Commit
•
890dde7
1
Parent(s):
f252fd7
End of training
Browse files- README.md +2 -2
- all_results.json +13 -0
- eval_results.json +8 -0
- runs/Apr02_01-48-53_6529928adb6f/events.out.tfevents.1712024211.6529928adb6f.1561.1 +3 -0
- train_results.json +8 -0
- trainer_state.json +540 -0
README.md
CHANGED
|
@@ -17,8 +17,8 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 17 |
|
| 18 |
This model is a fine-tuned version of [google/vit-base-patch16-224](https://huggingface.co/google/vit-base-patch16-224) on an unknown dataset.
|
| 19 |
It achieves the following results on the evaluation set:
|
| 20 |
-
- Loss: 0.
|
| 21 |
-
- Accuracy: 0.
|
| 22 |
|
| 23 |
## Model description
|
| 24 |
|
|
|
|
| 17 |
|
| 18 |
This model is a fine-tuned version of [google/vit-base-patch16-224](https://huggingface.co/google/vit-base-patch16-224) on an unknown dataset.
|
| 19 |
It achieves the following results on the evaluation set:
|
| 20 |
+
- Loss: 0.7345
|
| 21 |
+
- Accuracy: 0.8333
|
| 22 |
|
| 23 |
## Model description
|
| 24 |
|
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 37.94,
|
| 3 |
+
"eval_accuracy": 0.8333333333333334,
|
| 4 |
+
"eval_loss": 0.7344614863395691,
|
| 5 |
+
"eval_runtime": 1.6377,
|
| 6 |
+
"eval_samples_per_second": 36.637,
|
| 7 |
+
"eval_steps_per_second": 1.221,
|
| 8 |
+
"total_flos": 2.864620236542755e+18,
|
| 9 |
+
"train_loss": 0.20059432347818296,
|
| 10 |
+
"train_runtime": 1595.666,
|
| 11 |
+
"train_samples_per_second": 25.637,
|
| 12 |
+
"train_steps_per_second": 0.184
|
| 13 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 37.94,
|
| 3 |
+
"eval_accuracy": 0.8333333333333334,
|
| 4 |
+
"eval_loss": 0.7344614863395691,
|
| 5 |
+
"eval_runtime": 1.6377,
|
| 6 |
+
"eval_samples_per_second": 36.637,
|
| 7 |
+
"eval_steps_per_second": 1.221
|
| 8 |
+
}
|
runs/Apr02_01-48-53_6529928adb6f/events.out.tfevents.1712024211.6529928adb6f.1561.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5f450cd911659bdd1c7769e46cb0b66f05fa8a100b1d624e6dcfdcf70baf87e2
|
| 3 |
+
size 411
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 37.94,
|
| 3 |
+
"total_flos": 2.864620236542755e+18,
|
| 4 |
+
"train_loss": 0.20059432347818296,
|
| 5 |
+
"train_runtime": 1595.666,
|
| 6 |
+
"train_samples_per_second": 25.637,
|
| 7 |
+
"train_steps_per_second": 0.184
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,540 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.8333333333333334,
|
| 3 |
+
"best_model_checkpoint": "vit-base-patch16-224-dmae-va-U5-42/checkpoint-108",
|
| 4 |
+
"epoch": 37.935483870967744,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 294,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.9,
|
| 13 |
+
"eval_accuracy": 0.5,
|
| 14 |
+
"eval_loss": 1.28582763671875,
|
| 15 |
+
"eval_runtime": 1.6082,
|
| 16 |
+
"eval_samples_per_second": 37.31,
|
| 17 |
+
"eval_steps_per_second": 1.244,
|
| 18 |
+
"step": 7
|
| 19 |
+
},
|
| 20 |
+
{
|
| 21 |
+
"epoch": 1.55,
|
| 22 |
+
"grad_norm": 4.701062202453613,
|
| 23 |
+
"learning_rate": 2e-05,
|
| 24 |
+
"loss": 1.3455,
|
| 25 |
+
"step": 12
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 1.94,
|
| 29 |
+
"eval_accuracy": 0.48333333333333334,
|
| 30 |
+
"eval_loss": 1.1091300249099731,
|
| 31 |
+
"eval_runtime": 2.0547,
|
| 32 |
+
"eval_samples_per_second": 29.201,
|
| 33 |
+
"eval_steps_per_second": 0.973,
|
| 34 |
+
"step": 15
|
| 35 |
+
},
|
| 36 |
+
{
|
| 37 |
+
"epoch": 2.97,
|
| 38 |
+
"eval_accuracy": 0.5833333333333334,
|
| 39 |
+
"eval_loss": 0.8518259525299072,
|
| 40 |
+
"eval_runtime": 1.5248,
|
| 41 |
+
"eval_samples_per_second": 39.349,
|
| 42 |
+
"eval_steps_per_second": 1.312,
|
| 43 |
+
"step": 23
|
| 44 |
+
},
|
| 45 |
+
{
|
| 46 |
+
"epoch": 3.1,
|
| 47 |
+
"grad_norm": 2.8079495429992676,
|
| 48 |
+
"learning_rate": 4e-05,
|
| 49 |
+
"loss": 1.0067,
|
| 50 |
+
"step": 24
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 4.0,
|
| 54 |
+
"eval_accuracy": 0.7166666666666667,
|
| 55 |
+
"eval_loss": 0.7317402362823486,
|
| 56 |
+
"eval_runtime": 1.5389,
|
| 57 |
+
"eval_samples_per_second": 38.988,
|
| 58 |
+
"eval_steps_per_second": 1.3,
|
| 59 |
+
"step": 31
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 4.65,
|
| 63 |
+
"grad_norm": 3.14342999458313,
|
| 64 |
+
"learning_rate": 4.886363636363637e-05,
|
| 65 |
+
"loss": 0.6085,
|
| 66 |
+
"step": 36
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 4.9,
|
| 70 |
+
"eval_accuracy": 0.75,
|
| 71 |
+
"eval_loss": 0.694876492023468,
|
| 72 |
+
"eval_runtime": 1.5371,
|
| 73 |
+
"eval_samples_per_second": 39.035,
|
| 74 |
+
"eval_steps_per_second": 1.301,
|
| 75 |
+
"step": 38
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 5.94,
|
| 79 |
+
"eval_accuracy": 0.75,
|
| 80 |
+
"eval_loss": 0.6633104085922241,
|
| 81 |
+
"eval_runtime": 1.5303,
|
| 82 |
+
"eval_samples_per_second": 39.208,
|
| 83 |
+
"eval_steps_per_second": 1.307,
|
| 84 |
+
"step": 46
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 6.19,
|
| 88 |
+
"grad_norm": 2.108163833618164,
|
| 89 |
+
"learning_rate": 4.659090909090909e-05,
|
| 90 |
+
"loss": 0.3389,
|
| 91 |
+
"step": 48
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 6.97,
|
| 95 |
+
"eval_accuracy": 0.7666666666666667,
|
| 96 |
+
"eval_loss": 0.6791020035743713,
|
| 97 |
+
"eval_runtime": 1.5402,
|
| 98 |
+
"eval_samples_per_second": 38.956,
|
| 99 |
+
"eval_steps_per_second": 1.299,
|
| 100 |
+
"step": 54
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 7.74,
|
| 104 |
+
"grad_norm": 1.7717548608779907,
|
| 105 |
+
"learning_rate": 4.431818181818182e-05,
|
| 106 |
+
"loss": 0.1977,
|
| 107 |
+
"step": 60
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 8.0,
|
| 111 |
+
"eval_accuracy": 0.7333333333333333,
|
| 112 |
+
"eval_loss": 0.7010270357131958,
|
| 113 |
+
"eval_runtime": 1.6107,
|
| 114 |
+
"eval_samples_per_second": 37.252,
|
| 115 |
+
"eval_steps_per_second": 1.242,
|
| 116 |
+
"step": 62
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 8.9,
|
| 120 |
+
"eval_accuracy": 0.75,
|
| 121 |
+
"eval_loss": 0.697003960609436,
|
| 122 |
+
"eval_runtime": 1.5576,
|
| 123 |
+
"eval_samples_per_second": 38.522,
|
| 124 |
+
"eval_steps_per_second": 1.284,
|
| 125 |
+
"step": 69
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 9.29,
|
| 129 |
+
"grad_norm": 1.6967554092407227,
|
| 130 |
+
"learning_rate": 4.204545454545455e-05,
|
| 131 |
+
"loss": 0.1496,
|
| 132 |
+
"step": 72
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"epoch": 9.94,
|
| 136 |
+
"eval_accuracy": 0.8,
|
| 137 |
+
"eval_loss": 0.6983956098556519,
|
| 138 |
+
"eval_runtime": 1.5413,
|
| 139 |
+
"eval_samples_per_second": 38.929,
|
| 140 |
+
"eval_steps_per_second": 1.298,
|
| 141 |
+
"step": 77
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 10.84,
|
| 145 |
+
"grad_norm": 2.1696012020111084,
|
| 146 |
+
"learning_rate": 3.9772727272727275e-05,
|
| 147 |
+
"loss": 0.1194,
|
| 148 |
+
"step": 84
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 10.97,
|
| 152 |
+
"eval_accuracy": 0.7333333333333333,
|
| 153 |
+
"eval_loss": 0.9060508012771606,
|
| 154 |
+
"eval_runtime": 1.5628,
|
| 155 |
+
"eval_samples_per_second": 38.392,
|
| 156 |
+
"eval_steps_per_second": 1.28,
|
| 157 |
+
"step": 85
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 12.0,
|
| 161 |
+
"eval_accuracy": 0.75,
|
| 162 |
+
"eval_loss": 0.8720260858535767,
|
| 163 |
+
"eval_runtime": 1.5352,
|
| 164 |
+
"eval_samples_per_second": 39.084,
|
| 165 |
+
"eval_steps_per_second": 1.303,
|
| 166 |
+
"step": 93
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 12.39,
|
| 170 |
+
"grad_norm": 1.1283667087554932,
|
| 171 |
+
"learning_rate": 3.7500000000000003e-05,
|
| 172 |
+
"loss": 0.109,
|
| 173 |
+
"step": 96
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 12.9,
|
| 177 |
+
"eval_accuracy": 0.7833333333333333,
|
| 178 |
+
"eval_loss": 0.8438921570777893,
|
| 179 |
+
"eval_runtime": 1.549,
|
| 180 |
+
"eval_samples_per_second": 38.735,
|
| 181 |
+
"eval_steps_per_second": 1.291,
|
| 182 |
+
"step": 100
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 13.94,
|
| 186 |
+
"grad_norm": 1.2174146175384521,
|
| 187 |
+
"learning_rate": 3.522727272727273e-05,
|
| 188 |
+
"loss": 0.0902,
|
| 189 |
+
"step": 108
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 13.94,
|
| 193 |
+
"eval_accuracy": 0.8333333333333334,
|
| 194 |
+
"eval_loss": 0.7344614863395691,
|
| 195 |
+
"eval_runtime": 1.5138,
|
| 196 |
+
"eval_samples_per_second": 39.635,
|
| 197 |
+
"eval_steps_per_second": 1.321,
|
| 198 |
+
"step": 108
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 14.97,
|
| 202 |
+
"eval_accuracy": 0.7833333333333333,
|
| 203 |
+
"eval_loss": 0.8419939279556274,
|
| 204 |
+
"eval_runtime": 2.094,
|
| 205 |
+
"eval_samples_per_second": 28.654,
|
| 206 |
+
"eval_steps_per_second": 0.955,
|
| 207 |
+
"step": 116
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 15.48,
|
| 211 |
+
"grad_norm": 2.3597609996795654,
|
| 212 |
+
"learning_rate": 3.295454545454545e-05,
|
| 213 |
+
"loss": 0.0938,
|
| 214 |
+
"step": 120
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"epoch": 16.0,
|
| 218 |
+
"eval_accuracy": 0.75,
|
| 219 |
+
"eval_loss": 0.799355149269104,
|
| 220 |
+
"eval_runtime": 1.6114,
|
| 221 |
+
"eval_samples_per_second": 37.235,
|
| 222 |
+
"eval_steps_per_second": 1.241,
|
| 223 |
+
"step": 124
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 16.9,
|
| 227 |
+
"eval_accuracy": 0.8,
|
| 228 |
+
"eval_loss": 0.834147036075592,
|
| 229 |
+
"eval_runtime": 1.5793,
|
| 230 |
+
"eval_samples_per_second": 37.991,
|
| 231 |
+
"eval_steps_per_second": 1.266,
|
| 232 |
+
"step": 131
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 17.03,
|
| 236 |
+
"grad_norm": 3.3779869079589844,
|
| 237 |
+
"learning_rate": 3.068181818181818e-05,
|
| 238 |
+
"loss": 0.0862,
|
| 239 |
+
"step": 132
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 17.94,
|
| 243 |
+
"eval_accuracy": 0.8,
|
| 244 |
+
"eval_loss": 0.7238650918006897,
|
| 245 |
+
"eval_runtime": 1.5475,
|
| 246 |
+
"eval_samples_per_second": 38.771,
|
| 247 |
+
"eval_steps_per_second": 1.292,
|
| 248 |
+
"step": 139
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 18.58,
|
| 252 |
+
"grad_norm": 1.2446733713150024,
|
| 253 |
+
"learning_rate": 2.8409090909090912e-05,
|
| 254 |
+
"loss": 0.0864,
|
| 255 |
+
"step": 144
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 18.97,
|
| 259 |
+
"eval_accuracy": 0.7833333333333333,
|
| 260 |
+
"eval_loss": 0.8485052585601807,
|
| 261 |
+
"eval_runtime": 1.5613,
|
| 262 |
+
"eval_samples_per_second": 38.428,
|
| 263 |
+
"eval_steps_per_second": 1.281,
|
| 264 |
+
"step": 147
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 20.0,
|
| 268 |
+
"eval_accuracy": 0.8,
|
| 269 |
+
"eval_loss": 0.8948166370391846,
|
| 270 |
+
"eval_runtime": 1.558,
|
| 271 |
+
"eval_samples_per_second": 38.512,
|
| 272 |
+
"eval_steps_per_second": 1.284,
|
| 273 |
+
"step": 155
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 20.13,
|
| 277 |
+
"grad_norm": 1.548230767250061,
|
| 278 |
+
"learning_rate": 2.6136363636363637e-05,
|
| 279 |
+
"loss": 0.065,
|
| 280 |
+
"step": 156
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 20.9,
|
| 284 |
+
"eval_accuracy": 0.8166666666666667,
|
| 285 |
+
"eval_loss": 0.8681192994117737,
|
| 286 |
+
"eval_runtime": 1.5486,
|
| 287 |
+
"eval_samples_per_second": 38.745,
|
| 288 |
+
"eval_steps_per_second": 1.292,
|
| 289 |
+
"step": 162
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 21.68,
|
| 293 |
+
"grad_norm": 1.3194923400878906,
|
| 294 |
+
"learning_rate": 2.3863636363636365e-05,
|
| 295 |
+
"loss": 0.0793,
|
| 296 |
+
"step": 168
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 21.94,
|
| 300 |
+
"eval_accuracy": 0.8166666666666667,
|
| 301 |
+
"eval_loss": 0.8225926756858826,
|
| 302 |
+
"eval_runtime": 1.5256,
|
| 303 |
+
"eval_samples_per_second": 39.329,
|
| 304 |
+
"eval_steps_per_second": 1.311,
|
| 305 |
+
"step": 170
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 22.97,
|
| 309 |
+
"eval_accuracy": 0.8333333333333334,
|
| 310 |
+
"eval_loss": 0.7495377063751221,
|
| 311 |
+
"eval_runtime": 1.5935,
|
| 312 |
+
"eval_samples_per_second": 37.653,
|
| 313 |
+
"eval_steps_per_second": 1.255,
|
| 314 |
+
"step": 178
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 23.23,
|
| 318 |
+
"grad_norm": 1.1563166379928589,
|
| 319 |
+
"learning_rate": 2.1590909090909093e-05,
|
| 320 |
+
"loss": 0.0629,
|
| 321 |
+
"step": 180
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 24.0,
|
| 325 |
+
"eval_accuracy": 0.7666666666666667,
|
| 326 |
+
"eval_loss": 0.8813876509666443,
|
| 327 |
+
"eval_runtime": 1.5206,
|
| 328 |
+
"eval_samples_per_second": 39.458,
|
| 329 |
+
"eval_steps_per_second": 1.315,
|
| 330 |
+
"step": 186
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"epoch": 24.77,
|
| 334 |
+
"grad_norm": 1.163840413093567,
|
| 335 |
+
"learning_rate": 1.9318181818181818e-05,
|
| 336 |
+
"loss": 0.0666,
|
| 337 |
+
"step": 192
|
| 338 |
+
},
|
| 339 |
+
{
|
| 340 |
+
"epoch": 24.9,
|
| 341 |
+
"eval_accuracy": 0.8166666666666667,
|
| 342 |
+
"eval_loss": 0.7739368081092834,
|
| 343 |
+
"eval_runtime": 1.5452,
|
| 344 |
+
"eval_samples_per_second": 38.83,
|
| 345 |
+
"eval_steps_per_second": 1.294,
|
| 346 |
+
"step": 193
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 25.94,
|
| 350 |
+
"eval_accuracy": 0.7833333333333333,
|
| 351 |
+
"eval_loss": 0.9246408939361572,
|
| 352 |
+
"eval_runtime": 1.5689,
|
| 353 |
+
"eval_samples_per_second": 38.244,
|
| 354 |
+
"eval_steps_per_second": 1.275,
|
| 355 |
+
"step": 201
|
| 356 |
+
},
|
| 357 |
+
{
|
| 358 |
+
"epoch": 26.32,
|
| 359 |
+
"grad_norm": 0.9550792574882507,
|
| 360 |
+
"learning_rate": 1.7045454545454546e-05,
|
| 361 |
+
"loss": 0.0571,
|
| 362 |
+
"step": 204
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 26.97,
|
| 366 |
+
"eval_accuracy": 0.8333333333333334,
|
| 367 |
+
"eval_loss": 0.8077472448348999,
|
| 368 |
+
"eval_runtime": 1.5388,
|
| 369 |
+
"eval_samples_per_second": 38.992,
|
| 370 |
+
"eval_steps_per_second": 1.3,
|
| 371 |
+
"step": 209
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 27.87,
|
| 375 |
+
"grad_norm": 1.725164532661438,
|
| 376 |
+
"learning_rate": 1.4772727272727274e-05,
|
| 377 |
+
"loss": 0.0519,
|
| 378 |
+
"step": 216
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 28.0,
|
| 382 |
+
"eval_accuracy": 0.7833333333333333,
|
| 383 |
+
"eval_loss": 0.8974602818489075,
|
| 384 |
+
"eval_runtime": 1.5489,
|
| 385 |
+
"eval_samples_per_second": 38.738,
|
| 386 |
+
"eval_steps_per_second": 1.291,
|
| 387 |
+
"step": 217
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 28.9,
|
| 391 |
+
"eval_accuracy": 0.7833333333333333,
|
| 392 |
+
"eval_loss": 0.9199196100234985,
|
| 393 |
+
"eval_runtime": 1.5274,
|
| 394 |
+
"eval_samples_per_second": 39.283,
|
| 395 |
+
"eval_steps_per_second": 1.309,
|
| 396 |
+
"step": 224
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 29.42,
|
| 400 |
+
"grad_norm": 0.8022506833076477,
|
| 401 |
+
"learning_rate": 1.25e-05,
|
| 402 |
+
"loss": 0.0523,
|
| 403 |
+
"step": 228
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 29.94,
|
| 407 |
+
"eval_accuracy": 0.8,
|
| 408 |
+
"eval_loss": 0.8512372374534607,
|
| 409 |
+
"eval_runtime": 1.5327,
|
| 410 |
+
"eval_samples_per_second": 39.145,
|
| 411 |
+
"eval_steps_per_second": 1.305,
|
| 412 |
+
"step": 232
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 30.97,
|
| 416 |
+
"grad_norm": 1.4496326446533203,
|
| 417 |
+
"learning_rate": 1.0227272727272729e-05,
|
| 418 |
+
"loss": 0.0548,
|
| 419 |
+
"step": 240
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 30.97,
|
| 423 |
+
"eval_accuracy": 0.8166666666666667,
|
| 424 |
+
"eval_loss": 0.9376980066299438,
|
| 425 |
+
"eval_runtime": 1.5279,
|
| 426 |
+
"eval_samples_per_second": 39.268,
|
| 427 |
+
"eval_steps_per_second": 1.309,
|
| 428 |
+
"step": 240
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 32.0,
|
| 432 |
+
"eval_accuracy": 0.8166666666666667,
|
| 433 |
+
"eval_loss": 0.8212655782699585,
|
| 434 |
+
"eval_runtime": 1.5365,
|
| 435 |
+
"eval_samples_per_second": 39.049,
|
| 436 |
+
"eval_steps_per_second": 1.302,
|
| 437 |
+
"step": 248
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 32.52,
|
| 441 |
+
"grad_norm": 2.0979461669921875,
|
| 442 |
+
"learning_rate": 7.954545454545455e-06,
|
| 443 |
+
"loss": 0.0576,
|
| 444 |
+
"step": 252
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 32.9,
|
| 448 |
+
"eval_accuracy": 0.8166666666666667,
|
| 449 |
+
"eval_loss": 0.8384222388267517,
|
| 450 |
+
"eval_runtime": 1.5251,
|
| 451 |
+
"eval_samples_per_second": 39.342,
|
| 452 |
+
"eval_steps_per_second": 1.311,
|
| 453 |
+
"step": 255
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 33.94,
|
| 457 |
+
"eval_accuracy": 0.8,
|
| 458 |
+
"eval_loss": 0.866399347782135,
|
| 459 |
+
"eval_runtime": 1.5559,
|
| 460 |
+
"eval_samples_per_second": 38.564,
|
| 461 |
+
"eval_steps_per_second": 1.285,
|
| 462 |
+
"step": 263
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 34.06,
|
| 466 |
+
"grad_norm": 1.5001689195632935,
|
| 467 |
+
"learning_rate": 5.681818181818182e-06,
|
| 468 |
+
"loss": 0.0381,
|
| 469 |
+
"step": 264
|
| 470 |
+
},
|
| 471 |
+
{
|
| 472 |
+
"epoch": 34.97,
|
| 473 |
+
"eval_accuracy": 0.8,
|
| 474 |
+
"eval_loss": 0.8817654252052307,
|
| 475 |
+
"eval_runtime": 1.5241,
|
| 476 |
+
"eval_samples_per_second": 39.366,
|
| 477 |
+
"eval_steps_per_second": 1.312,
|
| 478 |
+
"step": 271
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 35.61,
|
| 482 |
+
"grad_norm": 0.628817617893219,
|
| 483 |
+
"learning_rate": 3.409090909090909e-06,
|
| 484 |
+
"loss": 0.0338,
|
| 485 |
+
"step": 276
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 36.0,
|
| 489 |
+
"eval_accuracy": 0.7833333333333333,
|
| 490 |
+
"eval_loss": 0.9105542302131653,
|
| 491 |
+
"eval_runtime": 1.5534,
|
| 492 |
+
"eval_samples_per_second": 38.625,
|
| 493 |
+
"eval_steps_per_second": 1.287,
|
| 494 |
+
"step": 279
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 36.9,
|
| 498 |
+
"eval_accuracy": 0.7833333333333333,
|
| 499 |
+
"eval_loss": 0.9056997299194336,
|
| 500 |
+
"eval_runtime": 1.5384,
|
| 501 |
+
"eval_samples_per_second": 39.002,
|
| 502 |
+
"eval_steps_per_second": 1.3,
|
| 503 |
+
"step": 286
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 37.16,
|
| 507 |
+
"grad_norm": 0.9884141683578491,
|
| 508 |
+
"learning_rate": 1.1363636363636364e-06,
|
| 509 |
+
"loss": 0.0443,
|
| 510 |
+
"step": 288
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 37.94,
|
| 514 |
+
"eval_accuracy": 0.7833333333333333,
|
| 515 |
+
"eval_loss": 0.9011732339859009,
|
| 516 |
+
"eval_runtime": 1.6397,
|
| 517 |
+
"eval_samples_per_second": 36.593,
|
| 518 |
+
"eval_steps_per_second": 1.22,
|
| 519 |
+
"step": 294
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 37.94,
|
| 523 |
+
"step": 294,
|
| 524 |
+
"total_flos": 2.864620236542755e+18,
|
| 525 |
+
"train_loss": 0.20059432347818296,
|
| 526 |
+
"train_runtime": 1595.666,
|
| 527 |
+
"train_samples_per_second": 25.637,
|
| 528 |
+
"train_steps_per_second": 0.184
|
| 529 |
+
}
|
| 530 |
+
],
|
| 531 |
+
"logging_steps": 12,
|
| 532 |
+
"max_steps": 294,
|
| 533 |
+
"num_input_tokens_seen": 0,
|
| 534 |
+
"num_train_epochs": 42,
|
| 535 |
+
"save_steps": 500,
|
| 536 |
+
"total_flos": 2.864620236542755e+18,
|
| 537 |
+
"train_batch_size": 32,
|
| 538 |
+
"trial_name": null,
|
| 539 |
+
"trial_params": null
|
| 540 |
+
}
|