All Dunn!!!
Browse files
README.md
CHANGED
|
@@ -14,6 +14,9 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 14 |
# trocr-large-printed-cmc7_tesseract_MICR_ocr
|
| 15 |
|
| 16 |
This model is a fine-tuned version of [microsoft/trocr-large-printed](https://huggingface.co/microsoft/trocr-large-printed) on an unknown dataset.
|
|
|
|
|
|
|
|
|
|
| 17 |
|
| 18 |
## Model description
|
| 19 |
|
|
|
|
| 14 |
# trocr-large-printed-cmc7_tesseract_MICR_ocr
|
| 15 |
|
| 16 |
This model is a fine-tuned version of [microsoft/trocr-large-printed](https://huggingface.co/microsoft/trocr-large-printed) on an unknown dataset.
|
| 17 |
+
It achieves the following results on the evaluation set:
|
| 18 |
+
- Loss: 0.2143
|
| 19 |
+
- Cer: 0.0027
|
| 20 |
|
| 21 |
## Model description
|
| 22 |
|
generation_config.json
CHANGED
|
@@ -1,5 +1,4 @@
|
|
| 1 |
{
|
| 2 |
-
"_from_model_config": true,
|
| 3 |
"bos_token_id": 0,
|
| 4 |
"decoder_start_token_id": 2,
|
| 5 |
"early_stopping": true,
|
|
|
|
| 1 |
{
|
|
|
|
| 2 |
"bos_token_id": 0,
|
| 3 |
"decoder_start_token_id": 2,
|
| 4 |
"early_stopping": true,
|
runs/Dec19_16-56-04_c5bc465dc86b/events.out.tfevents.1734636742.c5bc465dc86b.456.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ab5d6756ee1e331651b0c84b0779dcc79cc5b3c384c9deea2fab0f902756b786
|
| 3 |
+
size 406
|
trainer_state.json
ADDED
|
@@ -0,0 +1,749 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 5.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 3120,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0016025641025641025,
|
| 13 |
+
"grad_norm": 213.62364196777344,
|
| 14 |
+
"learning_rate": 4.998397435897436e-05,
|
| 15 |
+
"loss": 9.3378,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.049679487179487176,
|
| 20 |
+
"grad_norm": 43.086360931396484,
|
| 21 |
+
"learning_rate": 4.950320512820513e-05,
|
| 22 |
+
"loss": 2.5252,
|
| 23 |
+
"step": 31
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.09935897435897435,
|
| 27 |
+
"grad_norm": 48.219451904296875,
|
| 28 |
+
"learning_rate": 4.9006410256410256e-05,
|
| 29 |
+
"loss": 1.5213,
|
| 30 |
+
"step": 62
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.14903846153846154,
|
| 34 |
+
"grad_norm": 30.23198127746582,
|
| 35 |
+
"learning_rate": 4.850961538461539e-05,
|
| 36 |
+
"loss": 1.9636,
|
| 37 |
+
"step": 93
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.1987179487179487,
|
| 41 |
+
"grad_norm": 52.96772766113281,
|
| 42 |
+
"learning_rate": 4.8012820512820516e-05,
|
| 43 |
+
"loss": 1.3545,
|
| 44 |
+
"step": 124
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.2483974358974359,
|
| 48 |
+
"grad_norm": 48.79069900512695,
|
| 49 |
+
"learning_rate": 4.751602564102564e-05,
|
| 50 |
+
"loss": 1.4998,
|
| 51 |
+
"step": 155
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.2980769230769231,
|
| 55 |
+
"grad_norm": 38.8032341003418,
|
| 56 |
+
"learning_rate": 4.701923076923077e-05,
|
| 57 |
+
"loss": 1.4654,
|
| 58 |
+
"step": 186
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.34775641025641024,
|
| 62 |
+
"grad_norm": 106.62429809570312,
|
| 63 |
+
"learning_rate": 4.65224358974359e-05,
|
| 64 |
+
"loss": 1.8027,
|
| 65 |
+
"step": 217
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.3974358974358974,
|
| 69 |
+
"grad_norm": 25.370744705200195,
|
| 70 |
+
"learning_rate": 4.602564102564102e-05,
|
| 71 |
+
"loss": 1.4394,
|
| 72 |
+
"step": 248
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.44711538461538464,
|
| 76 |
+
"grad_norm": 82.52824401855469,
|
| 77 |
+
"learning_rate": 4.5528846153846157e-05,
|
| 78 |
+
"loss": 1.4573,
|
| 79 |
+
"step": 279
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.4967948717948718,
|
| 83 |
+
"grad_norm": 19.148210525512695,
|
| 84 |
+
"learning_rate": 4.503205128205128e-05,
|
| 85 |
+
"loss": 1.2547,
|
| 86 |
+
"step": 310
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.5464743589743589,
|
| 90 |
+
"grad_norm": 46.984779357910156,
|
| 91 |
+
"learning_rate": 4.453525641025642e-05,
|
| 92 |
+
"loss": 1.3297,
|
| 93 |
+
"step": 341
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.5961538461538461,
|
| 97 |
+
"grad_norm": 45.50709915161133,
|
| 98 |
+
"learning_rate": 4.403846153846154e-05,
|
| 99 |
+
"loss": 1.3359,
|
| 100 |
+
"step": 372
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.6458333333333334,
|
| 104 |
+
"grad_norm": 36.45267105102539,
|
| 105 |
+
"learning_rate": 4.354166666666667e-05,
|
| 106 |
+
"loss": 1.2466,
|
| 107 |
+
"step": 403
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.6955128205128205,
|
| 111 |
+
"grad_norm": 20.8362979888916,
|
| 112 |
+
"learning_rate": 4.30448717948718e-05,
|
| 113 |
+
"loss": 1.1206,
|
| 114 |
+
"step": 434
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.7451923076923077,
|
| 118 |
+
"grad_norm": 64.35557556152344,
|
| 119 |
+
"learning_rate": 4.2548076923076924e-05,
|
| 120 |
+
"loss": 1.3676,
|
| 121 |
+
"step": 465
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.7948717948717948,
|
| 125 |
+
"grad_norm": 12.930615425109863,
|
| 126 |
+
"learning_rate": 4.205128205128206e-05,
|
| 127 |
+
"loss": 1.3871,
|
| 128 |
+
"step": 496
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.844551282051282,
|
| 132 |
+
"grad_norm": 51.6870002746582,
|
| 133 |
+
"learning_rate": 4.1554487179487184e-05,
|
| 134 |
+
"loss": 1.0038,
|
| 135 |
+
"step": 527
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.8942307692307693,
|
| 139 |
+
"grad_norm": 14.677131652832031,
|
| 140 |
+
"learning_rate": 4.105769230769231e-05,
|
| 141 |
+
"loss": 1.2283,
|
| 142 |
+
"step": 558
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.9439102564102564,
|
| 146 |
+
"grad_norm": 43.24152755737305,
|
| 147 |
+
"learning_rate": 4.056089743589744e-05,
|
| 148 |
+
"loss": 1.0872,
|
| 149 |
+
"step": 589
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.9935897435897436,
|
| 153 |
+
"grad_norm": 25.641794204711914,
|
| 154 |
+
"learning_rate": 4.006410256410257e-05,
|
| 155 |
+
"loss": 0.9991,
|
| 156 |
+
"step": 620
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 1.0432692307692308,
|
| 160 |
+
"grad_norm": 36.622703552246094,
|
| 161 |
+
"learning_rate": 3.956730769230769e-05,
|
| 162 |
+
"loss": 0.9941,
|
| 163 |
+
"step": 651
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 1.092948717948718,
|
| 167 |
+
"grad_norm": 17.249757766723633,
|
| 168 |
+
"learning_rate": 3.9070512820512824e-05,
|
| 169 |
+
"loss": 0.9877,
|
| 170 |
+
"step": 682
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 1.142628205128205,
|
| 174 |
+
"grad_norm": 31.591552734375,
|
| 175 |
+
"learning_rate": 3.857371794871795e-05,
|
| 176 |
+
"loss": 0.8917,
|
| 177 |
+
"step": 713
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 1.1923076923076923,
|
| 181 |
+
"grad_norm": 28.23312759399414,
|
| 182 |
+
"learning_rate": 3.807692307692308e-05,
|
| 183 |
+
"loss": 0.9264,
|
| 184 |
+
"step": 744
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 1.2419871794871795,
|
| 188 |
+
"grad_norm": 37.707183837890625,
|
| 189 |
+
"learning_rate": 3.7580128205128204e-05,
|
| 190 |
+
"loss": 0.849,
|
| 191 |
+
"step": 775
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 1.2916666666666667,
|
| 195 |
+
"grad_norm": 12.67361831665039,
|
| 196 |
+
"learning_rate": 3.708333333333334e-05,
|
| 197 |
+
"loss": 0.8255,
|
| 198 |
+
"step": 806
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 1.3413461538461537,
|
| 202 |
+
"grad_norm": 39.1461181640625,
|
| 203 |
+
"learning_rate": 3.658653846153846e-05,
|
| 204 |
+
"loss": 0.7372,
|
| 205 |
+
"step": 837
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 1.391025641025641,
|
| 209 |
+
"grad_norm": 33.27256393432617,
|
| 210 |
+
"learning_rate": 3.608974358974359e-05,
|
| 211 |
+
"loss": 0.8227,
|
| 212 |
+
"step": 868
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.4407051282051282,
|
| 216 |
+
"grad_norm": 545.9265747070312,
|
| 217 |
+
"learning_rate": 3.559294871794872e-05,
|
| 218 |
+
"loss": 0.7941,
|
| 219 |
+
"step": 899
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 1.4903846153846154,
|
| 223 |
+
"grad_norm": 23.55372428894043,
|
| 224 |
+
"learning_rate": 3.5096153846153845e-05,
|
| 225 |
+
"loss": 0.8686,
|
| 226 |
+
"step": 930
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 1.5400641025641026,
|
| 230 |
+
"grad_norm": 22.64493179321289,
|
| 231 |
+
"learning_rate": 3.459935897435898e-05,
|
| 232 |
+
"loss": 0.7059,
|
| 233 |
+
"step": 961
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 1.5897435897435899,
|
| 237 |
+
"grad_norm": 79.4326171875,
|
| 238 |
+
"learning_rate": 3.4102564102564105e-05,
|
| 239 |
+
"loss": 0.7567,
|
| 240 |
+
"step": 992
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 1.6394230769230769,
|
| 244 |
+
"grad_norm": 13.156254768371582,
|
| 245 |
+
"learning_rate": 3.360576923076923e-05,
|
| 246 |
+
"loss": 0.7138,
|
| 247 |
+
"step": 1023
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 1.689102564102564,
|
| 251 |
+
"grad_norm": 39.97475051879883,
|
| 252 |
+
"learning_rate": 3.310897435897436e-05,
|
| 253 |
+
"loss": 0.676,
|
| 254 |
+
"step": 1054
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.7387820512820513,
|
| 258 |
+
"grad_norm": 24.62538719177246,
|
| 259 |
+
"learning_rate": 3.261217948717949e-05,
|
| 260 |
+
"loss": 0.683,
|
| 261 |
+
"step": 1085
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 1.7884615384615383,
|
| 265 |
+
"grad_norm": 18.064680099487305,
|
| 266 |
+
"learning_rate": 3.211538461538462e-05,
|
| 267 |
+
"loss": 0.5746,
|
| 268 |
+
"step": 1116
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 1.8381410256410255,
|
| 272 |
+
"grad_norm": 10.56497573852539,
|
| 273 |
+
"learning_rate": 3.1618589743589746e-05,
|
| 274 |
+
"loss": 0.5881,
|
| 275 |
+
"step": 1147
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 1.8878205128205128,
|
| 279 |
+
"grad_norm": 43.007911682128906,
|
| 280 |
+
"learning_rate": 3.112179487179487e-05,
|
| 281 |
+
"loss": 0.78,
|
| 282 |
+
"step": 1178
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 1.9375,
|
| 286 |
+
"grad_norm": 25.748680114746094,
|
| 287 |
+
"learning_rate": 3.0625000000000006e-05,
|
| 288 |
+
"loss": 0.592,
|
| 289 |
+
"step": 1209
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 1.9871794871794872,
|
| 293 |
+
"grad_norm": 20.511089324951172,
|
| 294 |
+
"learning_rate": 3.012820512820513e-05,
|
| 295 |
+
"loss": 0.6807,
|
| 296 |
+
"step": 1240
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 2.0368589743589745,
|
| 300 |
+
"grad_norm": 26.275846481323242,
|
| 301 |
+
"learning_rate": 2.963141025641026e-05,
|
| 302 |
+
"loss": 0.6935,
|
| 303 |
+
"step": 1271
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 2.0865384615384617,
|
| 307 |
+
"grad_norm": 69.18566131591797,
|
| 308 |
+
"learning_rate": 2.913461538461539e-05,
|
| 309 |
+
"loss": 0.511,
|
| 310 |
+
"step": 1302
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 2.136217948717949,
|
| 314 |
+
"grad_norm": 4.620207786560059,
|
| 315 |
+
"learning_rate": 2.8637820512820513e-05,
|
| 316 |
+
"loss": 0.5338,
|
| 317 |
+
"step": 1333
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 2.185897435897436,
|
| 321 |
+
"grad_norm": 44.329566955566406,
|
| 322 |
+
"learning_rate": 2.8141025641025643e-05,
|
| 323 |
+
"loss": 0.4471,
|
| 324 |
+
"step": 1364
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 2.235576923076923,
|
| 328 |
+
"grad_norm": 24.95821762084961,
|
| 329 |
+
"learning_rate": 2.7644230769230773e-05,
|
| 330 |
+
"loss": 0.5295,
|
| 331 |
+
"step": 1395
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 2.28525641025641,
|
| 335 |
+
"grad_norm": 32.0108757019043,
|
| 336 |
+
"learning_rate": 2.7147435897435896e-05,
|
| 337 |
+
"loss": 0.5075,
|
| 338 |
+
"step": 1426
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 2.3349358974358974,
|
| 342 |
+
"grad_norm": 15.610992431640625,
|
| 343 |
+
"learning_rate": 2.6650641025641026e-05,
|
| 344 |
+
"loss": 0.4898,
|
| 345 |
+
"step": 1457
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 2.3846153846153846,
|
| 349 |
+
"grad_norm": 6.399534225463867,
|
| 350 |
+
"learning_rate": 2.6153846153846157e-05,
|
| 351 |
+
"loss": 0.4273,
|
| 352 |
+
"step": 1488
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 2.434294871794872,
|
| 356 |
+
"grad_norm": 22.35305404663086,
|
| 357 |
+
"learning_rate": 2.565705128205128e-05,
|
| 358 |
+
"loss": 0.4826,
|
| 359 |
+
"step": 1519
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 2.483974358974359,
|
| 363 |
+
"grad_norm": 26.82170867919922,
|
| 364 |
+
"learning_rate": 2.516025641025641e-05,
|
| 365 |
+
"loss": 0.4472,
|
| 366 |
+
"step": 1550
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 2.5336538461538463,
|
| 370 |
+
"grad_norm": 31.885419845581055,
|
| 371 |
+
"learning_rate": 2.466346153846154e-05,
|
| 372 |
+
"loss": 0.3986,
|
| 373 |
+
"step": 1581
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 2.5833333333333335,
|
| 377 |
+
"grad_norm": 21.020639419555664,
|
| 378 |
+
"learning_rate": 2.4166666666666667e-05,
|
| 379 |
+
"loss": 0.4856,
|
| 380 |
+
"step": 1612
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 2.6330128205128203,
|
| 384 |
+
"grad_norm": 10.991323471069336,
|
| 385 |
+
"learning_rate": 2.3669871794871794e-05,
|
| 386 |
+
"loss": 0.5008,
|
| 387 |
+
"step": 1643
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 2.6826923076923075,
|
| 391 |
+
"grad_norm": 17.077383041381836,
|
| 392 |
+
"learning_rate": 2.3173076923076924e-05,
|
| 393 |
+
"loss": 0.3585,
|
| 394 |
+
"step": 1674
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 2.7323717948717947,
|
| 398 |
+
"grad_norm": 6.274226188659668,
|
| 399 |
+
"learning_rate": 2.2676282051282054e-05,
|
| 400 |
+
"loss": 0.4101,
|
| 401 |
+
"step": 1705
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 2.782051282051282,
|
| 405 |
+
"grad_norm": 7.574007511138916,
|
| 406 |
+
"learning_rate": 2.217948717948718e-05,
|
| 407 |
+
"loss": 0.3604,
|
| 408 |
+
"step": 1736
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 2.831730769230769,
|
| 412 |
+
"grad_norm": 15.08122730255127,
|
| 413 |
+
"learning_rate": 2.168269230769231e-05,
|
| 414 |
+
"loss": 0.407,
|
| 415 |
+
"step": 1767
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 2.8814102564102564,
|
| 419 |
+
"grad_norm": 32.56257247924805,
|
| 420 |
+
"learning_rate": 2.1185897435897437e-05,
|
| 421 |
+
"loss": 0.4002,
|
| 422 |
+
"step": 1798
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 2.9310897435897436,
|
| 426 |
+
"grad_norm": 34.24089431762695,
|
| 427 |
+
"learning_rate": 2.0689102564102564e-05,
|
| 428 |
+
"loss": 0.429,
|
| 429 |
+
"step": 1829
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 2.980769230769231,
|
| 433 |
+
"grad_norm": 5.287622928619385,
|
| 434 |
+
"learning_rate": 2.0192307692307694e-05,
|
| 435 |
+
"loss": 0.352,
|
| 436 |
+
"step": 1860
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 3.030448717948718,
|
| 440 |
+
"grad_norm": 33.49539566040039,
|
| 441 |
+
"learning_rate": 1.969551282051282e-05,
|
| 442 |
+
"loss": 0.3348,
|
| 443 |
+
"step": 1891
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 3.0801282051282053,
|
| 447 |
+
"grad_norm": 3.116619825363159,
|
| 448 |
+
"learning_rate": 1.919871794871795e-05,
|
| 449 |
+
"loss": 0.2613,
|
| 450 |
+
"step": 1922
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 3.1298076923076925,
|
| 454 |
+
"grad_norm": 5.825009346008301,
|
| 455 |
+
"learning_rate": 1.8701923076923078e-05,
|
| 456 |
+
"loss": 0.279,
|
| 457 |
+
"step": 1953
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 3.1794871794871793,
|
| 461 |
+
"grad_norm": 2.2447965145111084,
|
| 462 |
+
"learning_rate": 1.8205128205128204e-05,
|
| 463 |
+
"loss": 0.3405,
|
| 464 |
+
"step": 1984
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 3.2291666666666665,
|
| 468 |
+
"grad_norm": 6.297438144683838,
|
| 469 |
+
"learning_rate": 1.7708333333333335e-05,
|
| 470 |
+
"loss": 0.3294,
|
| 471 |
+
"step": 2015
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 3.2788461538461537,
|
| 475 |
+
"grad_norm": 17.42154884338379,
|
| 476 |
+
"learning_rate": 1.721153846153846e-05,
|
| 477 |
+
"loss": 0.296,
|
| 478 |
+
"step": 2046
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 3.328525641025641,
|
| 482 |
+
"grad_norm": 11.834112167358398,
|
| 483 |
+
"learning_rate": 1.6714743589743588e-05,
|
| 484 |
+
"loss": 0.3767,
|
| 485 |
+
"step": 2077
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 3.378205128205128,
|
| 489 |
+
"grad_norm": 2.2308290004730225,
|
| 490 |
+
"learning_rate": 1.6217948717948718e-05,
|
| 491 |
+
"loss": 0.2352,
|
| 492 |
+
"step": 2108
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 3.4278846153846154,
|
| 496 |
+
"grad_norm": 13.97610855102539,
|
| 497 |
+
"learning_rate": 1.5721153846153845e-05,
|
| 498 |
+
"loss": 0.2937,
|
| 499 |
+
"step": 2139
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 3.4775641025641026,
|
| 503 |
+
"grad_norm": 4.005247116088867,
|
| 504 |
+
"learning_rate": 1.5224358974358973e-05,
|
| 505 |
+
"loss": 0.283,
|
| 506 |
+
"step": 2170
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 3.52724358974359,
|
| 510 |
+
"grad_norm": 5.180654525756836,
|
| 511 |
+
"learning_rate": 1.4727564102564103e-05,
|
| 512 |
+
"loss": 0.2773,
|
| 513 |
+
"step": 2201
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 3.5769230769230766,
|
| 517 |
+
"grad_norm": 6.487146854400635,
|
| 518 |
+
"learning_rate": 1.423076923076923e-05,
|
| 519 |
+
"loss": 0.2665,
|
| 520 |
+
"step": 2232
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 3.626602564102564,
|
| 524 |
+
"grad_norm": 2.202457904815674,
|
| 525 |
+
"learning_rate": 1.373397435897436e-05,
|
| 526 |
+
"loss": 0.2541,
|
| 527 |
+
"step": 2263
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 3.676282051282051,
|
| 531 |
+
"grad_norm": 47.195499420166016,
|
| 532 |
+
"learning_rate": 1.3237179487179489e-05,
|
| 533 |
+
"loss": 0.2362,
|
| 534 |
+
"step": 2294
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 3.7259615384615383,
|
| 538 |
+
"grad_norm": 21.26088523864746,
|
| 539 |
+
"learning_rate": 1.2740384615384615e-05,
|
| 540 |
+
"loss": 0.2331,
|
| 541 |
+
"step": 2325
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 3.7756410256410255,
|
| 545 |
+
"grad_norm": 16.77073860168457,
|
| 546 |
+
"learning_rate": 1.2243589743589744e-05,
|
| 547 |
+
"loss": 0.2134,
|
| 548 |
+
"step": 2356
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 3.8253205128205128,
|
| 552 |
+
"grad_norm": 2.246936559677124,
|
| 553 |
+
"learning_rate": 1.1746794871794872e-05,
|
| 554 |
+
"loss": 0.2437,
|
| 555 |
+
"step": 2387
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 3.875,
|
| 559 |
+
"grad_norm": 1.4183766841888428,
|
| 560 |
+
"learning_rate": 1.125e-05,
|
| 561 |
+
"loss": 0.2012,
|
| 562 |
+
"step": 2418
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 3.9246794871794872,
|
| 566 |
+
"grad_norm": 3.9523444175720215,
|
| 567 |
+
"learning_rate": 1.0753205128205129e-05,
|
| 568 |
+
"loss": 0.2204,
|
| 569 |
+
"step": 2449
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 3.9743589743589745,
|
| 573 |
+
"grad_norm": 3.564764976501465,
|
| 574 |
+
"learning_rate": 1.0256410256410256e-05,
|
| 575 |
+
"loss": 0.2034,
|
| 576 |
+
"step": 2480
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 4.024038461538462,
|
| 580 |
+
"grad_norm": 1.7849295139312744,
|
| 581 |
+
"learning_rate": 9.759615384615384e-06,
|
| 582 |
+
"loss": 0.1819,
|
| 583 |
+
"step": 2511
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 4.073717948717949,
|
| 587 |
+
"grad_norm": 1.7893842458724976,
|
| 588 |
+
"learning_rate": 9.262820512820514e-06,
|
| 589 |
+
"loss": 0.1585,
|
| 590 |
+
"step": 2542
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 4.123397435897436,
|
| 594 |
+
"grad_norm": 3.4071810245513916,
|
| 595 |
+
"learning_rate": 8.766025641025641e-06,
|
| 596 |
+
"loss": 0.1596,
|
| 597 |
+
"step": 2573
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 4.173076923076923,
|
| 601 |
+
"grad_norm": 28.766258239746094,
|
| 602 |
+
"learning_rate": 8.26923076923077e-06,
|
| 603 |
+
"loss": 0.1563,
|
| 604 |
+
"step": 2604
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 4.222756410256411,
|
| 608 |
+
"grad_norm": 3.066229820251465,
|
| 609 |
+
"learning_rate": 7.772435897435898e-06,
|
| 610 |
+
"loss": 0.1612,
|
| 611 |
+
"step": 2635
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 4.272435897435898,
|
| 615 |
+
"grad_norm": 1.4539713859558105,
|
| 616 |
+
"learning_rate": 7.275641025641026e-06,
|
| 617 |
+
"loss": 0.151,
|
| 618 |
+
"step": 2666
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 4.322115384615385,
|
| 622 |
+
"grad_norm": 3.9048240184783936,
|
| 623 |
+
"learning_rate": 6.778846153846154e-06,
|
| 624 |
+
"loss": 0.1514,
|
| 625 |
+
"step": 2697
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 4.371794871794872,
|
| 629 |
+
"grad_norm": 1.2496715784072876,
|
| 630 |
+
"learning_rate": 6.282051282051282e-06,
|
| 631 |
+
"loss": 0.1632,
|
| 632 |
+
"step": 2728
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 4.421474358974359,
|
| 636 |
+
"grad_norm": 1.516411542892456,
|
| 637 |
+
"learning_rate": 5.785256410256411e-06,
|
| 638 |
+
"loss": 0.1497,
|
| 639 |
+
"step": 2759
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 4.471153846153846,
|
| 643 |
+
"grad_norm": 2.217137575149536,
|
| 644 |
+
"learning_rate": 5.288461538461538e-06,
|
| 645 |
+
"loss": 0.1514,
|
| 646 |
+
"step": 2790
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 4.520833333333333,
|
| 650 |
+
"grad_norm": 1.7278540134429932,
|
| 651 |
+
"learning_rate": 4.791666666666667e-06,
|
| 652 |
+
"loss": 0.1409,
|
| 653 |
+
"step": 2821
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 4.57051282051282,
|
| 657 |
+
"grad_norm": 2.3950905799865723,
|
| 658 |
+
"learning_rate": 4.294871794871795e-06,
|
| 659 |
+
"loss": 0.1524,
|
| 660 |
+
"step": 2852
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 4.6201923076923075,
|
| 664 |
+
"grad_norm": 2.2806148529052734,
|
| 665 |
+
"learning_rate": 3.798076923076923e-06,
|
| 666 |
+
"loss": 0.1498,
|
| 667 |
+
"step": 2883
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 4.669871794871795,
|
| 671 |
+
"grad_norm": 2.3029630184173584,
|
| 672 |
+
"learning_rate": 3.3012820512820517e-06,
|
| 673 |
+
"loss": 0.1471,
|
| 674 |
+
"step": 2914
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 4.719551282051282,
|
| 678 |
+
"grad_norm": 1.592606782913208,
|
| 679 |
+
"learning_rate": 2.8044871794871797e-06,
|
| 680 |
+
"loss": 0.14,
|
| 681 |
+
"step": 2945
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 4.769230769230769,
|
| 685 |
+
"grad_norm": 3.534292221069336,
|
| 686 |
+
"learning_rate": 2.307692307692308e-06,
|
| 687 |
+
"loss": 0.1426,
|
| 688 |
+
"step": 2976
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 4.818910256410256,
|
| 692 |
+
"grad_norm": 2.344146966934204,
|
| 693 |
+
"learning_rate": 1.810897435897436e-06,
|
| 694 |
+
"loss": 0.1399,
|
| 695 |
+
"step": 3007
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 4.868589743589744,
|
| 699 |
+
"grad_norm": 11.425804138183594,
|
| 700 |
+
"learning_rate": 1.3141025641025643e-06,
|
| 701 |
+
"loss": 0.1374,
|
| 702 |
+
"step": 3038
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 4.918269230769231,
|
| 706 |
+
"grad_norm": 2.704406499862671,
|
| 707 |
+
"learning_rate": 8.173076923076924e-07,
|
| 708 |
+
"loss": 0.1389,
|
| 709 |
+
"step": 3069
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 4.967948717948718,
|
| 713 |
+
"grad_norm": 6.8166728019714355,
|
| 714 |
+
"learning_rate": 3.205128205128205e-07,
|
| 715 |
+
"loss": 0.1721,
|
| 716 |
+
"step": 3100
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 5.0,
|
| 720 |
+
"step": 3120,
|
| 721 |
+
"total_flos": 1.845867535870722e+19,
|
| 722 |
+
"train_loss": 0.6135548925552613,
|
| 723 |
+
"train_runtime": 6226.8155,
|
| 724 |
+
"train_samples_per_second": 2.003,
|
| 725 |
+
"train_steps_per_second": 0.501
|
| 726 |
+
}
|
| 727 |
+
],
|
| 728 |
+
"logging_steps": 31,
|
| 729 |
+
"max_steps": 3120,
|
| 730 |
+
"num_input_tokens_seen": 0,
|
| 731 |
+
"num_train_epochs": 5,
|
| 732 |
+
"save_steps": 500,
|
| 733 |
+
"stateful_callbacks": {
|
| 734 |
+
"TrainerControl": {
|
| 735 |
+
"args": {
|
| 736 |
+
"should_epoch_stop": false,
|
| 737 |
+
"should_evaluate": false,
|
| 738 |
+
"should_log": false,
|
| 739 |
+
"should_save": true,
|
| 740 |
+
"should_training_stop": true
|
| 741 |
+
},
|
| 742 |
+
"attributes": {}
|
| 743 |
+
}
|
| 744 |
+
},
|
| 745 |
+
"total_flos": 1.845867535870722e+19,
|
| 746 |
+
"train_batch_size": 4,
|
| 747 |
+
"trial_name": null,
|
| 748 |
+
"trial_params": null
|
| 749 |
+
}
|