| 1534 |
+
"eval_accuracy": 0.6931216931216931,
|
| 1535 |
+
"eval_loss": 1.3879644870758057,
|
| 1536 |
+
"eval_runtime": 3.5148,
|
| 1537 |
+
"eval_samples_per_second": 53.773,
|
| 1538 |
+
"eval_steps_per_second": 6.828,
|
| 1539 |
+
"step": 1020
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 9.63,
|
| 1543 |
+
"learning_rate": 7.978971962616823e-05,
|
| 1544 |
+
"loss": 0.0225,
|
| 1545 |
+
"step": 1030
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 9.63,
|
| 1549 |
+
"eval_accuracy": 0.6931216931216931,
|
| 1550 |
+
"eval_loss": 1.3944264650344849,
|
| 1551 |
+
"eval_runtime": 3.2816,
|
| 1552 |
+
"eval_samples_per_second": 57.594,
|
| 1553 |
+
"eval_steps_per_second": 7.314,
|
| 1554 |
+
"step": 1030
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 9.72,
|
| 1558 |
+
"learning_rate": 7.862149532710281e-05,
|
| 1559 |
+
"loss": 0.0348,
|
| 1560 |
+
"step": 1040
|
| 1561 |
+
},
|
| 1562 |
+
{
|
| 1563 |
+
"epoch": 9.72,
|
| 1564 |
+
"eval_accuracy": 0.7142857142857143,
|
| 1565 |
+
"eval_loss": 1.3256537914276123,
|
| 1566 |
+
"eval_runtime": 3.2964,
|
| 1567 |
+
"eval_samples_per_second": 57.334,
|
| 1568 |
+
"eval_steps_per_second": 7.281,
|
| 1569 |
+
"step": 1040
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 9.81,
|
| 1573 |
+
"learning_rate": 7.74532710280374e-05,
|
| 1574 |
+
"loss": 0.0354,
|
| 1575 |
+
"step": 1050
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 9.81,
|
| 1579 |
+
"eval_accuracy": 0.746031746031746,
|
| 1580 |
+
"eval_loss": 1.1538487672805786,
|
| 1581 |
+
"eval_runtime": 3.3622,
|
| 1582 |
+
"eval_samples_per_second": 56.213,
|
| 1583 |
+
"eval_steps_per_second": 7.138,
|
| 1584 |
+
"step": 1050
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 9.91,
|
| 1588 |
+
"learning_rate": 7.628504672897196e-05,
|
| 1589 |
+
"loss": 0.0412,
|
| 1590 |
+
"step": 1060
|
| 1591 |
+
},
|
| 1592 |
+
{
|
| 1593 |
+
"epoch": 9.91,
|
| 1594 |
+
"eval_accuracy": 0.7248677248677249,
|
| 1595 |
+
"eval_loss": 1.237243890762329,
|
| 1596 |
+
"eval_runtime": 3.4338,
|
| 1597 |
+
"eval_samples_per_second": 55.041,
|
| 1598 |
+
"eval_steps_per_second": 6.989,
|
| 1599 |
+
"step": 1060
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 10.0,
|
| 1603 |
+
"learning_rate": 7.511682242990654e-05,
|
| 1604 |
+
"loss": 0.055,
|
| 1605 |
+
"step": 1070
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 10.0,
|
| 1609 |
+
"eval_accuracy": 0.708994708994709,
|
| 1610 |
+
"eval_loss": 1.2265738248825073,
|
| 1611 |
+
"eval_runtime": 3.4414,
|
| 1612 |
+
"eval_samples_per_second": 54.919,
|
| 1613 |
+
"eval_steps_per_second": 6.974,
|
| 1614 |
+
"step": 1070
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"epoch": 10.09,
|
| 1618 |
+
"learning_rate": 7.394859813084113e-05,
|
| 1619 |
+
"loss": 0.0115,
|
| 1620 |
+
"step": 1080
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 10.09,
|
| 1624 |
+
"eval_accuracy": 0.7248677248677249,
|
| 1625 |
+
"eval_loss": 1.235258936882019,
|
| 1626 |
+
"eval_runtime": 3.3434,
|
| 1627 |
+
"eval_samples_per_second": 56.53,
|
| 1628 |
+
"eval_steps_per_second": 7.178,
|
| 1629 |
+
"step": 1080
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 10.19,
|
| 1633 |
+
"learning_rate": 7.278037383177571e-05,
|
| 1634 |
+
"loss": 0.011,
|
| 1635 |
+
"step": 1090
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 10.19,
|
| 1639 |
+
"eval_accuracy": 0.7248677248677249,
|
| 1640 |
+
"eval_loss": 1.2655125856399536,
|
| 1641 |
+
"eval_runtime": 3.2569,
|
| 1642 |
+
"eval_samples_per_second": 58.031,
|
| 1643 |
+
"eval_steps_per_second": 7.369,
|
| 1644 |
+
"step": 1090
|
| 1645 |
+
},
|
| 1646 |
+
{
|
| 1647 |
+
"epoch": 10.28,
|
| 1648 |
+
"learning_rate": 7.161214953271029e-05,
|
| 1649 |
+
"loss": 0.0105,
|
| 1650 |
+
"step": 1100
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 10.28,
|
| 1654 |
+
"eval_accuracy": 0.7354497354497355,
|
| 1655 |
+
"eval_loss": 1.2830661535263062,
|
| 1656 |
+
"eval_runtime": 3.3769,
|
| 1657 |
+
"eval_samples_per_second": 55.968,
|
| 1658 |
+
"eval_steps_per_second": 7.107,
|
| 1659 |
+
"step": 1100
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 10.37,
|
| 1663 |
+
"learning_rate": 7.044392523364486e-05,
|
| 1664 |
+
"loss": 0.0248,
|
| 1665 |
+
"step": 1110
|
| 1666 |
+
},
|
| 1667 |
+
{
|
| 1668 |
+
"epoch": 10.37,
|
| 1669 |
+
"eval_accuracy": 0.7142857142857143,
|
| 1670 |
+
"eval_loss": 1.3138365745544434,
|
| 1671 |
+
"eval_runtime": 3.2282,
|
| 1672 |
+
"eval_samples_per_second": 58.547,
|
| 1673 |
+
"eval_steps_per_second": 7.435,
|
| 1674 |
+
"step": 1110
|
| 1675 |
+
},
|
| 1676 |
+
{
|
| 1677 |
+
"epoch": 10.47,
|
| 1678 |
+
"learning_rate": 6.927570093457945e-05,
|
| 1679 |
+
"loss": 0.0287,
|
| 1680 |
+
"step": 1120
|
| 1681 |
+
},
|
| 1682 |
+
{
|
| 1683 |
+
"epoch": 10.47,
|
| 1684 |
+
"eval_accuracy": 0.7195767195767195,
|
| 1685 |
+
"eval_loss": 1.247241735458374,
|
| 1686 |
+
"eval_runtime": 3.7473,
|
| 1687 |
+
"eval_samples_per_second": 50.437,
|
| 1688 |
+
"eval_steps_per_second": 6.405,
|
| 1689 |
+
"step": 1120
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 10.56,
|
| 1693 |
+
"learning_rate": 6.810747663551402e-05,
|
| 1694 |
+
"loss": 0.017,
|
| 1695 |
+
"step": 1130
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 10.56,
|
| 1699 |
+
"eval_accuracy": 0.7619047619047619,
|
| 1700 |
+
"eval_loss": 1.1517435312271118,
|
| 1701 |
+
"eval_runtime": 3.3479,
|
| 1702 |
+
"eval_samples_per_second": 56.453,
|
| 1703 |
+
"eval_steps_per_second": 7.169,
|
| 1704 |
+
"step": 1130
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 10.65,
|
| 1708 |
+
"learning_rate": 6.69392523364486e-05,
|
| 1709 |
+
"loss": 0.0326,
|
| 1710 |
+
"step": 1140
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 10.65,
|
| 1714 |
+
"eval_accuracy": 0.7513227513227513,
|
| 1715 |
+
"eval_loss": 1.1729180812835693,
|
| 1716 |
+
"eval_runtime": 3.563,
|
| 1717 |
+
"eval_samples_per_second": 53.045,
|
| 1718 |
+
"eval_steps_per_second": 6.736,
|
| 1719 |
+
"step": 1140
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 10.75,
|
| 1723 |
+
"learning_rate": 6.577102803738317e-05,
|
| 1724 |
+
"loss": 0.0298,
|
| 1725 |
+
"step": 1150
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 10.75,
|
| 1729 |
+
"eval_accuracy": 0.746031746031746,
|
| 1730 |
+
"eval_loss": 1.1990931034088135,
|
| 1731 |
+
"eval_runtime": 3.1999,
|
| 1732 |
+
"eval_samples_per_second": 59.065,
|
| 1733 |
+
"eval_steps_per_second": 7.5,
|
| 1734 |
+
"step": 1150
|
| 1735 |
+
},
|
| 1736 |
+
{
|
| 1737 |
+
"epoch": 10.84,
|
| 1738 |
+
"learning_rate": 6.460280373831777e-05,
|
| 1739 |
+
"loss": 0.0087,
|
| 1740 |
+
"step": 1160
|
| 1741 |
+
},
|
| 1742 |
+
{
|
| 1743 |
+
"epoch": 10.84,
|
| 1744 |
+
"eval_accuracy": 0.7195767195767195,
|
| 1745 |
+
"eval_loss": 1.1964682340621948,
|
| 1746 |
+
"eval_runtime": 3.3211,
|
| 1747 |
+
"eval_samples_per_second": 56.908,
|
| 1748 |
+
"eval_steps_per_second": 7.226,
|
| 1749 |
+
"step": 1160
|
| 1750 |
+
},
|
| 1751 |
+
{
|
| 1752 |
+
"epoch": 10.93,
|
| 1753 |
+
"learning_rate": 6.343457943925233e-05,
|
| 1754 |
+
"loss": 0.0104,
|
| 1755 |
+
"step": 1170
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 10.93,
|
| 1759 |
+
"eval_accuracy": 0.7301587301587301,
|
| 1760 |
+
"eval_loss": 1.200627088546753,
|
| 1761 |
+
"eval_runtime": 3.2863,
|
| 1762 |
+
"eval_samples_per_second": 57.511,
|
| 1763 |
+
"eval_steps_per_second": 7.303,
|
| 1764 |
+
"step": 1170
|
| 1765 |
+
},
|
| 1766 |
+
{
|
| 1767 |
+
"epoch": 11.03,
|
| 1768 |
+
"learning_rate": 6.226635514018691e-05,
|
| 1769 |
+
"loss": 0.0176,
|
| 1770 |
+
"step": 1180
|
| 1771 |
+
},
|
| 1772 |
+
{
|
| 1773 |
+
"epoch": 11.03,
|
| 1774 |
+
"eval_accuracy": 0.7195767195767195,
|
| 1775 |
+
"eval_loss": 1.2818912267684937,
|
| 1776 |
+
"eval_runtime": 3.3882,
|
| 1777 |
+
"eval_samples_per_second": 55.783,
|
| 1778 |
+
"eval_steps_per_second": 7.083,
|
| 1779 |
+
"step": 1180
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 11.12,
|
| 1783 |
+
"learning_rate": 6.109813084112151e-05,
|
| 1784 |
+
"loss": 0.0088,
|
| 1785 |
+
"step": 1190
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 11.12,
|
| 1789 |
+
"eval_accuracy": 0.7248677248677249,
|
| 1790 |
+
"eval_loss": 1.2860090732574463,
|
| 1791 |
+
"eval_runtime": 3.4546,
|
| 1792 |
+
"eval_samples_per_second": 54.709,
|
| 1793 |
+
"eval_steps_per_second": 6.947,
|
| 1794 |
+
"step": 1190
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 11.21,
|
| 1798 |
+
"learning_rate": 5.992990654205608e-05,
|
| 1799 |
+
"loss": 0.0218,
|
| 1800 |
+
"step": 1200
|
| 1801 |
+
},
|
| 1802 |
+
{
|
| 1803 |
+
"epoch": 11.21,
|
| 1804 |
+
"eval_accuracy": 0.7407407407407407,
|
| 1805 |
+
"eval_loss": 1.1996196508407593,
|
| 1806 |
+
"eval_runtime": 3.4627,
|
| 1807 |
+
"eval_samples_per_second": 54.582,
|
| 1808 |
+
"eval_steps_per_second": 6.931,
|
| 1809 |
+
"step": 1200
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 11.31,
|
| 1813 |
+
"learning_rate": 5.876168224299065e-05,
|
| 1814 |
+
"loss": 0.011,
|
| 1815 |
+
"step": 1210
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 11.31,
|
| 1819 |
+
"eval_accuracy": 0.7407407407407407,
|
| 1820 |
+
"eval_loss": 1.1904582977294922,
|
| 1821 |
+
"eval_runtime": 3.6413,
|
| 1822 |
+
"eval_samples_per_second": 51.905,
|
| 1823 |
+
"eval_steps_per_second": 6.591,
|
| 1824 |
+
"step": 1210
|
| 1825 |
+
},
|
| 1826 |
+
{
|
| 1827 |
+
"epoch": 11.4,
|
| 1828 |
+
"learning_rate": 5.7593457943925235e-05,
|
| 1829 |
+
"loss": 0.0195,
|
| 1830 |
+
"step": 1220
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 11.4,
|
| 1834 |
+
"eval_accuracy": 0.746031746031746,
|
| 1835 |
+
"eval_loss": 1.1777358055114746,
|
| 1836 |
+
"eval_runtime": 3.4353,
|
| 1837 |
+
"eval_samples_per_second": 55.016,
|
| 1838 |
+
"eval_steps_per_second": 6.986,
|
| 1839 |
+
"step": 1220
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 11.5,
|
| 1843 |
+
"learning_rate": 5.6425233644859823e-05,
|
| 1844 |
+
"loss": 0.012,
|
| 1845 |
+
"step": 1230
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 11.5,
|
| 1849 |
+
"eval_accuracy": 0.7566137566137566,
|
| 1850 |
+
"eval_loss": 1.1417244672775269,
|
| 1851 |
+
"eval_runtime": 3.4149,
|
| 1852 |
+
"eval_samples_per_second": 55.346,
|
| 1853 |
+
"eval_steps_per_second": 7.028,
|
| 1854 |
+
"step": 1230
|
| 1855 |
+
},
|
| 1856 |
+
{
|
| 1857 |
+
"epoch": 11.59,
|
| 1858 |
+
"learning_rate": 5.52570093457944e-05,
|
| 1859 |
+
"loss": 0.0075,
|
| 1860 |
+
"step": 1240
|
| 1861 |
+
},
|
| 1862 |
+
{
|
| 1863 |
+
"epoch": 11.59,
|
| 1864 |
+
"eval_accuracy": 0.7619047619047619,
|
| 1865 |
+
"eval_loss": 1.1428831815719604,
|
| 1866 |
+
"eval_runtime": 3.4904,
|
| 1867 |
+
"eval_samples_per_second": 54.148,
|
| 1868 |
+
"eval_steps_per_second": 6.876,
|
| 1869 |
+
"step": 1240
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 11.68,
|
| 1873 |
+
"learning_rate": 5.408878504672897e-05,
|
| 1874 |
+
"loss": 0.0131,
|
| 1875 |
+
"step": 1250
|
| 1876 |
+
},
|
| 1877 |
+
{
|
| 1878 |
+
"epoch": 11.68,
|
| 1879 |
+
"eval_accuracy": 0.7671957671957672,
|
| 1880 |
+
"eval_loss": 1.1381064653396606,
|
| 1881 |
+
"eval_runtime": 3.3671,
|
| 1882 |
+
"eval_samples_per_second": 56.131,
|
| 1883 |
+
"eval_steps_per_second": 7.128,
|
| 1884 |
+
"step": 1250
|
| 1885 |
+
},
|
| 1886 |
+
{
|
| 1887 |
+
"epoch": 11.78,
|
| 1888 |
+
"learning_rate": 5.292056074766355e-05,
|
| 1889 |
+
"loss": 0.0078,
|
| 1890 |
+
"step": 1260
|
| 1891 |
+
},
|
| 1892 |
+
{
|
| 1893 |
+
"epoch": 11.78,
|
| 1894 |
+
"eval_accuracy": 0.7566137566137566,
|
| 1895 |
+
"eval_loss": 1.156213641166687,
|
| 1896 |
+
"eval_runtime": 3.7773,
|
| 1897 |
+
"eval_samples_per_second": 50.036,
|
| 1898 |
+
"eval_steps_per_second": 6.354,
|
| 1899 |
+
"step": 1260
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 11.87,
|
| 1903 |
+
"learning_rate": 5.175233644859814e-05,
|
| 1904 |
+
"loss": 0.0071,
|
| 1905 |
+
"step": 1270
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 11.87,
|
| 1909 |
+
"eval_accuracy": 0.7619047619047619,
|
| 1910 |
+
"eval_loss": 1.1708426475524902,
|
| 1911 |
+
"eval_runtime": 3.4015,
|
| 1912 |
+
"eval_samples_per_second": 55.564,
|
| 1913 |
+
"eval_steps_per_second": 7.056,
|
| 1914 |
+
"step": 1270
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 11.96,
|
| 1918 |
+
"learning_rate": 5.058411214953271e-05,
|
| 1919 |
+
"loss": 0.04,
|
| 1920 |
+
"step": 1280
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 11.96,
|
| 1924 |
+
"eval_accuracy": 0.7513227513227513,
|
| 1925 |
+
"eval_loss": 1.1964821815490723,
|
| 1926 |
+
"eval_runtime": 3.6594,
|
| 1927 |
+
"eval_samples_per_second": 51.648,
|
| 1928 |
+
"eval_steps_per_second": 6.559,
|
| 1929 |
+
"step": 1280
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 12.06,
|
| 1933 |
+
"learning_rate": 4.941588785046729e-05,
|
| 1934 |
+
"loss": 0.0066,
|
| 1935 |
+
"step": 1290
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 12.06,
|
| 1939 |
+
"eval_accuracy": 0.7354497354497355,
|
| 1940 |
+
"eval_loss": 1.2294511795043945,
|
| 1941 |
+
"eval_runtime": 3.2967,
|
| 1942 |
+
"eval_samples_per_second": 57.329,
|
| 1943 |
+
"eval_steps_per_second": 7.28,
|
| 1944 |
+
"step": 1290
|
| 1945 |
+
},
|
| 1946 |
+
{
|
| 1947 |
+
"epoch": 12.15,
|
| 1948 |
+
"learning_rate": 4.8247663551401875e-05,
|
| 1949 |
+
"loss": 0.0179,
|
| 1950 |
+
"step": 1300
|
| 1951 |
+
},
|
| 1952 |
+
{
|
| 1953 |
+
"epoch": 12.15,
|
| 1954 |
+
"eval_accuracy": 0.7354497354497355,
|
| 1955 |
+
"eval_loss": 1.2337099313735962,
|
| 1956 |
+
"eval_runtime": 3.344,
|
| 1957 |
+
"eval_samples_per_second": 56.519,
|
| 1958 |
+
"eval_steps_per_second": 7.177,
|
| 1959 |
+
"step": 1300
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 12.24,
|
| 1963 |
+
"learning_rate": 4.707943925233645e-05,
|
| 1964 |
+
"loss": 0.0072,
|
| 1965 |
+
"step": 1310
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 12.24,
|
| 1969 |
+
"eval_accuracy": 0.7407407407407407,
|
| 1970 |
+
"eval_loss": 1.2375584840774536,
|
| 1971 |
+
"eval_runtime": 3.3433,
|
| 1972 |
+
"eval_samples_per_second": 56.531,
|
| 1973 |
+
"eval_steps_per_second": 7.179,
|
| 1974 |
+
"step": 1310
|
| 1975 |
+
},
|
| 1976 |
+
{
|
| 1977 |
+
"epoch": 12.34,
|
| 1978 |
+
"learning_rate": 4.591121495327103e-05,
|
| 1979 |
+
"loss": 0.0189,
|
| 1980 |
+
"step": 1320
|
| 1981 |
+
},
|
| 1982 |
+
{
|
| 1983 |
+
"epoch": 12.34,
|
| 1984 |
+
"eval_accuracy": 0.7354497354497355,
|
| 1985 |
+
"eval_loss": 1.2402395009994507,
|
| 1986 |
+
"eval_runtime": 3.4557,
|
| 1987 |
+
"eval_samples_per_second": 54.692,
|
| 1988 |
+
"eval_steps_per_second": 6.945,
|
| 1989 |
+
"step": 1320
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 12.43,
|
| 1993 |
+
"learning_rate": 4.4742990654205606e-05,
|
| 1994 |
+
"loss": 0.0067,
|
| 1995 |
+
"step": 1330
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 12.43,
|
| 1999 |
+
"eval_accuracy": 0.7407407407407407,
|
| 2000 |
+
"eval_loss": 1.2426130771636963,
|
| 2001 |
+
"eval_runtime": 3.468,
|
| 2002 |
+
"eval_samples_per_second": 54.498,
|
| 2003 |
+
"eval_steps_per_second": 6.92,
|
| 2004 |
+
"step": 1330
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 12.52,
|
| 2008 |
+
"learning_rate": 4.357476635514019e-05,
|
| 2009 |
+
"loss": 0.014,
|
| 2010 |
+
"step": 1340
|
| 2011 |
+
},
|
| 2012 |
+
{
|
| 2013 |
+
"epoch": 12.52,
|
| 2014 |
+
"eval_accuracy": 0.746031746031746,
|
| 2015 |
+
"eval_loss": 1.219921588897705,
|
| 2016 |
+
"eval_runtime": 3.4683,
|
| 2017 |
+
"eval_samples_per_second": 54.494,
|
| 2018 |
+
"eval_steps_per_second": 6.92,
|
| 2019 |
+
"step": 1340
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 12.62,
|
| 2023 |
+
"learning_rate": 4.240654205607476e-05,
|
| 2024 |
+
"loss": 0.0065,
|
| 2025 |
+
"step": 1350
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 12.62,
|
| 2029 |
+
"eval_accuracy": 0.7513227513227513,
|
| 2030 |
+
"eval_loss": 1.2070165872573853,
|
| 2031 |
+
"eval_runtime": 3.3345,
|
| 2032 |
+
"eval_samples_per_second": 56.68,
|
| 2033 |
+
"eval_steps_per_second": 7.197,
|
| 2034 |
+
"step": 1350
|
| 2035 |
+
},
|
| 2036 |
+
{
|
| 2037 |
+
"epoch": 12.71,
|
| 2038 |
+
"learning_rate": 4.123831775700935e-05,
|
| 2039 |
+
"loss": 0.0119,
|
| 2040 |
+
"step": 1360
|
| 2041 |
+
},
|
| 2042 |
+
{
|
| 2043 |
+
"epoch": 12.71,
|
| 2044 |
+
"eval_accuracy": 0.7513227513227513,
|
| 2045 |
+
"eval_loss": 1.2172141075134277,
|
| 2046 |
+
"eval_runtime": 3.335,
|
| 2047 |
+
"eval_samples_per_second": 56.672,
|
| 2048 |
+
"eval_steps_per_second": 7.196,
|
| 2049 |
+
"step": 1360
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 12.8,
|
| 2053 |
+
"learning_rate": 4.0070093457943926e-05,
|
| 2054 |
+
"loss": 0.0065,
|
| 2055 |
+
"step": 1370
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 12.8,
|
| 2059 |
+
"eval_accuracy": 0.746031746031746,
|
| 2060 |
+
"eval_loss": 1.2299124002456665,
|
| 2061 |
+
"eval_runtime": 3.4554,
|
| 2062 |
+
"eval_samples_per_second": 54.697,
|
| 2063 |
+
"eval_steps_per_second": 6.946,
|
| 2064 |
+
"step": 1370
|
| 2065 |
+
},
|
| 2066 |
+
{
|
| 2067 |
+
"epoch": 12.9,
|
| 2068 |
+
"learning_rate": 3.890186915887851e-05,
|
| 2069 |
+
"loss": 0.0139,
|
| 2070 |
+
"step": 1380
|
| 2071 |
+
},
|
| 2072 |
+
{
|
| 2073 |
+
"epoch": 12.9,
|
| 2074 |
+
"eval_accuracy": 0.7513227513227513,
|
| 2075 |
+
"eval_loss": 1.2095272541046143,
|
| 2076 |
+
"eval_runtime": 3.5215,
|
| 2077 |
+
"eval_samples_per_second": 53.67,
|
| 2078 |
+
"eval_steps_per_second": 6.815,
|
| 2079 |
+
"step": 1380
|
| 2080 |
+
},
|
| 2081 |
+
{
|
| 2082 |
+
"epoch": 12.99,
|
| 2083 |
+
"learning_rate": 3.773364485981308e-05,
|
| 2084 |
+
"loss": 0.0195,
|
| 2085 |
+
"step": 1390
|
| 2086 |
+
},
|
| 2087 |
+
{
|
| 2088 |
+
"epoch": 12.99,
|
| 2089 |
+
"eval_accuracy": 0.7513227513227513,
|
| 2090 |
+
"eval_loss": 1.1914021968841553,
|
| 2091 |
+
"eval_runtime": 3.3578,
|
| 2092 |
+
"eval_samples_per_second": 56.287,
|
| 2093 |
+
"eval_steps_per_second": 7.148,
|
| 2094 |
+
"step": 1390
|
| 2095 |
+
},
|
| 2096 |
+
{
|
| 2097 |
+
"epoch": 13.08,
|
| 2098 |
+
"learning_rate": 3.6565420560747665e-05,
|
| 2099 |
+
"loss": 0.0102,
|
| 2100 |
+
"step": 1400
|
| 2101 |
+
},
|
| 2102 |
+
{
|
| 2103 |
+
"epoch": 13.08,
|
| 2104 |
+
"eval_accuracy": 0.7513227513227513,
|
| 2105 |
+
"eval_loss": 1.1972215175628662,
|
| 2106 |
+
"eval_runtime": 3.4283,
|
| 2107 |
+
"eval_samples_per_second": 55.13,
|
| 2108 |
+
"eval_steps_per_second": 7.001,
|
| 2109 |
+
"step": 1400
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 13.18,
|
| 2113 |
+
"learning_rate": 3.5397196261682246e-05,
|
| 2114 |
+
"loss": 0.0162,
|
| 2115 |
+
"step": 1410
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 13.18,
|
| 2119 |
+
"eval_accuracy": 0.7566137566137566,
|
| 2120 |
+
"eval_loss": 1.2006089687347412,
|
| 2121 |
+
"eval_runtime": 3.2014,
|
| 2122 |
+
"eval_samples_per_second": 59.036,
|
| 2123 |
+
"eval_steps_per_second": 7.497,
|
| 2124 |
+
"step": 1410
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 13.27,
|
| 2128 |
+
"learning_rate": 3.422897196261682e-05,
|
| 2129 |
+
"loss": 0.0057,
|
| 2130 |
+
"step": 1420
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 13.27,
|
| 2134 |
+
"eval_accuracy": 0.7566137566137566,
|
| 2135 |
+
"eval_loss": 1.213531494140625,
|
| 2136 |
+
"eval_runtime": 3.2962,
|
| 2137 |
+
"eval_samples_per_second": 57.34,
|
| 2138 |
+
"eval_steps_per_second": 7.281,
|
| 2139 |
+
"step": 1420
|
| 2140 |
+
},
|
| 2141 |
+
{
|
| 2142 |
+
"epoch": 13.36,
|
| 2143 |
+
"learning_rate": 3.30607476635514e-05,
|
| 2144 |
+
"loss": 0.0099,
|
| 2145 |
+
"step": 1430
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 13.36,
|
| 2149 |
+
"eval_accuracy": 0.7566137566137566,
|
| 2150 |
+
"eval_loss": 1.2060108184814453,
|
| 2151 |
+
"eval_runtime": 3.4059,
|
| 2152 |
+
"eval_samples_per_second": 55.493,
|
| 2153 |
+
"eval_steps_per_second": 7.047,
|
| 2154 |
+
"step": 1430
|
| 2155 |
+
},
|
| 2156 |
+
{
|
| 2157 |
+
"epoch": 13.46,
|
| 2158 |
+
"learning_rate": 3.1892523364485985e-05,
|
| 2159 |
+
"loss": 0.0092,
|
| 2160 |
+
"step": 1440
|
| 2161 |
+
},
|
| 2162 |
+
{
|
| 2163 |
+
"epoch": 13.46,
|
| 2164 |
+
"eval_accuracy": 0.7513227513227513,
|
| 2165 |
+
"eval_loss": 1.209419846534729,
|
| 2166 |
+
"eval_runtime": 3.3619,
|
| 2167 |
+
"eval_samples_per_second": 56.218,
|
| 2168 |
+
"eval_steps_per_second": 7.139,
|
| 2169 |
+
"step": 1440
|
| 2170 |
+
},
|
| 2171 |
+
{
|
| 2172 |
+
"epoch": 13.55,
|
| 2173 |
+
"learning_rate": 3.0724299065420566e-05,
|
| 2174 |
+
"loss": 0.0059,
|
| 2175 |
+
"step": 1450
|
| 2176 |
+
},
|
| 2177 |
+
{
|
| 2178 |
+
"epoch": 13.55,
|
| 2179 |
+
"eval_accuracy": 0.746031746031746,
|
| 2180 |
+
"eval_loss": 1.21532142162323,
|
| 2181 |
+
"eval_runtime": 3.4612,
|
| 2182 |
+
"eval_samples_per_second": 54.605,
|
| 2183 |
+
"eval_steps_per_second": 6.934,
|
| 2184 |
+
"step": 1450
|
| 2185 |
+
},
|
| 2186 |
+
{
|
| 2187 |
+
"epoch": 13.64,
|
| 2188 |
+
"learning_rate": 2.955607476635514e-05,
|
| 2189 |
+
"loss": 0.0132,
|
| 2190 |
+
"step": 1460
|
| 2191 |
+
},
|
| 2192 |
+
{
|
| 2193 |
+
"epoch": 13.64,
|
| 2194 |
+
"eval_accuracy": 0.7513227513227513,
|
| 2195 |
+
"eval_loss": 1.2271349430084229,
|
| 2196 |
+
"eval_runtime": 3.3079,
|
| 2197 |
+
"eval_samples_per_second": 57.136,
|
| 2198 |
+
"eval_steps_per_second": 7.255,
|
| 2199 |
+
"step": 1460
|
| 2200 |
+
},
|
| 2201 |
+
{
|
| 2202 |
+
"epoch": 13.74,
|
| 2203 |
+
"learning_rate": 2.8387850467289723e-05,
|
| 2204 |
+
"loss": 0.0224,
|
| 2205 |
+
"step": 1470
|
| 2206 |
+
},
|
| 2207 |
+
{
|
| 2208 |
+
"epoch": 13.74,
|
| 2209 |
+
"eval_accuracy": 0.746031746031746,
|
| 2210 |
+
"eval_loss": 1.2394039630889893,
|
| 2211 |
+
"eval_runtime": 3.2507,
|
| 2212 |
+
"eval_samples_per_second": 58.141,
|
| 2213 |
+
"eval_steps_per_second": 7.383,
|
| 2214 |
+
"step": 1470
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 13.83,
|
| 2218 |
+
"learning_rate": 2.7219626168224298e-05,
|
| 2219 |
+
"loss": 0.0116,
|
| 2220 |
+
"step": 1480
|
| 2221 |
+
},
|
| 2222 |
+
{
|
| 2223 |
+
"epoch": 13.83,
|
| 2224 |
+
"eval_accuracy": 0.746031746031746,
|
| 2225 |
+
"eval_loss": 1.2353821992874146,
|
| 2226 |
+
"eval_runtime": 3.2965,
|
| 2227 |
+
"eval_samples_per_second": 57.333,
|
| 2228 |
+
"eval_steps_per_second": 7.28,
|
| 2229 |
+
"step": 1480
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 13.93,
|
| 2233 |
+
"learning_rate": 2.605140186915888e-05,
|
| 2234 |
+
"loss": 0.0096,
|
| 2235 |
+
"step": 1490
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 13.93,
|
| 2239 |
+
"eval_accuracy": 0.746031746031746,
|
| 2240 |
+
"eval_loss": 1.2316288948059082,
|
| 2241 |
+
"eval_runtime": 3.263,
|
| 2242 |
+
"eval_samples_per_second": 57.923,
|
| 2243 |
+
"eval_steps_per_second": 7.355,
|
| 2244 |
+
"step": 1490
|
| 2245 |
+
},
|
| 2246 |
+
{
|
| 2247 |
+
"epoch": 14.02,
|
| 2248 |
+
"learning_rate": 2.488317757009346e-05,
|
| 2249 |
+
"loss": 0.0055,
|
| 2250 |
+
"step": 1500
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"epoch": 14.02,
|
| 2254 |
+
"eval_accuracy": 0.746031746031746,
|
| 2255 |
+
"eval_loss": 1.2332398891448975,
|
| 2256 |
+
"eval_runtime": 3.4255,
|
| 2257 |
+
"eval_samples_per_second": 55.175,
|
| 2258 |
+
"eval_steps_per_second": 7.006,
|
| 2259 |
+
"step": 1500
|
| 2260 |
+
},
|
| 2261 |
+
{
|
| 2262 |
+
"epoch": 14.11,
|
| 2263 |
+
"learning_rate": 2.371495327102804e-05,
|
| 2264 |
+
"loss": 0.009,
|
| 2265 |
+
"step": 1510
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 14.11,
|
| 2269 |
+
"eval_accuracy": 0.746031746031746,
|
| 2270 |
+
"eval_loss": 1.2354868650436401,
|
| 2271 |
+
"eval_runtime": 3.3723,
|
| 2272 |
+
"eval_samples_per_second": 56.045,
|
| 2273 |
+
"eval_steps_per_second": 7.117,
|
| 2274 |
+
"step": 1510
|
| 2275 |
+
},
|
| 2276 |
+
{
|
| 2277 |
+
"epoch": 14.21,
|
| 2278 |
+
"learning_rate": 2.2546728971962618e-05,
|
| 2279 |
+
"loss": 0.0058,
|
| 2280 |
+
"step": 1520
|
| 2281 |
+
},
|
| 2282 |
+
{
|
| 2283 |
+
"epoch": 14.21,
|
| 2284 |
+
"eval_accuracy": 0.746031746031746,
|
| 2285 |
+
"eval_loss": 1.2446562051773071,
|
| 2286 |
+
"eval_runtime": 3.5668,
|
| 2287 |
+
"eval_samples_per_second": 52.988,
|
| 2288 |
+
"eval_steps_per_second": 6.729,
|
| 2289 |
+
"step": 1520
|
| 2290 |
+
},
|
| 2291 |
+
{
|
| 2292 |
+
"epoch": 14.3,
|
| 2293 |
+
"learning_rate": 2.13785046728972e-05,
|
| 2294 |
+
"loss": 0.01,
|
| 2295 |
+
"step": 1530
|
| 2296 |
+
},
|
| 2297 |
+
{
|
| 2298 |
+
"epoch": 14.3,
|
| 2299 |
+
"eval_accuracy": 0.746031746031746,
|
| 2300 |
+
"eval_loss": 1.2436867952346802,
|
| 2301 |
+
"eval_runtime": 3.2984,
|
| 2302 |
+
"eval_samples_per_second": 57.3,
|
| 2303 |
+
"eval_steps_per_second": 7.276,
|
| 2304 |
+
"step": 1530
|
| 2305 |
+
},
|
| 2306 |
+
{
|
| 2307 |
+
"epoch": 14.39,
|
| 2308 |
+
"learning_rate": 2.0210280373831778e-05,
|
| 2309 |
+
"loss": 0.0055,
|
| 2310 |
+
"step": 1540
|
| 2311 |
+
},
|
| 2312 |
+
{
|
| 2313 |
+
"epoch": 14.39,
|
| 2314 |
+
"eval_accuracy": 0.746031746031746,
|
| 2315 |
+
"eval_loss": 1.2422124147415161,
|
| 2316 |
+
"eval_runtime": 3.3462,
|
| 2317 |
+
"eval_samples_per_second": 56.481,
|
| 2318 |
+
"eval_steps_per_second": 7.172,
|
| 2319 |
+
"step": 1540
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 14.49,
|
| 2323 |
+
"learning_rate": 1.9042056074766356e-05,
|
| 2324 |
+
"loss": 0.0187,
|
| 2325 |
+
"step": 1550
|
| 2326 |
+
},
|
| 2327 |
+
{
|
| 2328 |
+
"epoch": 14.49,
|
| 2329 |
+
"eval_accuracy": 0.7513227513227513,
|
| 2330 |
+
"eval_loss": 1.221534252166748,
|
| 2331 |
+
"eval_runtime": 3.4706,
|
| 2332 |
+
"eval_samples_per_second": 54.457,
|
| 2333 |
+
"eval_steps_per_second": 6.915,
|
| 2334 |
+
"step": 1550
|
| 2335 |
+
},
|
| 2336 |
+
{
|
| 2337 |
+
"epoch": 14.58,
|
| 2338 |
+
"learning_rate": 1.7873831775700935e-05,
|
| 2339 |
+
"loss": 0.0103,
|
| 2340 |
+
"step": 1560
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 14.58,
|
| 2344 |
+
"eval_accuracy": 0.7513227513227513,
|
| 2345 |
+
"eval_loss": 1.217819333076477,
|
| 2346 |
+
"eval_runtime": 3.2699,
|
| 2347 |
+
"eval_samples_per_second": 57.799,
|
| 2348 |
+
"eval_steps_per_second": 7.34,
|
| 2349 |
+
"step": 1560
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 14.67,
|
| 2353 |
+
"learning_rate": 1.6705607476635516e-05,
|
| 2354 |
+
"loss": 0.0053,
|
| 2355 |
+
"step": 1570
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 14.67,
|
| 2359 |
+
"eval_accuracy": 0.746031746031746,
|
| 2360 |
+
"eval_loss": 1.2216895818710327,
|
| 2361 |
+
"eval_runtime": 3.2903,
|
| 2362 |
+
"eval_samples_per_second": 57.442,
|
| 2363 |
+
"eval_steps_per_second": 7.294,
|
| 2364 |
+
"step": 1570
|
| 2365 |
+
},
|
| 2366 |
+
{
|
| 2367 |
+
"epoch": 14.77,
|
| 2368 |
+
"learning_rate": 1.5537383177570095e-05,
|
| 2369 |
+
"loss": 0.01,
|
| 2370 |
+
"step": 1580
|
| 2371 |
+
},
|
| 2372 |
+
{
|
| 2373 |
+
"epoch": 14.77,
|
| 2374 |
+
"eval_accuracy": 0.746031746031746,
|
| 2375 |
+
"eval_loss": 1.2266777753829956,
|
| 2376 |
+
"eval_runtime": 3.2812,
|
| 2377 |
+
"eval_samples_per_second": 57.602,
|
| 2378 |
+
"eval_steps_per_second": 7.315,
|
| 2379 |
+
"step": 1580
|
| 2380 |
+
},
|
| 2381 |
+
{
|
| 2382 |
+
"epoch": 14.86,
|
| 2383 |
+
"learning_rate": 1.4369158878504675e-05,
|
| 2384 |
+
"loss": 0.0238,
|
| 2385 |
+
"step": 1590
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"epoch": 14.86,
|
| 2389 |
+
"eval_accuracy": 0.746031746031746,
|
| 2390 |
+
"eval_loss": 1.227875828742981,
|
| 2391 |
+
"eval_runtime": 3.3613,
|
| 2392 |
+
"eval_samples_per_second": 56.228,
|
| 2393 |
+
"eval_steps_per_second": 7.14,
|
| 2394 |
+
"step": 1590
|
| 2395 |
+
},
|
| 2396 |
+
{
|
| 2397 |
+
"epoch": 14.95,
|
| 2398 |
+
"learning_rate": 1.3200934579439253e-05,
|
| 2399 |
+
"loss": 0.0091,
|
| 2400 |
+
"step": 1600
|
| 2401 |
+
},
|
| 2402 |
+
{
|
| 2403 |
+
"epoch": 14.95,
|
| 2404 |
+
"eval_accuracy": 0.746031746031746,
|
| 2405 |
+
"eval_loss": 1.224220871925354,
|
| 2406 |
+
"eval_runtime": 3.2469,
|
| 2407 |
+
"eval_samples_per_second": 58.209,
|
| 2408 |
+
"eval_steps_per_second": 7.392,
|
| 2409 |
+
"step": 1600
|
| 2410 |
+
},
|
| 2411 |
+
{
|
| 2412 |
+
"epoch": 15.05,
|
| 2413 |
+
"learning_rate": 1.2032710280373833e-05,
|
| 2414 |
+
"loss": 0.0053,
|
| 2415 |
+
"step": 1610
|
| 2416 |
+
},
|
| 2417 |
+
{
|
| 2418 |
+
"epoch": 15.05,
|
| 2419 |
+
"eval_accuracy": 0.7513227513227513,
|
| 2420 |
+
"eval_loss": 1.2231720685958862,
|
| 2421 |
+
"eval_runtime": 3.1879,
|
| 2422 |
+
"eval_samples_per_second": 59.287,
|
| 2423 |
+
"eval_steps_per_second": 7.528,
|
| 2424 |
+
"step": 1610
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 15.14,
|
| 2428 |
+
"learning_rate": 1.0864485981308411e-05,
|
| 2429 |
+
"loss": 0.0101,
|
| 2430 |
+
"step": 1620
|
| 2431 |
+
},
|
| 2432 |
+
{
|
| 2433 |
+
"epoch": 15.14,
|
| 2434 |
+
"eval_accuracy": 0.746031746031746,
|
| 2435 |
+
"eval_loss": 1.2256537675857544,
|
| 2436 |
+
"eval_runtime": 3.2787,
|
| 2437 |
+
"eval_samples_per_second": 57.646,
|
| 2438 |
+
"eval_steps_per_second": 7.32,
|
| 2439 |
+
"step": 1620
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 15.23,
|
| 2443 |
+
"learning_rate": 9.696261682242991e-06,
|
| 2444 |
+
"loss": 0.0189,
|
| 2445 |
+
"step": 1630
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 15.23,
|
| 2449 |
+
"eval_accuracy": 0.746031746031746,
|
| 2450 |
+
"eval_loss": 1.227663278579712,
|
| 2451 |
+
"eval_runtime": 3.188,
|
| 2452 |
+
"eval_samples_per_second": 59.286,
|
| 2453 |
+
"eval_steps_per_second": 7.528,
|
| 2454 |
+
"step": 1630
|
| 2455 |
+
},
|
| 2456 |
+
{
|
| 2457 |
+
"epoch": 15.33,
|
| 2458 |
+
"learning_rate": 8.52803738317757e-06,
|
| 2459 |
+
"loss": 0.0056,
|
| 2460 |
+
"step": 1640
|
| 2461 |
+
},
|
| 2462 |
+
{
|
| 2463 |
+
"epoch": 15.33,
|
| 2464 |
+
"eval_accuracy": 0.746031746031746,
|
| 2465 |
+
"eval_loss": 1.233586072921753,
|
| 2466 |
+
"eval_runtime": 3.2945,
|
| 2467 |
+
"eval_samples_per_second": 57.368,
|
| 2468 |
+
"eval_steps_per_second": 7.285,
|
| 2469 |
+
"step": 1640
|
| 2470 |
+
},
|
| 2471 |
+
{
|
| 2472 |
+
"epoch": 15.42,
|
| 2473 |
+
"learning_rate": 7.3598130841121496e-06,
|
| 2474 |
+
"loss": 0.0052,
|
| 2475 |
+
"step": 1650
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"epoch": 15.42,
|
| 2479 |
+
"eval_accuracy": 0.746031746031746,
|
| 2480 |
+
"eval_loss": 1.235275149345398,
|
| 2481 |
+
"eval_runtime": 3.3479,
|
| 2482 |
+
"eval_samples_per_second": 56.453,
|
| 2483 |
+
"eval_steps_per_second": 7.169,
|
| 2484 |
+
"step": 1650
|
| 2485 |
+
},
|
| 2486 |
+
{
|
| 2487 |
+
"epoch": 15.51,
|
| 2488 |
+
"learning_rate": 6.1915887850467296e-06,
|
| 2489 |
+
"loss": 0.0054,
|
| 2490 |
+
"step": 1660
|
| 2491 |
+
},
|
| 2492 |
+
{
|
| 2493 |
+
"epoch": 15.51,
|
| 2494 |
+
"eval_accuracy": 0.746031746031746,
|
| 2495 |
+
"eval_loss": 1.2359087467193604,
|
| 2496 |
+
"eval_runtime": 3.3879,
|
| 2497 |
+
"eval_samples_per_second": 55.786,
|
| 2498 |
+
"eval_steps_per_second": 7.084,
|
| 2499 |
+
"step": 1660
|
| 2500 |
+
},
|
| 2501 |
+
{
|
| 2502 |
+
"epoch": 15.61,
|
| 2503 |
+
"learning_rate": 5.023364485981309e-06,
|
| 2504 |
+
"loss": 0.0054,
|
| 2505 |
+
"step": 1670
|
| 2506 |
+
},
|
| 2507 |
+
{
|
| 2508 |
+
"epoch": 15.61,
|
| 2509 |
+
"eval_accuracy": 0.746031746031746,
|
| 2510 |
+
"eval_loss": 1.2361844778060913,
|
| 2511 |
+
"eval_runtime": 3.1737,
|
| 2512 |
+
"eval_samples_per_second": 59.551,
|
| 2513 |
+
"eval_steps_per_second": 7.562,
|
| 2514 |
+
"step": 1670
|
| 2515 |
+
},
|
| 2516 |
+
{
|
| 2517 |
+
"epoch": 15.7,
|
| 2518 |
+
"learning_rate": 3.855140186915888e-06,
|
| 2519 |
+
"loss": 0.0102,
|
| 2520 |
+
"step": 1680
|
| 2521 |
+
},
|
| 2522 |
+
{
|
| 2523 |
+
"epoch": 15.7,
|
| 2524 |
+
"eval_accuracy": 0.7513227513227513,
|
| 2525 |
+
"eval_loss": 1.2348004579544067,
|
| 2526 |
+
"eval_runtime": 3.2806,
|
| 2527 |
+
"eval_samples_per_second": 57.612,
|
| 2528 |
+
"eval_steps_per_second": 7.316,
|
| 2529 |
+
"step": 1680
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 15.79,
|
| 2533 |
+
"learning_rate": 2.6869158878504674e-06,
|
| 2534 |
+
"loss": 0.0193,
|
| 2535 |
+
"step": 1690
|
| 2536 |
+
},
|
| 2537 |
+
{
|
| 2538 |
+
"epoch": 15.79,
|
| 2539 |
+
"eval_accuracy": 0.7513227513227513,
|
| 2540 |
+
"eval_loss": 1.2326399087905884,
|
| 2541 |
+
"eval_runtime": 3.2413,
|
| 2542 |
+
"eval_samples_per_second": 58.309,
|
| 2543 |
+
"eval_steps_per_second": 7.404,
|
| 2544 |
+
"step": 1690
|
| 2545 |
+
},
|
| 2546 |
+
{
|
| 2547 |
+
"epoch": 15.89,
|
| 2548 |
+
"learning_rate": 1.5186915887850468e-06,
|
| 2549 |
+
"loss": 0.0104,
|
| 2550 |
+
"step": 1700
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 15.89,
|
| 2554 |
+
"eval_accuracy": 0.7513227513227513,
|
| 2555 |
+
"eval_loss": 1.2315497398376465,
|
| 2556 |
+
"eval_runtime": 3.3871,
|
| 2557 |
+
"eval_samples_per_second": 55.8,
|
| 2558 |
+
"eval_steps_per_second": 7.086,
|
| 2559 |
+
"step": 1700
|
| 2560 |
+
},
|
| 2561 |
+
{
|
| 2562 |
+
"epoch": 15.98,
|
| 2563 |
+
"learning_rate": 3.5046728971962617e-07,
|
| 2564 |
+
"loss": 0.0095,
|
| 2565 |
+
"step": 1710
|
| 2566 |
+
},
|
| 2567 |
+
{
|
| 2568 |
+
"epoch": 15.98,
|
| 2569 |
+
"eval_accuracy": 0.7513227513227513,
|
| 2570 |
+
"eval_loss": 1.2312463521957397,
|
| 2571 |
+
"eval_runtime": 3.1744,
|
| 2572 |
+
"eval_samples_per_second": 59.54,
|
| 2573 |
+
"eval_steps_per_second": 7.561,
|
| 2574 |
+
"step": 1710
|
| 2575 |
+
},
|
| 2576 |
+
{
|
| 2577 |
+
"epoch": 16.0,
|
| 2578 |
+
"step": 1712,
|
| 2579 |
+
"total_flos": 2.1042510824822538e+18,
|
| 2580 |
+
"train_loss": 0.32865437336057146,
|
| 2581 |
+
"train_runtime": 1354.021,
|
| 2582 |
+
"train_samples_per_second": 20.053,
|
| 2583 |
+
"train_steps_per_second": 1.264
|
| 2584 |
+
}
|
| 2585 |
+
],
|
| 2586 |
+
"logging_steps": 10,
|
| 2587 |
+
"max_steps": 1712,
|
| 2588 |
+
"num_train_epochs": 16,
|
| 2589 |
+
"save_steps": 10,
|
| 2590 |
+
"total_flos": 2.1042510824822538e+18,
|
| 2591 |
+
"trial_name": null,
|
| 2592 |
+
"trial_params": null
|
| 2593 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a2afdd3dcd8d497caf494075688f15e2f0c7b9c69f932134c70b4930b8907ec1
|
| 3 |
+
size 4472
|