Model save
Browse files- README.md +66 -36
- all_results.json +8 -0
- logs/events.out.tfevents.1743190651.4955dc82343d.1607.0 +3 -0
- train_results.json +8 -0
- trainer_state.json +522 -0
- training_args.bin +1 -1
- training_metrics.xlsx +0 -0
README.md
CHANGED
|
@@ -18,8 +18,8 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 18 |
|
| 19 |
This model is a fine-tuned version of [WinKawaks/vit-tiny-patch16-224](https://huggingface.co/WinKawaks/vit-tiny-patch16-224) on an unknown dataset.
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
-
- Loss: 0.
|
| 22 |
-
- Accuracy: 0.
|
| 23 |
|
| 24 |
## Model description
|
| 25 |
|
|
@@ -45,43 +45,73 @@ The following hyperparameters were used during training:
|
|
| 45 |
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
|
| 46 |
- lr_scheduler_type: cosine_with_restarts
|
| 47 |
- lr_scheduler_warmup_steps: 256
|
| 48 |
-
- num_epochs:
|
| 49 |
- mixed_precision_training: Native AMP
|
| 50 |
|
| 51 |
### Training results
|
| 52 |
|
| 53 |
-
| Training Loss | Epoch | Step | Validation Loss |
|
| 54 |
-
|:-------------:|:-----:|:----:|:---------------
|
| 55 |
-
| 2.0564 | 0.5 | 64 |
|
| 56 |
-
| 1.0767 | 1.0 | 128 | 0.
|
| 57 |
-
| 0.4917 | 1.5 | 192 | 0.
|
| 58 |
-
| 0.285 | 2.0 | 256 | 0.
|
| 59 |
-
| 0.0902 | 2.5 | 320 | 0.
|
| 60 |
-
| 0.0588 | 3.0 | 384 | 0.
|
| 61 |
-
| 0.0155 | 3.5 | 448 | 0.
|
| 62 |
-
| 0.0066 | 4.0 | 512 | 0.
|
| 63 |
-
| 0.0017 | 4.5 | 576 | 0.
|
| 64 |
-
| 0.0009 | 5.0 | 640 | 0.
|
| 65 |
-
| 0.0006 | 5.5 | 704 | 0.
|
| 66 |
-
| 0.0005 | 6.0 | 768 | 0.
|
| 67 |
-
| 0.0005 | 6.5 | 832 | 0.
|
| 68 |
-
| 0.0004 | 7.0 | 896 | 0.
|
| 69 |
-
| 0.0003 | 7.5 | 960 | 0.
|
| 70 |
-
| 0.0002 | 8.0 | 1024 | 0.
|
| 71 |
-
| 0.0002 | 8.5 | 1088 | 0.
|
| 72 |
-
| 0.0002 | 9.0 | 1152 | 0.
|
| 73 |
-
| 0.0001 | 9.5 | 1216 | 0.
|
| 74 |
-
| 0.0001 | 10.0 | 1280 | 0.
|
| 75 |
-
| 0.0001 | 10.5 | 1344 | 0.
|
| 76 |
-
| 0.0001 | 11.0 | 1408 | 0.
|
| 77 |
-
| 0.0001 | 11.5 | 1472 | 0.
|
| 78 |
-
| 0.0001 | 12.0 | 1536 | 0.
|
| 79 |
-
| 0.0001 | 12.5 | 1600 | 0.
|
| 80 |
-
| 0.0001 | 13.0 | 1664 | 0.
|
| 81 |
-
| 0.0001 | 13.5 | 1728 | 0.
|
| 82 |
-
| 0.0001 | 14.0 | 1792 | 0.
|
| 83 |
-
| 0.0001 | 14.5 | 1856 | 0.
|
| 84 |
-
| 0.0001 | 15.0 | 1920 | 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 85 |
|
| 86 |
|
| 87 |
### Framework versions
|
|
@@ -89,4 +119,4 @@ The following hyperparameters were used during training:
|
|
| 89 |
- Transformers 4.48.3
|
| 90 |
- Pytorch 2.5.1+cu124
|
| 91 |
- Datasets 3.3.2
|
| 92 |
-
- Tokenizers 0.21.
|
|
|
|
| 18 |
|
| 19 |
This model is a fine-tuned version of [WinKawaks/vit-tiny-patch16-224](https://huggingface.co/WinKawaks/vit-tiny-patch16-224) on an unknown dataset.
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 0.3674
|
| 22 |
+
- Accuracy: 0.9262
|
| 23 |
|
| 24 |
## Model description
|
| 25 |
|
|
|
|
| 45 |
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
|
| 46 |
- lr_scheduler_type: cosine_with_restarts
|
| 47 |
- lr_scheduler_warmup_steps: 256
|
| 48 |
+
- num_epochs: 30
|
| 49 |
- mixed_precision_training: Native AMP
|
| 50 |
|
| 51 |
### Training results
|
| 52 |
|
| 53 |
+
| Training Loss | Epoch | Step | Accuracy | Validation Loss |
|
| 54 |
+
|:-------------:|:-----:|:----:|:--------:|:---------------:|
|
| 55 |
+
| 2.0564 | 0.5 | 64 | 0.4899 | 1.4541 |
|
| 56 |
+
| 1.0767 | 1.0 | 128 | 0.7651 | 0.6909 |
|
| 57 |
+
| 0.4917 | 1.5 | 192 | 0.8322 | 0.4307 |
|
| 58 |
+
| 0.285 | 2.0 | 256 | 0.9027 | 0.2932 |
|
| 59 |
+
| 0.0902 | 2.5 | 320 | 0.8993 | 0.3134 |
|
| 60 |
+
| 0.0588 | 3.0 | 384 | 0.9161 | 0.3076 |
|
| 61 |
+
| 0.0155 | 3.5 | 448 | 0.9396 | 0.2627 |
|
| 62 |
+
| 0.0066 | 4.0 | 512 | 0.9295 | 0.2992 |
|
| 63 |
+
| 0.0017 | 4.5 | 576 | 0.9228 | 0.2936 |
|
| 64 |
+
| 0.0009 | 5.0 | 640 | 0.9228 | 0.2961 |
|
| 65 |
+
| 0.0006 | 5.5 | 704 | 0.9228 | 0.3005 |
|
| 66 |
+
| 0.0005 | 6.0 | 768 | 0.9228 | 0.3004 |
|
| 67 |
+
| 0.0005 | 6.5 | 832 | 0.9262 | 0.2867 |
|
| 68 |
+
| 0.0004 | 7.0 | 896 | 0.9295 | 0.2977 |
|
| 69 |
+
| 0.0003 | 7.5 | 960 | 0.9295 | 0.2944 |
|
| 70 |
+
| 0.0002 | 8.0 | 1024 | 0.9295 | 0.3074 |
|
| 71 |
+
| 0.0002 | 8.5 | 1088 | 0.9329 | 0.3053 |
|
| 72 |
+
| 0.0002 | 9.0 | 1152 | 0.9295 | 0.3098 |
|
| 73 |
+
| 0.0001 | 9.5 | 1216 | 0.9295 | 0.3102 |
|
| 74 |
+
| 0.0001 | 10.0 | 1280 | 0.9262 | 0.3105 |
|
| 75 |
+
| 0.0001 | 10.5 | 1344 | 0.9262 | 0.3105 |
|
| 76 |
+
| 0.0001 | 11.0 | 1408 | 0.9262 | 0.3202 |
|
| 77 |
+
| 0.0001 | 11.5 | 1472 | 0.9295 | 0.3183 |
|
| 78 |
+
| 0.0001 | 12.0 | 1536 | 0.9329 | 0.3131 |
|
| 79 |
+
| 0.0001 | 12.5 | 1600 | 0.9295 | 0.3157 |
|
| 80 |
+
| 0.0001 | 13.0 | 1664 | 0.9228 | 0.3238 |
|
| 81 |
+
| 0.0001 | 13.5 | 1728 | 0.9228 | 0.3220 |
|
| 82 |
+
| 0.0001 | 14.0 | 1792 | 0.9228 | 0.3266 |
|
| 83 |
+
| 0.0001 | 14.5 | 1856 | 0.9228 | 0.3274 |
|
| 84 |
+
| 0.0001 | 15.0 | 1920 | 0.9228 | 0.3269 |
|
| 85 |
+
| 0.0001 | 15.5 | 1984 | 0.3267 | 0.9262 |
|
| 86 |
+
| 0.0001 | 16.0 | 2048 | 0.3298 | 0.9228 |
|
| 87 |
+
| 0.0001 | 16.5 | 2112 | 0.3330 | 0.9228 |
|
| 88 |
+
| 0.0001 | 17.0 | 2176 | 0.3337 | 0.9228 |
|
| 89 |
+
| 0.0001 | 17.5 | 2240 | 0.3337 | 0.9228 |
|
| 90 |
+
| 0.0001 | 18.0 | 2304 | 0.3355 | 0.9228 |
|
| 91 |
+
| 0.0 | 18.5 | 2368 | 0.3346 | 0.9228 |
|
| 92 |
+
| 0.0 | 19.0 | 2432 | 0.3360 | 0.9228 |
|
| 93 |
+
| 0.0 | 19.5 | 2496 | 0.3368 | 0.9228 |
|
| 94 |
+
| 0.0 | 20.0 | 2560 | 0.3365 | 0.9228 |
|
| 95 |
+
| 0.0 | 20.5 | 2624 | 0.3364 | 0.9228 |
|
| 96 |
+
| 0.0 | 21.0 | 2688 | 0.3412 | 0.9228 |
|
| 97 |
+
| 0.0 | 21.5 | 2752 | 0.3414 | 0.9228 |
|
| 98 |
+
| 0.0 | 22.0 | 2816 | 0.3435 | 0.9262 |
|
| 99 |
+
| 0.0 | 22.5 | 2880 | 0.3557 | 0.9228 |
|
| 100 |
+
| 0.0 | 23.0 | 2944 | 0.3490 | 0.9295 |
|
| 101 |
+
| 0.0 | 23.5 | 3008 | 0.3564 | 0.9262 |
|
| 102 |
+
| 0.0 | 24.0 | 3072 | 0.3545 | 0.9295 |
|
| 103 |
+
| 0.0 | 24.5 | 3136 | 0.3577 | 0.9262 |
|
| 104 |
+
| 0.0 | 25.0 | 3200 | 0.3597 | 0.9262 |
|
| 105 |
+
| 0.0 | 25.5 | 3264 | 0.3632 | 0.9262 |
|
| 106 |
+
| 0.0 | 26.0 | 3328 | 0.3627 | 0.9262 |
|
| 107 |
+
| 0.0 | 26.5 | 3392 | 0.3650 | 0.9262 |
|
| 108 |
+
| 0.0 | 27.0 | 3456 | 0.3664 | 0.9262 |
|
| 109 |
+
| 0.0 | 27.5 | 3520 | 0.3664 | 0.9262 |
|
| 110 |
+
| 0.0 | 28.0 | 3584 | 0.3666 | 0.9262 |
|
| 111 |
+
| 0.0 | 28.5 | 3648 | 0.3666 | 0.9262 |
|
| 112 |
+
| 0.0 | 29.0 | 3712 | 0.3670 | 0.9262 |
|
| 113 |
+
| 0.0 | 29.5 | 3776 | 0.3673 | 0.9262 |
|
| 114 |
+
| 0.0 | 30.0 | 3840 | 0.3674 | 0.9262 |
|
| 115 |
|
| 116 |
|
| 117 |
### Framework versions
|
|
|
|
| 119 |
- Transformers 4.48.3
|
| 120 |
- Pytorch 2.5.1+cu124
|
| 121 |
- Datasets 3.3.2
|
| 122 |
+
- Tokenizers 0.21.1
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 15.0,
|
| 3 |
+
"total_flos": 6.132781352484864e+17,
|
| 4 |
+
"train_loss": 0.1362585227402936,
|
| 5 |
+
"train_runtime": 4195.4959,
|
| 6 |
+
"train_samples_per_second": 29.289,
|
| 7 |
+
"train_steps_per_second": 0.458
|
| 8 |
+
}
|
logs/events.out.tfevents.1743190651.4955dc82343d.1607.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6c7af3195b7de6bb3689f57bca3872a46b11102d5d98ff9a890b9cc3db0c36ba
|
| 3 |
+
size 22083
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 15.0,
|
| 3 |
+
"total_flos": 6.132781352484864e+17,
|
| 4 |
+
"train_loss": 0.1362585227402936,
|
| 5 |
+
"train_runtime": 4195.4959,
|
| 6 |
+
"train_samples_per_second": 29.289,
|
| 7 |
+
"train_steps_per_second": 0.458
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,522 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.2626972794532776,
|
| 3 |
+
"best_model_checkpoint": "./drive/Shareddrives/CS198-Drones/[v4] Training Output/vit-tiny-patch16-224_rice-leaf-disease-augmented-v4_fft/checkpoint-448",
|
| 4 |
+
"epoch": 15.0,
|
| 5 |
+
"eval_steps": 64,
|
| 6 |
+
"global_step": 1920,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.5,
|
| 13 |
+
"grad_norm": 9.17007064819336,
|
| 14 |
+
"learning_rate": 7.265625e-06,
|
| 15 |
+
"loss": 2.0564,
|
| 16 |
+
"step": 64
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.5,
|
| 20 |
+
"eval_accuracy": 0.4899328859060403,
|
| 21 |
+
"eval_loss": 1.4541155099868774,
|
| 22 |
+
"eval_runtime": 8.722,
|
| 23 |
+
"eval_samples_per_second": 34.167,
|
| 24 |
+
"eval_steps_per_second": 0.573,
|
| 25 |
+
"step": 64
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 1.0,
|
| 29 |
+
"grad_norm": 13.332218170166016,
|
| 30 |
+
"learning_rate": 1.4765625e-05,
|
| 31 |
+
"loss": 1.0767,
|
| 32 |
+
"step": 128
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 1.0,
|
| 36 |
+
"eval_accuracy": 0.7651006711409396,
|
| 37 |
+
"eval_loss": 0.6909474730491638,
|
| 38 |
+
"eval_runtime": 8.6475,
|
| 39 |
+
"eval_samples_per_second": 34.461,
|
| 40 |
+
"eval_steps_per_second": 0.578,
|
| 41 |
+
"step": 128
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 1.5,
|
| 45 |
+
"grad_norm": 9.764442443847656,
|
| 46 |
+
"learning_rate": 2.2265625e-05,
|
| 47 |
+
"loss": 0.4917,
|
| 48 |
+
"step": 192
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 1.5,
|
| 52 |
+
"eval_accuracy": 0.8322147651006712,
|
| 53 |
+
"eval_loss": 0.4307171106338501,
|
| 54 |
+
"eval_runtime": 8.6026,
|
| 55 |
+
"eval_samples_per_second": 34.641,
|
| 56 |
+
"eval_steps_per_second": 0.581,
|
| 57 |
+
"step": 192
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 2.0,
|
| 61 |
+
"grad_norm": 9.665606498718262,
|
| 62 |
+
"learning_rate": 2.9765625e-05,
|
| 63 |
+
"loss": 0.285,
|
| 64 |
+
"step": 256
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"epoch": 2.0,
|
| 68 |
+
"eval_accuracy": 0.9026845637583892,
|
| 69 |
+
"eval_loss": 0.2932307720184326,
|
| 70 |
+
"eval_runtime": 7.7763,
|
| 71 |
+
"eval_samples_per_second": 38.322,
|
| 72 |
+
"eval_steps_per_second": 0.643,
|
| 73 |
+
"step": 256
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 2.5,
|
| 77 |
+
"grad_norm": 8.143477439880371,
|
| 78 |
+
"learning_rate": 2.9084596206825315e-05,
|
| 79 |
+
"loss": 0.0902,
|
| 80 |
+
"step": 320
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 2.5,
|
| 84 |
+
"eval_accuracy": 0.8993288590604027,
|
| 85 |
+
"eval_loss": 0.31344401836395264,
|
| 86 |
+
"eval_runtime": 7.522,
|
| 87 |
+
"eval_samples_per_second": 39.617,
|
| 88 |
+
"eval_steps_per_second": 0.665,
|
| 89 |
+
"step": 320
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 3.0,
|
| 93 |
+
"grad_norm": 2.6106486320495605,
|
| 94 |
+
"learning_rate": 2.633961484257573e-05,
|
| 95 |
+
"loss": 0.0588,
|
| 96 |
+
"step": 384
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 3.0,
|
| 100 |
+
"eval_accuracy": 0.9161073825503355,
|
| 101 |
+
"eval_loss": 0.3075862526893616,
|
| 102 |
+
"eval_runtime": 8.3334,
|
| 103 |
+
"eval_samples_per_second": 35.76,
|
| 104 |
+
"eval_steps_per_second": 0.6,
|
| 105 |
+
"step": 384
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 3.5,
|
| 109 |
+
"grad_norm": 1.9543571472167969,
|
| 110 |
+
"learning_rate": 2.212085192038453e-05,
|
| 111 |
+
"loss": 0.0155,
|
| 112 |
+
"step": 448
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 3.5,
|
| 116 |
+
"eval_accuracy": 0.9395973154362416,
|
| 117 |
+
"eval_loss": 0.2626972794532776,
|
| 118 |
+
"eval_runtime": 8.3127,
|
| 119 |
+
"eval_samples_per_second": 35.849,
|
| 120 |
+
"eval_steps_per_second": 0.601,
|
| 121 |
+
"step": 448
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 4.0,
|
| 125 |
+
"grad_norm": 0.38745102286338806,
|
| 126 |
+
"learning_rate": 1.6976609572058592e-05,
|
| 127 |
+
"loss": 0.0066,
|
| 128 |
+
"step": 512
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 4.0,
|
| 132 |
+
"eval_accuracy": 0.9295302013422819,
|
| 133 |
+
"eval_loss": 0.299156129360199,
|
| 134 |
+
"eval_runtime": 7.2879,
|
| 135 |
+
"eval_samples_per_second": 40.89,
|
| 136 |
+
"eval_steps_per_second": 0.686,
|
| 137 |
+
"step": 512
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 4.5,
|
| 141 |
+
"grad_norm": 0.10160534083843231,
|
| 142 |
+
"learning_rate": 1.1575472190259976e-05,
|
| 143 |
+
"loss": 0.0017,
|
| 144 |
+
"step": 576
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 4.5,
|
| 148 |
+
"eval_accuracy": 0.9228187919463087,
|
| 149 |
+
"eval_loss": 0.2935960590839386,
|
| 150 |
+
"eval_runtime": 8.4642,
|
| 151 |
+
"eval_samples_per_second": 35.207,
|
| 152 |
+
"eval_steps_per_second": 0.591,
|
| 153 |
+
"step": 576
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 5.0,
|
| 157 |
+
"grad_norm": 0.13512970507144928,
|
| 158 |
+
"learning_rate": 6.619412176671753e-06,
|
| 159 |
+
"loss": 0.0009,
|
| 160 |
+
"step": 640
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 5.0,
|
| 164 |
+
"eval_accuracy": 0.9228187919463087,
|
| 165 |
+
"eval_loss": 0.29606831073760986,
|
| 166 |
+
"eval_runtime": 8.1466,
|
| 167 |
+
"eval_samples_per_second": 36.58,
|
| 168 |
+
"eval_steps_per_second": 0.614,
|
| 169 |
+
"step": 640
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 5.5,
|
| 173 |
+
"grad_norm": 0.08882313966751099,
|
| 174 |
+
"learning_rate": 2.7525563336129812e-06,
|
| 175 |
+
"loss": 0.0006,
|
| 176 |
+
"step": 704
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 5.5,
|
| 180 |
+
"eval_accuracy": 0.9228187919463087,
|
| 181 |
+
"eval_loss": 0.30046388506889343,
|
| 182 |
+
"eval_runtime": 7.478,
|
| 183 |
+
"eval_samples_per_second": 39.85,
|
| 184 |
+
"eval_steps_per_second": 0.669,
|
| 185 |
+
"step": 704
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 6.0,
|
| 189 |
+
"grad_norm": 0.08427230268716812,
|
| 190 |
+
"learning_rate": 4.774703044353035e-07,
|
| 191 |
+
"loss": 0.0005,
|
| 192 |
+
"step": 768
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 6.0,
|
| 196 |
+
"eval_accuracy": 0.9228187919463087,
|
| 197 |
+
"eval_loss": 0.3003771901130676,
|
| 198 |
+
"eval_runtime": 8.3653,
|
| 199 |
+
"eval_samples_per_second": 35.623,
|
| 200 |
+
"eval_steps_per_second": 0.598,
|
| 201 |
+
"step": 768
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 6.5,
|
| 205 |
+
"grad_norm": 0.03683856502175331,
|
| 206 |
+
"learning_rate": 2.9910158634081504e-05,
|
| 207 |
+
"loss": 0.0005,
|
| 208 |
+
"step": 832
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 6.5,
|
| 212 |
+
"eval_accuracy": 0.9261744966442953,
|
| 213 |
+
"eval_loss": 0.2867479920387268,
|
| 214 |
+
"eval_runtime": 7.215,
|
| 215 |
+
"eval_samples_per_second": 41.303,
|
| 216 |
+
"eval_steps_per_second": 0.693,
|
| 217 |
+
"step": 832
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 7.0,
|
| 221 |
+
"grad_norm": 0.04806307703256607,
|
| 222 |
+
"learning_rate": 2.8359951312200077e-05,
|
| 223 |
+
"loss": 0.0004,
|
| 224 |
+
"step": 896
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 7.0,
|
| 228 |
+
"eval_accuracy": 0.9295302013422819,
|
| 229 |
+
"eval_loss": 0.2976870834827423,
|
| 230 |
+
"eval_runtime": 8.3401,
|
| 231 |
+
"eval_samples_per_second": 35.731,
|
| 232 |
+
"eval_steps_per_second": 0.6,
|
| 233 |
+
"step": 896
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 7.5,
|
| 237 |
+
"grad_norm": 0.033956822007894516,
|
| 238 |
+
"learning_rate": 2.5073384322705278e-05,
|
| 239 |
+
"loss": 0.0003,
|
| 240 |
+
"step": 960
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 7.5,
|
| 244 |
+
"eval_accuracy": 0.9295302013422819,
|
| 245 |
+
"eval_loss": 0.2943709194660187,
|
| 246 |
+
"eval_runtime": 7.6673,
|
| 247 |
+
"eval_samples_per_second": 38.867,
|
| 248 |
+
"eval_steps_per_second": 0.652,
|
| 249 |
+
"step": 960
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 8.0,
|
| 253 |
+
"grad_norm": 0.031262967735528946,
|
| 254 |
+
"learning_rate": 2.0477604608884026e-05,
|
| 255 |
+
"loss": 0.0002,
|
| 256 |
+
"step": 1024
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 8.0,
|
| 260 |
+
"eval_accuracy": 0.9295302013422819,
|
| 261 |
+
"eval_loss": 0.30740392208099365,
|
| 262 |
+
"eval_runtime": 8.2818,
|
| 263 |
+
"eval_samples_per_second": 35.982,
|
| 264 |
+
"eval_steps_per_second": 0.604,
|
| 265 |
+
"step": 1024
|
| 266 |
+
},
|
| 267 |
+
{
|
| 268 |
+
"epoch": 8.5,
|
| 269 |
+
"grad_norm": 0.01630540005862713,
|
| 270 |
+
"learning_rate": 1.516991423792483e-05,
|
| 271 |
+
"loss": 0.0002,
|
| 272 |
+
"step": 1088
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 8.5,
|
| 276 |
+
"eval_accuracy": 0.9328859060402684,
|
| 277 |
+
"eval_loss": 0.3053071200847626,
|
| 278 |
+
"eval_runtime": 8.2512,
|
| 279 |
+
"eval_samples_per_second": 36.116,
|
| 280 |
+
"eval_steps_per_second": 0.606,
|
| 281 |
+
"step": 1088
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 9.0,
|
| 285 |
+
"grad_norm": 0.016130153089761734,
|
| 286 |
+
"learning_rate": 9.840140535762432e-06,
|
| 287 |
+
"loss": 0.0002,
|
| 288 |
+
"step": 1152
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 9.0,
|
| 292 |
+
"eval_accuracy": 0.9295302013422819,
|
| 293 |
+
"eval_loss": 0.3097546696662903,
|
| 294 |
+
"eval_runtime": 8.418,
|
| 295 |
+
"eval_samples_per_second": 35.4,
|
| 296 |
+
"eval_steps_per_second": 0.594,
|
| 297 |
+
"step": 1152
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 9.5,
|
| 301 |
+
"grad_norm": 0.014978409744799137,
|
| 302 |
+
"learning_rate": 5.180980944002794e-06,
|
| 303 |
+
"loss": 0.0001,
|
| 304 |
+
"step": 1216
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 9.5,
|
| 308 |
+
"eval_accuracy": 0.9295302013422819,
|
| 309 |
+
"eval_loss": 0.310248464345932,
|
| 310 |
+
"eval_runtime": 7.2498,
|
| 311 |
+
"eval_samples_per_second": 41.105,
|
| 312 |
+
"eval_steps_per_second": 0.69,
|
| 313 |
+
"step": 1216
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 10.0,
|
| 317 |
+
"grad_norm": 0.013572459109127522,
|
| 318 |
+
"learning_rate": 1.7979748550475833e-06,
|
| 319 |
+
"loss": 0.0001,
|
| 320 |
+
"step": 1280
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 10.0,
|
| 324 |
+
"eval_accuracy": 0.9261744966442953,
|
| 325 |
+
"eval_loss": 0.3105408847332001,
|
| 326 |
+
"eval_runtime": 8.3297,
|
| 327 |
+
"eval_samples_per_second": 35.776,
|
| 328 |
+
"eval_steps_per_second": 0.6,
|
| 329 |
+
"step": 1280
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 10.5,
|
| 333 |
+
"grad_norm": 0.010590254329144955,
|
| 334 |
+
"learning_rate": 1.3080316225364152e-07,
|
| 335 |
+
"loss": 0.0001,
|
| 336 |
+
"step": 1344
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 10.5,
|
| 340 |
+
"eval_accuracy": 0.9261744966442953,
|
| 341 |
+
"eval_loss": 0.3105214238166809,
|
| 342 |
+
"eval_runtime": 7.2656,
|
| 343 |
+
"eval_samples_per_second": 41.015,
|
| 344 |
+
"eval_steps_per_second": 0.688,
|
| 345 |
+
"step": 1344
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 11.0,
|
| 349 |
+
"grad_norm": 0.011299003846943378,
|
| 350 |
+
"learning_rate": 2.9603855973006482e-05,
|
| 351 |
+
"loss": 0.0001,
|
| 352 |
+
"step": 1408
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 11.0,
|
| 356 |
+
"eval_accuracy": 0.9261744966442953,
|
| 357 |
+
"eval_loss": 0.32019779086112976,
|
| 358 |
+
"eval_runtime": 8.1649,
|
| 359 |
+
"eval_samples_per_second": 36.498,
|
| 360 |
+
"eval_steps_per_second": 0.612,
|
| 361 |
+
"step": 1408
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 11.5,
|
| 365 |
+
"grad_norm": 0.016378453001379967,
|
| 366 |
+
"learning_rate": 2.7440488243452587e-05,
|
| 367 |
+
"loss": 0.0001,
|
| 368 |
+
"step": 1472
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 11.5,
|
| 372 |
+
"eval_accuracy": 0.9295302013422819,
|
| 373 |
+
"eval_loss": 0.3183320462703705,
|
| 374 |
+
"eval_runtime": 8.0678,
|
| 375 |
+
"eval_samples_per_second": 36.937,
|
| 376 |
+
"eval_steps_per_second": 0.62,
|
| 377 |
+
"step": 1472
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 12.0,
|
| 381 |
+
"grad_norm": 0.007818573154509068,
|
| 382 |
+
"learning_rate": 2.3660261176123762e-05,
|
| 383 |
+
"loss": 0.0001,
|
| 384 |
+
"step": 1536
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 12.0,
|
| 388 |
+
"eval_accuracy": 0.9328859060402684,
|
| 389 |
+
"eval_loss": 0.3130977153778076,
|
| 390 |
+
"eval_runtime": 8.2508,
|
| 391 |
+
"eval_samples_per_second": 36.118,
|
| 392 |
+
"eval_steps_per_second": 0.606,
|
| 393 |
+
"step": 1536
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 12.5,
|
| 397 |
+
"grad_norm": 0.00784409698098898,
|
| 398 |
+
"learning_rate": 1.875448148769462e-05,
|
| 399 |
+
"loss": 0.0001,
|
| 400 |
+
"step": 1600
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 12.5,
|
| 404 |
+
"eval_accuracy": 0.9295302013422819,
|
| 405 |
+
"eval_loss": 0.3157329261302948,
|
| 406 |
+
"eval_runtime": 7.1968,
|
| 407 |
+
"eval_samples_per_second": 41.407,
|
| 408 |
+
"eval_steps_per_second": 0.695,
|
| 409 |
+
"step": 1600
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 13.0,
|
| 413 |
+
"grad_norm": 0.006332057528197765,
|
| 414 |
+
"learning_rate": 1.3360741171588585e-05,
|
| 415 |
+
"loss": 0.0001,
|
| 416 |
+
"step": 1664
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 13.0,
|
| 420 |
+
"eval_accuracy": 0.9228187919463087,
|
| 421 |
+
"eval_loss": 0.3237887918949127,
|
| 422 |
+
"eval_runtime": 8.2082,
|
| 423 |
+
"eval_samples_per_second": 36.305,
|
| 424 |
+
"eval_steps_per_second": 0.609,
|
| 425 |
+
"step": 1664
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 13.5,
|
| 429 |
+
"grad_norm": 0.006929404567927122,
|
| 430 |
+
"learning_rate": 8.180051251245103e-06,
|
| 431 |
+
"loss": 0.0001,
|
| 432 |
+
"step": 1728
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 13.5,
|
| 436 |
+
"eval_accuracy": 0.9228187919463087,
|
| 437 |
+
"eval_loss": 0.3219589591026306,
|
| 438 |
+
"eval_runtime": 7.2368,
|
| 439 |
+
"eval_samples_per_second": 41.179,
|
| 440 |
+
"eval_steps_per_second": 0.691,
|
| 441 |
+
"step": 1728
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 14.0,
|
| 445 |
+
"grad_norm": 0.005624026525765657,
|
| 446 |
+
"learning_rate": 3.885733119675617e-06,
|
| 447 |
+
"loss": 0.0001,
|
| 448 |
+
"step": 1792
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 14.0,
|
| 452 |
+
"eval_accuracy": 0.9228187919463087,
|
| 453 |
+
"eval_loss": 0.3266230821609497,
|
| 454 |
+
"eval_runtime": 7.1645,
|
| 455 |
+
"eval_samples_per_second": 41.594,
|
| 456 |
+
"eval_steps_per_second": 0.698,
|
| 457 |
+
"step": 1792
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 14.5,
|
| 461 |
+
"grad_norm": 0.006068665534257889,
|
| 462 |
+
"learning_rate": 1.0359086314671929e-06,
|
| 463 |
+
"loss": 0.0001,
|
| 464 |
+
"step": 1856
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 14.5,
|
| 468 |
+
"eval_accuracy": 0.9228187919463087,
|
| 469 |
+
"eval_loss": 0.32735249400138855,
|
| 470 |
+
"eval_runtime": 7.4237,
|
| 471 |
+
"eval_samples_per_second": 40.142,
|
| 472 |
+
"eval_steps_per_second": 0.674,
|
| 473 |
+
"step": 1856
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 15.0,
|
| 477 |
+
"grad_norm": 0.007146658841520548,
|
| 478 |
+
"learning_rate": 9.62392481628771e-10,
|
| 479 |
+
"loss": 0.0001,
|
| 480 |
+
"step": 1920
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 15.0,
|
| 484 |
+
"eval_accuracy": 0.9228187919463087,
|
| 485 |
+
"eval_loss": 0.32691627740859985,
|
| 486 |
+
"eval_runtime": 7.1631,
|
| 487 |
+
"eval_samples_per_second": 41.602,
|
| 488 |
+
"eval_steps_per_second": 0.698,
|
| 489 |
+
"step": 1920
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 15.0,
|
| 493 |
+
"step": 1920,
|
| 494 |
+
"total_flos": 6.132781352484864e+17,
|
| 495 |
+
"train_loss": 0.1362585227402936,
|
| 496 |
+
"train_runtime": 4195.4959,
|
| 497 |
+
"train_samples_per_second": 29.289,
|
| 498 |
+
"train_steps_per_second": 0.458
|
| 499 |
+
}
|
| 500 |
+
],
|
| 501 |
+
"logging_steps": 64,
|
| 502 |
+
"max_steps": 1920,
|
| 503 |
+
"num_input_tokens_seen": 0,
|
| 504 |
+
"num_train_epochs": 15,
|
| 505 |
+
"save_steps": 64,
|
| 506 |
+
"stateful_callbacks": {
|
| 507 |
+
"TrainerControl": {
|
| 508 |
+
"args": {
|
| 509 |
+
"should_epoch_stop": false,
|
| 510 |
+
"should_evaluate": false,
|
| 511 |
+
"should_log": false,
|
| 512 |
+
"should_save": true,
|
| 513 |
+
"should_training_stop": true
|
| 514 |
+
},
|
| 515 |
+
"attributes": {}
|
| 516 |
+
}
|
| 517 |
+
},
|
| 518 |
+
"total_flos": 6.132781352484864e+17,
|
| 519 |
+
"train_batch_size": 64,
|
| 520 |
+
"trial_name": null,
|
| 521 |
+
"trial_params": null
|
| 522 |
+
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5496
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6637a4431fabd444509438423930c601c6a4ae8966189a2b6318b9bbda03930c
|
| 3 |
size 5496
|
training_metrics.xlsx
CHANGED
|
Binary files a/training_metrics.xlsx and b/training_metrics.xlsx differ
|
|
|