Commit
•
8e1117c
1
Parent(s):
125ef8a
End of training
Browse files- README.md +5 -5
- all_results.json +16 -0
- eval_results.json +11 -0
- train_results.json +8 -0
- trainer_state.json +591 -0
README.md
CHANGED
|
@@ -21,11 +21,11 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 21 |
|
| 22 |
This model is a fine-tuned version of [microsoft/swin-large-patch4-window7-224-in22k](https://huggingface.co/microsoft/swin-large-patch4-window7-224-in22k) on an unknown dataset.
|
| 23 |
It achieves the following results on the evaluation set:
|
| 24 |
-
- Loss: 0.
|
| 25 |
-
- Accuracy: 0.
|
| 26 |
-
- Precision: 0.
|
| 27 |
-
- Recall: 0.
|
| 28 |
-
- F1: 0.
|
| 29 |
|
| 30 |
## Model description
|
| 31 |
|
|
|
|
| 21 |
|
| 22 |
This model is a fine-tuned version of [microsoft/swin-large-patch4-window7-224-in22k](https://huggingface.co/microsoft/swin-large-patch4-window7-224-in22k) on an unknown dataset.
|
| 23 |
It achieves the following results on the evaluation set:
|
| 24 |
+
- Loss: 0.1159
|
| 25 |
+
- Accuracy: 0.9588
|
| 26 |
+
- Precision: 0.9599
|
| 27 |
+
- Recall: 0.9401
|
| 28 |
+
- F1: 0.9492
|
| 29 |
|
| 30 |
## Model description
|
| 31 |
|
all_results.json
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.88,
|
| 3 |
+
"eval_accuracy": 0.9587628865979382,
|
| 4 |
+
"eval_f1": 0.9492478744277306,
|
| 5 |
+
"eval_loss": 0.1159210205078125,
|
| 6 |
+
"eval_precision": 0.9598554797270173,
|
| 7 |
+
"eval_recall": 0.9400691509796388,
|
| 8 |
+
"eval_runtime": 9.465,
|
| 9 |
+
"eval_samples_per_second": 61.49,
|
| 10 |
+
"eval_steps_per_second": 3.909,
|
| 11 |
+
"total_flos": 7.138406630605308e+18,
|
| 12 |
+
"train_loss": 0.19363035304205758,
|
| 13 |
+
"train_runtime": 1305.2945,
|
| 14 |
+
"train_samples_per_second": 31.234,
|
| 15 |
+
"train_steps_per_second": 0.483
|
| 16 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.88,
|
| 3 |
+
"eval_accuracy": 0.9587628865979382,
|
| 4 |
+
"eval_f1": 0.9492478744277306,
|
| 5 |
+
"eval_loss": 0.1159210205078125,
|
| 6 |
+
"eval_precision": 0.9598554797270173,
|
| 7 |
+
"eval_recall": 0.9400691509796388,
|
| 8 |
+
"eval_runtime": 9.465,
|
| 9 |
+
"eval_samples_per_second": 61.49,
|
| 10 |
+
"eval_steps_per_second": 3.909
|
| 11 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.88,
|
| 3 |
+
"total_flos": 7.138406630605308e+18,
|
| 4 |
+
"train_loss": 0.19363035304205758,
|
| 5 |
+
"train_runtime": 1305.2945,
|
| 6 |
+
"train_samples_per_second": 31.234,
|
| 7 |
+
"train_steps_per_second": 0.483
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,591 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.9793991416309012,
|
| 3 |
+
"best_model_checkpoint": "swin-large-patch4-window7-224-in22k-finetuned-lora-medmnistv2/checkpoint-255",
|
| 4 |
+
"epoch": 9.882352941176471,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 630,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.16,
|
| 13 |
+
"grad_norm": 0.9944203495979309,
|
| 14 |
+
"learning_rate": 0.004920634920634921,
|
| 15 |
+
"loss": 0.6411,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.31,
|
| 20 |
+
"grad_norm": 1.986366868019104,
|
| 21 |
+
"learning_rate": 0.004841269841269842,
|
| 22 |
+
"loss": 0.3966,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.47,
|
| 27 |
+
"grad_norm": 1.2478023767471313,
|
| 28 |
+
"learning_rate": 0.0047619047619047615,
|
| 29 |
+
"loss": 0.2835,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.63,
|
| 34 |
+
"grad_norm": 0.9372655749320984,
|
| 35 |
+
"learning_rate": 0.004682539682539683,
|
| 36 |
+
"loss": 0.2194,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.78,
|
| 41 |
+
"grad_norm": 0.834052562713623,
|
| 42 |
+
"learning_rate": 0.004603174603174603,
|
| 43 |
+
"loss": 0.4268,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.94,
|
| 48 |
+
"grad_norm": 0.8572702407836914,
|
| 49 |
+
"learning_rate": 0.004523809523809524,
|
| 50 |
+
"loss": 0.3305,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.99,
|
| 55 |
+
"eval_accuracy": 0.936480686695279,
|
| 56 |
+
"eval_f1": 0.9119407558733402,
|
| 57 |
+
"eval_loss": 0.16002865135669708,
|
| 58 |
+
"eval_precision": 0.9478312410980476,
|
| 59 |
+
"eval_recall": 0.886801607677731,
|
| 60 |
+
"eval_runtime": 18.303,
|
| 61 |
+
"eval_samples_per_second": 63.651,
|
| 62 |
+
"eval_steps_per_second": 3.988,
|
| 63 |
+
"step": 63
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.1,
|
| 67 |
+
"grad_norm": 1.092854619026184,
|
| 68 |
+
"learning_rate": 0.0044444444444444444,
|
| 69 |
+
"loss": 0.2775,
|
| 70 |
+
"step": 70
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.25,
|
| 74 |
+
"grad_norm": 1.632826328277588,
|
| 75 |
+
"learning_rate": 0.004365079365079365,
|
| 76 |
+
"loss": 0.2432,
|
| 77 |
+
"step": 80
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.41,
|
| 81 |
+
"grad_norm": 0.9421872496604919,
|
| 82 |
+
"learning_rate": 0.004285714285714286,
|
| 83 |
+
"loss": 0.2564,
|
| 84 |
+
"step": 90
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.57,
|
| 88 |
+
"grad_norm": 3.025440216064453,
|
| 89 |
+
"learning_rate": 0.004206349206349207,
|
| 90 |
+
"loss": 0.2726,
|
| 91 |
+
"step": 100
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.73,
|
| 95 |
+
"grad_norm": 0.8582188487052917,
|
| 96 |
+
"learning_rate": 0.0041269841269841265,
|
| 97 |
+
"loss": 0.2958,
|
| 98 |
+
"step": 110
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.88,
|
| 102 |
+
"grad_norm": 1.093877911567688,
|
| 103 |
+
"learning_rate": 0.004047619047619048,
|
| 104 |
+
"loss": 0.2335,
|
| 105 |
+
"step": 120
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.99,
|
| 109 |
+
"eval_accuracy": 0.9313304721030042,
|
| 110 |
+
"eval_f1": 0.9166446055107573,
|
| 111 |
+
"eval_loss": 0.15517598390579224,
|
| 112 |
+
"eval_precision": 0.8967545322648443,
|
| 113 |
+
"eval_recall": 0.9471582920716426,
|
| 114 |
+
"eval_runtime": 18.7183,
|
| 115 |
+
"eval_samples_per_second": 62.239,
|
| 116 |
+
"eval_steps_per_second": 3.9,
|
| 117 |
+
"step": 127
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 2.04,
|
| 121 |
+
"grad_norm": 1.588196873664856,
|
| 122 |
+
"learning_rate": 0.003968253968253968,
|
| 123 |
+
"loss": 0.1984,
|
| 124 |
+
"step": 130
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 2.2,
|
| 128 |
+
"grad_norm": 0.8592113256454468,
|
| 129 |
+
"learning_rate": 0.003896825396825397,
|
| 130 |
+
"loss": 0.2741,
|
| 131 |
+
"step": 140
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 2.35,
|
| 135 |
+
"grad_norm": 1.1290347576141357,
|
| 136 |
+
"learning_rate": 0.003817460317460317,
|
| 137 |
+
"loss": 0.2253,
|
| 138 |
+
"step": 150
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 2.51,
|
| 142 |
+
"grad_norm": 0.8879281878471375,
|
| 143 |
+
"learning_rate": 0.0037380952380952383,
|
| 144 |
+
"loss": 0.2726,
|
| 145 |
+
"step": 160
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 2.67,
|
| 149 |
+
"grad_norm": 0.8666319847106934,
|
| 150 |
+
"learning_rate": 0.0036587301587301586,
|
| 151 |
+
"loss": 0.2226,
|
| 152 |
+
"step": 170
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 2.82,
|
| 156 |
+
"grad_norm": 1.2298210859298706,
|
| 157 |
+
"learning_rate": 0.0035793650793650793,
|
| 158 |
+
"loss": 0.1967,
|
| 159 |
+
"step": 180
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 2.98,
|
| 163 |
+
"grad_norm": 0.7905568480491638,
|
| 164 |
+
"learning_rate": 0.0034999999999999996,
|
| 165 |
+
"loss": 0.1977,
|
| 166 |
+
"step": 190
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 3.0,
|
| 170 |
+
"eval_accuracy": 0.9733905579399141,
|
| 171 |
+
"eval_f1": 0.9659005824707372,
|
| 172 |
+
"eval_loss": 0.08546662330627441,
|
| 173 |
+
"eval_precision": 0.9607843137254902,
|
| 174 |
+
"eval_recall": 0.9713575096277278,
|
| 175 |
+
"eval_runtime": 18.7005,
|
| 176 |
+
"eval_samples_per_second": 62.298,
|
| 177 |
+
"eval_steps_per_second": 3.904,
|
| 178 |
+
"step": 191
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 3.14,
|
| 182 |
+
"grad_norm": 0.9821394681930542,
|
| 183 |
+
"learning_rate": 0.003420634920634921,
|
| 184 |
+
"loss": 0.2158,
|
| 185 |
+
"step": 200
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 3.29,
|
| 189 |
+
"grad_norm": 0.6182402968406677,
|
| 190 |
+
"learning_rate": 0.003341269841269841,
|
| 191 |
+
"loss": 0.1528,
|
| 192 |
+
"step": 210
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 3.45,
|
| 196 |
+
"grad_norm": 0.6537771821022034,
|
| 197 |
+
"learning_rate": 0.003261904761904762,
|
| 198 |
+
"loss": 0.2038,
|
| 199 |
+
"step": 220
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 3.61,
|
| 203 |
+
"grad_norm": 1.0976545810699463,
|
| 204 |
+
"learning_rate": 0.0031825396825396826,
|
| 205 |
+
"loss": 0.217,
|
| 206 |
+
"step": 230
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 3.76,
|
| 210 |
+
"grad_norm": 0.581605076789856,
|
| 211 |
+
"learning_rate": 0.0031031746031746034,
|
| 212 |
+
"loss": 0.2044,
|
| 213 |
+
"step": 240
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 3.92,
|
| 217 |
+
"grad_norm": 0.8219888806343079,
|
| 218 |
+
"learning_rate": 0.0030238095238095237,
|
| 219 |
+
"loss": 0.1746,
|
| 220 |
+
"step": 250
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 4.0,
|
| 224 |
+
"eval_accuracy": 0.9793991416309012,
|
| 225 |
+
"eval_f1": 0.972945356587167,
|
| 226 |
+
"eval_loss": 0.0870010182261467,
|
| 227 |
+
"eval_precision": 0.9794167490467448,
|
| 228 |
+
"eval_recall": 0.9669104162846139,
|
| 229 |
+
"eval_runtime": 18.7359,
|
| 230 |
+
"eval_samples_per_second": 62.18,
|
| 231 |
+
"eval_steps_per_second": 3.896,
|
| 232 |
+
"step": 255
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 4.08,
|
| 236 |
+
"grad_norm": 0.45729899406433105,
|
| 237 |
+
"learning_rate": 0.0029444444444444444,
|
| 238 |
+
"loss": 0.1785,
|
| 239 |
+
"step": 260
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 4.24,
|
| 243 |
+
"grad_norm": 0.9842467904090881,
|
| 244 |
+
"learning_rate": 0.002865079365079365,
|
| 245 |
+
"loss": 0.1507,
|
| 246 |
+
"step": 270
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.39,
|
| 250 |
+
"grad_norm": 2.326084852218628,
|
| 251 |
+
"learning_rate": 0.002785714285714286,
|
| 252 |
+
"loss": 0.2188,
|
| 253 |
+
"step": 280
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 4.55,
|
| 257 |
+
"grad_norm": 0.7227071523666382,
|
| 258 |
+
"learning_rate": 0.002706349206349206,
|
| 259 |
+
"loss": 0.1853,
|
| 260 |
+
"step": 290
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 4.71,
|
| 264 |
+
"grad_norm": 1.1356016397476196,
|
| 265 |
+
"learning_rate": 0.002626984126984127,
|
| 266 |
+
"loss": 0.1671,
|
| 267 |
+
"step": 300
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 4.86,
|
| 271 |
+
"grad_norm": 0.6211657524108887,
|
| 272 |
+
"learning_rate": 0.0025476190476190477,
|
| 273 |
+
"loss": 0.1797,
|
| 274 |
+
"step": 310
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 4.99,
|
| 278 |
+
"eval_accuracy": 0.9699570815450643,
|
| 279 |
+
"eval_f1": 0.961658208434536,
|
| 280 |
+
"eval_loss": 0.0828637108206749,
|
| 281 |
+
"eval_precision": 0.95490488153731,
|
| 282 |
+
"eval_recall": 0.969034629256067,
|
| 283 |
+
"eval_runtime": 18.2407,
|
| 284 |
+
"eval_samples_per_second": 63.868,
|
| 285 |
+
"eval_steps_per_second": 4.002,
|
| 286 |
+
"step": 318
|
| 287 |
+
},
|
| 288 |
+
{
|
| 289 |
+
"epoch": 5.02,
|
| 290 |
+
"grad_norm": 0.3646533787250519,
|
| 291 |
+
"learning_rate": 0.0024682539682539684,
|
| 292 |
+
"loss": 0.1636,
|
| 293 |
+
"step": 320
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 5.18,
|
| 297 |
+
"grad_norm": 0.5121908187866211,
|
| 298 |
+
"learning_rate": 0.002388888888888889,
|
| 299 |
+
"loss": 0.1463,
|
| 300 |
+
"step": 330
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"epoch": 5.33,
|
| 304 |
+
"grad_norm": 0.34603381156921387,
|
| 305 |
+
"learning_rate": 0.0023095238095238095,
|
| 306 |
+
"loss": 0.1284,
|
| 307 |
+
"step": 340
|
| 308 |
+
},
|
| 309 |
+
{
|
| 310 |
+
"epoch": 5.49,
|
| 311 |
+
"grad_norm": 1.5245040655136108,
|
| 312 |
+
"learning_rate": 0.0022301587301587302,
|
| 313 |
+
"loss": 0.1307,
|
| 314 |
+
"step": 350
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 5.65,
|
| 318 |
+
"grad_norm": 1.0758203268051147,
|
| 319 |
+
"learning_rate": 0.002150793650793651,
|
| 320 |
+
"loss": 0.1912,
|
| 321 |
+
"step": 360
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 5.8,
|
| 325 |
+
"grad_norm": 0.9437044858932495,
|
| 326 |
+
"learning_rate": 0.0020714285714285717,
|
| 327 |
+
"loss": 0.1942,
|
| 328 |
+
"step": 370
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 5.96,
|
| 332 |
+
"grad_norm": 0.6630730628967285,
|
| 333 |
+
"learning_rate": 0.001992063492063492,
|
| 334 |
+
"loss": 0.1436,
|
| 335 |
+
"step": 380
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 5.99,
|
| 339 |
+
"eval_accuracy": 0.9708154506437768,
|
| 340 |
+
"eval_f1": 0.9627915802446869,
|
| 341 |
+
"eval_loss": 0.0797128826379776,
|
| 342 |
+
"eval_precision": 0.9556120562130177,
|
| 343 |
+
"eval_recall": 0.9706793660981723,
|
| 344 |
+
"eval_runtime": 18.3826,
|
| 345 |
+
"eval_samples_per_second": 63.375,
|
| 346 |
+
"eval_steps_per_second": 3.971,
|
| 347 |
+
"step": 382
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 6.12,
|
| 351 |
+
"grad_norm": 0.5699043869972229,
|
| 352 |
+
"learning_rate": 0.0019126984126984126,
|
| 353 |
+
"loss": 0.1704,
|
| 354 |
+
"step": 390
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 6.27,
|
| 358 |
+
"grad_norm": 0.6287882328033447,
|
| 359 |
+
"learning_rate": 0.0018333333333333333,
|
| 360 |
+
"loss": 0.1034,
|
| 361 |
+
"step": 400
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 6.43,
|
| 365 |
+
"grad_norm": 0.4099302589893341,
|
| 366 |
+
"learning_rate": 0.0017539682539682538,
|
| 367 |
+
"loss": 0.1633,
|
| 368 |
+
"step": 410
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 6.59,
|
| 372 |
+
"grad_norm": 0.24668000638484955,
|
| 373 |
+
"learning_rate": 0.0016746031746031746,
|
| 374 |
+
"loss": 0.1477,
|
| 375 |
+
"step": 420
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 6.75,
|
| 379 |
+
"grad_norm": 0.6959215998649597,
|
| 380 |
+
"learning_rate": 0.001595238095238095,
|
| 381 |
+
"loss": 0.1578,
|
| 382 |
+
"step": 430
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 6.9,
|
| 386 |
+
"grad_norm": 0.4429934620857239,
|
| 387 |
+
"learning_rate": 0.0015158730158730158,
|
| 388 |
+
"loss": 0.1632,
|
| 389 |
+
"step": 440
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 7.0,
|
| 393 |
+
"eval_accuracy": 0.9699570815450643,
|
| 394 |
+
"eval_f1": 0.9621164611060785,
|
| 395 |
+
"eval_loss": 0.08162170648574829,
|
| 396 |
+
"eval_precision": 0.9507741239032943,
|
| 397 |
+
"eval_recall": 0.9754187297512072,
|
| 398 |
+
"eval_runtime": 18.5501,
|
| 399 |
+
"eval_samples_per_second": 62.803,
|
| 400 |
+
"eval_steps_per_second": 3.935,
|
| 401 |
+
"step": 446
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 7.06,
|
| 405 |
+
"grad_norm": 0.3358498513698578,
|
| 406 |
+
"learning_rate": 0.0014365079365079364,
|
| 407 |
+
"loss": 0.1628,
|
| 408 |
+
"step": 450
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 7.22,
|
| 412 |
+
"grad_norm": 1.001493215560913,
|
| 413 |
+
"learning_rate": 0.0013571428571428571,
|
| 414 |
+
"loss": 0.1495,
|
| 415 |
+
"step": 460
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 7.37,
|
| 419 |
+
"grad_norm": 1.0969635248184204,
|
| 420 |
+
"learning_rate": 0.0012777777777777776,
|
| 421 |
+
"loss": 0.1972,
|
| 422 |
+
"step": 470
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 7.53,
|
| 426 |
+
"grad_norm": 0.4329684376716614,
|
| 427 |
+
"learning_rate": 0.0011984126984126984,
|
| 428 |
+
"loss": 0.1538,
|
| 429 |
+
"step": 480
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 7.69,
|
| 433 |
+
"grad_norm": 0.426087886095047,
|
| 434 |
+
"learning_rate": 0.0011190476190476191,
|
| 435 |
+
"loss": 0.1291,
|
| 436 |
+
"step": 490
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 7.84,
|
| 440 |
+
"grad_norm": 0.24835826456546783,
|
| 441 |
+
"learning_rate": 0.0010396825396825396,
|
| 442 |
+
"loss": 0.1301,
|
| 443 |
+
"step": 500
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 8.0,
|
| 447 |
+
"grad_norm": 0.3314363956451416,
|
| 448 |
+
"learning_rate": 0.0009603174603174604,
|
| 449 |
+
"loss": 0.1125,
|
| 450 |
+
"step": 510
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 8.0,
|
| 454 |
+
"eval_accuracy": 0.9613733905579399,
|
| 455 |
+
"eval_f1": 0.9519463889441405,
|
| 456 |
+
"eval_loss": 0.10071194916963577,
|
| 457 |
+
"eval_precision": 0.9364616472251679,
|
| 458 |
+
"eval_recall": 0.9717395623204352,
|
| 459 |
+
"eval_runtime": 18.4952,
|
| 460 |
+
"eval_samples_per_second": 62.989,
|
| 461 |
+
"eval_steps_per_second": 3.947,
|
| 462 |
+
"step": 510
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 8.16,
|
| 466 |
+
"grad_norm": 0.4465714693069458,
|
| 467 |
+
"learning_rate": 0.000880952380952381,
|
| 468 |
+
"loss": 0.1279,
|
| 469 |
+
"step": 520
|
| 470 |
+
},
|
| 471 |
+
{
|
| 472 |
+
"epoch": 8.31,
|
| 473 |
+
"grad_norm": 0.7344756126403809,
|
| 474 |
+
"learning_rate": 0.0008015873015873017,
|
| 475 |
+
"loss": 0.1408,
|
| 476 |
+
"step": 530
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 8.47,
|
| 480 |
+
"grad_norm": 2.2782132625579834,
|
| 481 |
+
"learning_rate": 0.0007222222222222222,
|
| 482 |
+
"loss": 0.134,
|
| 483 |
+
"step": 540
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 8.63,
|
| 487 |
+
"grad_norm": 0.8408999443054199,
|
| 488 |
+
"learning_rate": 0.0006428571428571428,
|
| 489 |
+
"loss": 0.1193,
|
| 490 |
+
"step": 550
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 8.78,
|
| 494 |
+
"grad_norm": 0.4357975721359253,
|
| 495 |
+
"learning_rate": 0.0005634920634920636,
|
| 496 |
+
"loss": 0.114,
|
| 497 |
+
"step": 560
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 8.94,
|
| 501 |
+
"grad_norm": 0.6790524125099182,
|
| 502 |
+
"learning_rate": 0.00048412698412698415,
|
| 503 |
+
"loss": 0.1076,
|
| 504 |
+
"step": 570
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.99,
|
| 508 |
+
"eval_accuracy": 0.9690987124463519,
|
| 509 |
+
"eval_f1": 0.9612241124260354,
|
| 510 |
+
"eval_loss": 0.09004171937704086,
|
| 511 |
+
"eval_precision": 0.9481984892871503,
|
| 512 |
+
"eval_recall": 0.9769660431566722,
|
| 513 |
+
"eval_runtime": 18.3674,
|
| 514 |
+
"eval_samples_per_second": 63.427,
|
| 515 |
+
"eval_steps_per_second": 3.974,
|
| 516 |
+
"step": 573
|
| 517 |
+
},
|
| 518 |
+
{
|
| 519 |
+
"epoch": 9.1,
|
| 520 |
+
"grad_norm": 0.38498663902282715,
|
| 521 |
+
"learning_rate": 0.0004047619047619048,
|
| 522 |
+
"loss": 0.1174,
|
| 523 |
+
"step": 580
|
| 524 |
+
},
|
| 525 |
+
{
|
| 526 |
+
"epoch": 9.25,
|
| 527 |
+
"grad_norm": 0.7055544853210449,
|
| 528 |
+
"learning_rate": 0.0003253968253968254,
|
| 529 |
+
"loss": 0.0951,
|
| 530 |
+
"step": 590
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 9.41,
|
| 534 |
+
"grad_norm": 0.8157325387001038,
|
| 535 |
+
"learning_rate": 0.00024603174603174605,
|
| 536 |
+
"loss": 0.1397,
|
| 537 |
+
"step": 600
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 9.57,
|
| 541 |
+
"grad_norm": 0.6511685252189636,
|
| 542 |
+
"learning_rate": 0.00016666666666666666,
|
| 543 |
+
"loss": 0.1296,
|
| 544 |
+
"step": 610
|
| 545 |
+
},
|
| 546 |
+
{
|
| 547 |
+
"epoch": 9.73,
|
| 548 |
+
"grad_norm": 0.5405935049057007,
|
| 549 |
+
"learning_rate": 8.730158730158731e-05,
|
| 550 |
+
"loss": 0.1028,
|
| 551 |
+
"step": 620
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 9.88,
|
| 555 |
+
"grad_norm": 0.472501665353775,
|
| 556 |
+
"learning_rate": 7.936507936507936e-06,
|
| 557 |
+
"loss": 0.1188,
|
| 558 |
+
"step": 630
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"epoch": 9.88,
|
| 562 |
+
"eval_accuracy": 0.9622317596566523,
|
| 563 |
+
"eval_f1": 0.9529697156530292,
|
| 564 |
+
"eval_loss": 0.10635051876306534,
|
| 565 |
+
"eval_precision": 0.9377354615755036,
|
| 566 |
+
"eval_recall": 0.9723202824133504,
|
| 567 |
+
"eval_runtime": 18.4162,
|
| 568 |
+
"eval_samples_per_second": 63.26,
|
| 569 |
+
"eval_steps_per_second": 3.964,
|
| 570 |
+
"step": 630
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 9.88,
|
| 574 |
+
"step": 630,
|
| 575 |
+
"total_flos": 7.138406630605308e+18,
|
| 576 |
+
"train_loss": 0.19363035304205758,
|
| 577 |
+
"train_runtime": 1305.2945,
|
| 578 |
+
"train_samples_per_second": 31.234,
|
| 579 |
+
"train_steps_per_second": 0.483
|
| 580 |
+
}
|
| 581 |
+
],
|
| 582 |
+
"logging_steps": 10,
|
| 583 |
+
"max_steps": 630,
|
| 584 |
+
"num_input_tokens_seen": 0,
|
| 585 |
+
"num_train_epochs": 10,
|
| 586 |
+
"save_steps": 500,
|
| 587 |
+
"total_flos": 7.138406630605308e+18,
|
| 588 |
+
"train_batch_size": 16,
|
| 589 |
+
"trial_name": null,
|
| 590 |
+
"trial_params": null
|
| 591 |
+
}
|