Commit
•
362a524
1
Parent(s):
0bd219a
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 +673 -0
README.md
CHANGED
|
@@ -23,11 +23,11 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 23 |
|
| 24 |
This model is a fine-tuned version of [microsoft/beit-base-patch16-224-pt22k-ft22k](https://huggingface.co/microsoft/beit-base-patch16-224-pt22k-ft22k) on the medmnist-v2 dataset.
|
| 25 |
It achieves the following results on the evaluation set:
|
| 26 |
-
- Loss: 0.
|
| 27 |
-
- Accuracy: 0.
|
| 28 |
-
- Precision: 0.
|
| 29 |
-
- Recall: 0.
|
| 30 |
-
- F1: 0.
|
| 31 |
|
| 32 |
## Model description
|
| 33 |
|
|
|
|
| 23 |
|
| 24 |
This model is a fine-tuned version of [microsoft/beit-base-patch16-224-pt22k-ft22k](https://huggingface.co/microsoft/beit-base-patch16-224-pt22k-ft22k) on the medmnist-v2 dataset.
|
| 25 |
It achieves the following results on the evaluation set:
|
| 26 |
+
- Loss: 0.3960
|
| 27 |
+
- Accuracy: 0.8446
|
| 28 |
+
- Precision: 0.8354
|
| 29 |
+
- Recall: 0.8312
|
| 30 |
+
- F1: 0.8332
|
| 31 |
|
| 32 |
## Model description
|
| 33 |
|
all_results.json
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.898305084745763,
|
| 3 |
+
"eval_accuracy": 0.844551282051282,
|
| 4 |
+
"eval_f1": 0.8331674967682543,
|
| 5 |
+
"eval_loss": 0.3960355818271637,
|
| 6 |
+
"eval_precision": 0.8353898733896292,
|
| 7 |
+
"eval_recall": 0.8311965811965811,
|
| 8 |
+
"eval_runtime": 3.6206,
|
| 9 |
+
"eval_samples_per_second": 172.349,
|
| 10 |
+
"eval_steps_per_second": 10.772,
|
| 11 |
+
"total_flos": 3.6354082184735293e+18,
|
| 12 |
+
"train_loss": 0.42946596570210915,
|
| 13 |
+
"train_runtime": 617.6458,
|
| 14 |
+
"train_samples_per_second": 76.225,
|
| 15 |
+
"train_steps_per_second": 1.182
|
| 16 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.898305084745763,
|
| 3 |
+
"eval_accuracy": 0.844551282051282,
|
| 4 |
+
"eval_f1": 0.8331674967682543,
|
| 5 |
+
"eval_loss": 0.3960355818271637,
|
| 6 |
+
"eval_precision": 0.8353898733896292,
|
| 7 |
+
"eval_recall": 0.8311965811965811,
|
| 8 |
+
"eval_runtime": 3.6206,
|
| 9 |
+
"eval_samples_per_second": 172.349,
|
| 10 |
+
"eval_steps_per_second": 10.772
|
| 11 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.898305084745763,
|
| 3 |
+
"total_flos": 3.6354082184735293e+18,
|
| 4 |
+
"train_loss": 0.42946596570210915,
|
| 5 |
+
"train_runtime": 617.6458,
|
| 6 |
+
"train_samples_per_second": 76.225,
|
| 7 |
+
"train_steps_per_second": 1.182
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,673 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.8683206106870229,
|
| 3 |
+
"best_model_checkpoint": "beit-base-patch16-224-pt22k-ft22k-finetuned-lora-medmnistv2/checkpoint-590",
|
| 4 |
+
"epoch": 9.898305084745763,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 730,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.13559322033898305,
|
| 13 |
+
"grad_norm": 12.995793342590332,
|
| 14 |
+
"learning_rate": 0.004945205479452055,
|
| 15 |
+
"loss": 0.7148,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.2711864406779661,
|
| 20 |
+
"grad_norm": 1.2968316078186035,
|
| 21 |
+
"learning_rate": 0.004876712328767123,
|
| 22 |
+
"loss": 0.612,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.4067796610169492,
|
| 27 |
+
"grad_norm": 1.6618480682373047,
|
| 28 |
+
"learning_rate": 0.004808219178082192,
|
| 29 |
+
"loss": 0.5877,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.5423728813559322,
|
| 34 |
+
"grad_norm": 5.746555328369141,
|
| 35 |
+
"learning_rate": 0.00473972602739726,
|
| 36 |
+
"loss": 0.5768,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.6779661016949152,
|
| 41 |
+
"grad_norm": 1.3945930004119873,
|
| 42 |
+
"learning_rate": 0.0046712328767123295,
|
| 43 |
+
"loss": 0.5698,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.8135593220338984,
|
| 48 |
+
"grad_norm": 6.072067737579346,
|
| 49 |
+
"learning_rate": 0.004602739726027398,
|
| 50 |
+
"loss": 0.5479,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.9491525423728814,
|
| 55 |
+
"grad_norm": 0.36280161142349243,
|
| 56 |
+
"learning_rate": 0.004534246575342466,
|
| 57 |
+
"loss": 0.5947,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.9898305084745763,
|
| 62 |
+
"eval_accuracy": 0.7423664122137404,
|
| 63 |
+
"eval_f1": 0.42606790799561883,
|
| 64 |
+
"eval_loss": 0.5164713263511658,
|
| 65 |
+
"eval_precision": 0.3711832061068702,
|
| 66 |
+
"eval_recall": 0.5,
|
| 67 |
+
"eval_runtime": 2.9766,
|
| 68 |
+
"eval_samples_per_second": 176.038,
|
| 69 |
+
"eval_steps_per_second": 11.086,
|
| 70 |
+
"step": 73
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0847457627118644,
|
| 74 |
+
"grad_norm": 3.3023428916931152,
|
| 75 |
+
"learning_rate": 0.004465753424657534,
|
| 76 |
+
"loss": 0.5075,
|
| 77 |
+
"step": 80
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2203389830508475,
|
| 81 |
+
"grad_norm": 4.128161430358887,
|
| 82 |
+
"learning_rate": 0.004397260273972603,
|
| 83 |
+
"loss": 0.641,
|
| 84 |
+
"step": 90
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3559322033898304,
|
| 88 |
+
"grad_norm": 7.110351085662842,
|
| 89 |
+
"learning_rate": 0.004328767123287672,
|
| 90 |
+
"loss": 0.7467,
|
| 91 |
+
"step": 100
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4915254237288136,
|
| 95 |
+
"grad_norm": 0.7752661108970642,
|
| 96 |
+
"learning_rate": 0.00426027397260274,
|
| 97 |
+
"loss": 0.5226,
|
| 98 |
+
"step": 110
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6271186440677967,
|
| 102 |
+
"grad_norm": 0.3319437801837921,
|
| 103 |
+
"learning_rate": 0.004191780821917808,
|
| 104 |
+
"loss": 0.5337,
|
| 105 |
+
"step": 120
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7627118644067796,
|
| 109 |
+
"grad_norm": 1.7883654832839966,
|
| 110 |
+
"learning_rate": 0.004123287671232877,
|
| 111 |
+
"loss": 0.4819,
|
| 112 |
+
"step": 130
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.8983050847457628,
|
| 116 |
+
"grad_norm": 0.5220211148262024,
|
| 117 |
+
"learning_rate": 0.0040547945205479455,
|
| 118 |
+
"loss": 0.4888,
|
| 119 |
+
"step": 140
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.993220338983051,
|
| 123 |
+
"eval_accuracy": 0.8568702290076335,
|
| 124 |
+
"eval_f1": 0.8151240738562859,
|
| 125 |
+
"eval_loss": 0.3449835777282715,
|
| 126 |
+
"eval_precision": 0.8115699404761905,
|
| 127 |
+
"eval_recall": 0.8189564886222984,
|
| 128 |
+
"eval_runtime": 3.0609,
|
| 129 |
+
"eval_samples_per_second": 171.191,
|
| 130 |
+
"eval_steps_per_second": 10.781,
|
| 131 |
+
"step": 147
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 2.0338983050847457,
|
| 135 |
+
"grad_norm": 5.3052544593811035,
|
| 136 |
+
"learning_rate": 0.003986301369863014,
|
| 137 |
+
"loss": 0.4278,
|
| 138 |
+
"step": 150
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 2.169491525423729,
|
| 142 |
+
"grad_norm": 4.525509357452393,
|
| 143 |
+
"learning_rate": 0.003917808219178082,
|
| 144 |
+
"loss": 0.4448,
|
| 145 |
+
"step": 160
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 2.305084745762712,
|
| 149 |
+
"grad_norm": 6.519794940948486,
|
| 150 |
+
"learning_rate": 0.0038493150684931507,
|
| 151 |
+
"loss": 0.4817,
|
| 152 |
+
"step": 170
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 2.440677966101695,
|
| 156 |
+
"grad_norm": 1.3314294815063477,
|
| 157 |
+
"learning_rate": 0.0037808219178082194,
|
| 158 |
+
"loss": 0.4393,
|
| 159 |
+
"step": 180
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 2.576271186440678,
|
| 163 |
+
"grad_norm": 0.6930696368217468,
|
| 164 |
+
"learning_rate": 0.0037123287671232876,
|
| 165 |
+
"loss": 0.3842,
|
| 166 |
+
"step": 190
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 2.711864406779661,
|
| 170 |
+
"grad_norm": 3.826169729232788,
|
| 171 |
+
"learning_rate": 0.0036438356164383563,
|
| 172 |
+
"loss": 0.4206,
|
| 173 |
+
"step": 200
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 2.847457627118644,
|
| 177 |
+
"grad_norm": 0.5641322135925293,
|
| 178 |
+
"learning_rate": 0.0035753424657534246,
|
| 179 |
+
"loss": 0.3684,
|
| 180 |
+
"step": 210
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 2.983050847457627,
|
| 184 |
+
"grad_norm": 3.910162925720215,
|
| 185 |
+
"learning_rate": 0.0035068493150684932,
|
| 186 |
+
"loss": 0.4022,
|
| 187 |
+
"step": 220
|
| 188 |
+
},
|
| 189 |
+
{
|
| 190 |
+
"epoch": 2.9966101694915253,
|
| 191 |
+
"eval_accuracy": 0.833969465648855,
|
| 192 |
+
"eval_f1": 0.8079106043560151,
|
| 193 |
+
"eval_loss": 0.42247942090034485,
|
| 194 |
+
"eval_precision": 0.7914074166251867,
|
| 195 |
+
"eval_recall": 0.8567361706179186,
|
| 196 |
+
"eval_runtime": 3.2139,
|
| 197 |
+
"eval_samples_per_second": 163.043,
|
| 198 |
+
"eval_steps_per_second": 10.268,
|
| 199 |
+
"step": 221
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 3.1186440677966103,
|
| 203 |
+
"grad_norm": 1.5272403955459595,
|
| 204 |
+
"learning_rate": 0.0034383561643835615,
|
| 205 |
+
"loss": 0.4376,
|
| 206 |
+
"step": 230
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 3.2542372881355934,
|
| 210 |
+
"grad_norm": 0.9074549674987793,
|
| 211 |
+
"learning_rate": 0.00336986301369863,
|
| 212 |
+
"loss": 0.403,
|
| 213 |
+
"step": 240
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 3.389830508474576,
|
| 217 |
+
"grad_norm": 0.9038354158401489,
|
| 218 |
+
"learning_rate": 0.0033013698630136984,
|
| 219 |
+
"loss": 0.3898,
|
| 220 |
+
"step": 250
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 3.5254237288135593,
|
| 224 |
+
"grad_norm": 0.5530111193656921,
|
| 225 |
+
"learning_rate": 0.003232876712328767,
|
| 226 |
+
"loss": 0.3958,
|
| 227 |
+
"step": 260
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 3.6610169491525424,
|
| 231 |
+
"grad_norm": 3.4515538215637207,
|
| 232 |
+
"learning_rate": 0.0031643835616438354,
|
| 233 |
+
"loss": 0.4354,
|
| 234 |
+
"step": 270
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 3.7966101694915255,
|
| 238 |
+
"grad_norm": 5.728641986846924,
|
| 239 |
+
"learning_rate": 0.003095890410958904,
|
| 240 |
+
"loss": 0.4601,
|
| 241 |
+
"step": 280
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 3.9322033898305087,
|
| 245 |
+
"grad_norm": 0.6854091882705688,
|
| 246 |
+
"learning_rate": 0.0030273972602739723,
|
| 247 |
+
"loss": 0.4319,
|
| 248 |
+
"step": 290
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 4.0,
|
| 252 |
+
"eval_accuracy": 0.8587786259541985,
|
| 253 |
+
"eval_f1": 0.8292197382097493,
|
| 254 |
+
"eval_loss": 0.3599797189235687,
|
| 255 |
+
"eval_precision": 0.8122691052465915,
|
| 256 |
+
"eval_recall": 0.8589355422260307,
|
| 257 |
+
"eval_runtime": 2.983,
|
| 258 |
+
"eval_samples_per_second": 175.665,
|
| 259 |
+
"eval_steps_per_second": 11.063,
|
| 260 |
+
"step": 295
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 4.067796610169491,
|
| 264 |
+
"grad_norm": 4.648618221282959,
|
| 265 |
+
"learning_rate": 0.002958904109589041,
|
| 266 |
+
"loss": 0.3716,
|
| 267 |
+
"step": 300
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 4.203389830508475,
|
| 271 |
+
"grad_norm": 2.3614566326141357,
|
| 272 |
+
"learning_rate": 0.0028904109589041093,
|
| 273 |
+
"loss": 0.423,
|
| 274 |
+
"step": 310
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 4.338983050847458,
|
| 278 |
+
"grad_norm": 1.8763720989227295,
|
| 279 |
+
"learning_rate": 0.0028219178082191784,
|
| 280 |
+
"loss": 0.3737,
|
| 281 |
+
"step": 320
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 4.47457627118644,
|
| 285 |
+
"grad_norm": 3.2308576107025146,
|
| 286 |
+
"learning_rate": 0.002753424657534246,
|
| 287 |
+
"loss": 0.4613,
|
| 288 |
+
"step": 330
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 4.610169491525424,
|
| 292 |
+
"grad_norm": 1.1305418014526367,
|
| 293 |
+
"learning_rate": 0.0026849315068493153,
|
| 294 |
+
"loss": 0.377,
|
| 295 |
+
"step": 340
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 4.745762711864407,
|
| 299 |
+
"grad_norm": 0.8450165390968323,
|
| 300 |
+
"learning_rate": 0.0026164383561643836,
|
| 301 |
+
"loss": 0.3981,
|
| 302 |
+
"step": 350
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 4.88135593220339,
|
| 306 |
+
"grad_norm": 0.7082638144493103,
|
| 307 |
+
"learning_rate": 0.0025479452054794523,
|
| 308 |
+
"loss": 0.3836,
|
| 309 |
+
"step": 360
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 4.989830508474577,
|
| 313 |
+
"eval_accuracy": 0.851145038167939,
|
| 314 |
+
"eval_f1": 0.8233035899564225,
|
| 315 |
+
"eval_loss": 0.36647942662239075,
|
| 316 |
+
"eval_precision": 0.8054056666076059,
|
| 317 |
+
"eval_recall": 0.8610492240312292,
|
| 318 |
+
"eval_runtime": 3.005,
|
| 319 |
+
"eval_samples_per_second": 174.375,
|
| 320 |
+
"eval_steps_per_second": 10.982,
|
| 321 |
+
"step": 368
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 5.016949152542373,
|
| 325 |
+
"grad_norm": 0.8164511322975159,
|
| 326 |
+
"learning_rate": 0.0024794520547945205,
|
| 327 |
+
"loss": 0.3491,
|
| 328 |
+
"step": 370
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 5.1525423728813555,
|
| 332 |
+
"grad_norm": 2.7462785243988037,
|
| 333 |
+
"learning_rate": 0.002410958904109589,
|
| 334 |
+
"loss": 0.4279,
|
| 335 |
+
"step": 380
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 5.288135593220339,
|
| 339 |
+
"grad_norm": 0.1708032637834549,
|
| 340 |
+
"learning_rate": 0.0023424657534246575,
|
| 341 |
+
"loss": 0.3868,
|
| 342 |
+
"step": 390
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 5.423728813559322,
|
| 346 |
+
"grad_norm": 0.4240102469921112,
|
| 347 |
+
"learning_rate": 0.002273972602739726,
|
| 348 |
+
"loss": 0.416,
|
| 349 |
+
"step": 400
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 5.559322033898305,
|
| 353 |
+
"grad_norm": 0.3024919331073761,
|
| 354 |
+
"learning_rate": 0.002205479452054795,
|
| 355 |
+
"loss": 0.3745,
|
| 356 |
+
"step": 410
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 5.694915254237288,
|
| 360 |
+
"grad_norm": 0.4930131733417511,
|
| 361 |
+
"learning_rate": 0.002136986301369863,
|
| 362 |
+
"loss": 0.4183,
|
| 363 |
+
"step": 420
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 5.830508474576272,
|
| 367 |
+
"grad_norm": 2.3017737865448,
|
| 368 |
+
"learning_rate": 0.002068493150684932,
|
| 369 |
+
"loss": 0.3791,
|
| 370 |
+
"step": 430
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 5.966101694915254,
|
| 374 |
+
"grad_norm": 1.3229153156280518,
|
| 375 |
+
"learning_rate": 0.002,
|
| 376 |
+
"loss": 0.3887,
|
| 377 |
+
"step": 440
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 5.9932203389830505,
|
| 381 |
+
"eval_accuracy": 0.8645038167938931,
|
| 382 |
+
"eval_f1": 0.8382821499206712,
|
| 383 |
+
"eval_loss": 0.36667701601982117,
|
| 384 |
+
"eval_precision": 0.8196564265766058,
|
| 385 |
+
"eval_recall": 0.8748833666571456,
|
| 386 |
+
"eval_runtime": 3.2223,
|
| 387 |
+
"eval_samples_per_second": 162.616,
|
| 388 |
+
"eval_steps_per_second": 10.241,
|
| 389 |
+
"step": 442
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 6.101694915254237,
|
| 393 |
+
"grad_norm": 1.612464189529419,
|
| 394 |
+
"learning_rate": 0.0019315068493150687,
|
| 395 |
+
"loss": 0.3626,
|
| 396 |
+
"step": 450
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 6.237288135593221,
|
| 400 |
+
"grad_norm": 0.30710288882255554,
|
| 401 |
+
"learning_rate": 0.0018630136986301372,
|
| 402 |
+
"loss": 0.4014,
|
| 403 |
+
"step": 460
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 6.372881355932203,
|
| 407 |
+
"grad_norm": 1.5897902250289917,
|
| 408 |
+
"learning_rate": 0.0017945205479452057,
|
| 409 |
+
"loss": 0.3953,
|
| 410 |
+
"step": 470
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 6.508474576271187,
|
| 414 |
+
"grad_norm": 1.6094199419021606,
|
| 415 |
+
"learning_rate": 0.0017260273972602741,
|
| 416 |
+
"loss": 0.4189,
|
| 417 |
+
"step": 480
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 6.6440677966101696,
|
| 421 |
+
"grad_norm": 1.078356385231018,
|
| 422 |
+
"learning_rate": 0.0016575342465753426,
|
| 423 |
+
"loss": 0.4268,
|
| 424 |
+
"step": 490
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 6.779661016949152,
|
| 428 |
+
"grad_norm": 2.4723339080810547,
|
| 429 |
+
"learning_rate": 0.001589041095890411,
|
| 430 |
+
"loss": 0.4025,
|
| 431 |
+
"step": 500
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 6.915254237288136,
|
| 435 |
+
"grad_norm": 1.7719030380249023,
|
| 436 |
+
"learning_rate": 0.0015205479452054796,
|
| 437 |
+
"loss": 0.3947,
|
| 438 |
+
"step": 510
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 6.996610169491525,
|
| 442 |
+
"eval_accuracy": 0.8530534351145038,
|
| 443 |
+
"eval_f1": 0.8282881157569955,
|
| 444 |
+
"eval_loss": 0.39514508843421936,
|
| 445 |
+
"eval_precision": 0.8098487236054208,
|
| 446 |
+
"eval_recall": 0.8744263543749404,
|
| 447 |
+
"eval_runtime": 2.9797,
|
| 448 |
+
"eval_samples_per_second": 175.856,
|
| 449 |
+
"eval_steps_per_second": 11.075,
|
| 450 |
+
"step": 516
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 7.0508474576271185,
|
| 454 |
+
"grad_norm": 0.9901585578918457,
|
| 455 |
+
"learning_rate": 0.0014520547945205478,
|
| 456 |
+
"loss": 0.364,
|
| 457 |
+
"step": 520
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 7.186440677966102,
|
| 461 |
+
"grad_norm": 1.4664475917816162,
|
| 462 |
+
"learning_rate": 0.0013835616438356163,
|
| 463 |
+
"loss": 0.3738,
|
| 464 |
+
"step": 530
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 7.322033898305085,
|
| 468 |
+
"grad_norm": 1.8693777322769165,
|
| 469 |
+
"learning_rate": 0.0013150684931506848,
|
| 470 |
+
"loss": 0.409,
|
| 471 |
+
"step": 540
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 7.4576271186440675,
|
| 475 |
+
"grad_norm": 1.6105625629425049,
|
| 476 |
+
"learning_rate": 0.0012465753424657534,
|
| 477 |
+
"loss": 0.373,
|
| 478 |
+
"step": 550
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 7.593220338983051,
|
| 482 |
+
"grad_norm": 1.700050950050354,
|
| 483 |
+
"learning_rate": 0.001178082191780822,
|
| 484 |
+
"loss": 0.3899,
|
| 485 |
+
"step": 560
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 7.728813559322034,
|
| 489 |
+
"grad_norm": 0.6655128598213196,
|
| 490 |
+
"learning_rate": 0.0011095890410958904,
|
| 491 |
+
"loss": 0.3573,
|
| 492 |
+
"step": 570
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 7.864406779661017,
|
| 496 |
+
"grad_norm": 2.1642274856567383,
|
| 497 |
+
"learning_rate": 0.0010410958904109589,
|
| 498 |
+
"loss": 0.3828,
|
| 499 |
+
"step": 580
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 8.0,
|
| 503 |
+
"grad_norm": 1.8888564109802246,
|
| 504 |
+
"learning_rate": 0.0009726027397260274,
|
| 505 |
+
"loss": 0.3741,
|
| 506 |
+
"step": 590
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 8.0,
|
| 510 |
+
"eval_accuracy": 0.8683206106870229,
|
| 511 |
+
"eval_f1": 0.8398298890291713,
|
| 512 |
+
"eval_loss": 0.34485381841659546,
|
| 513 |
+
"eval_precision": 0.8229333868378812,
|
| 514 |
+
"eval_recall": 0.8677806341045415,
|
| 515 |
+
"eval_runtime": 2.988,
|
| 516 |
+
"eval_samples_per_second": 175.367,
|
| 517 |
+
"eval_steps_per_second": 11.044,
|
| 518 |
+
"step": 590
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.135593220338983,
|
| 522 |
+
"grad_norm": 1.7316805124282837,
|
| 523 |
+
"learning_rate": 0.0009041095890410959,
|
| 524 |
+
"loss": 0.4155,
|
| 525 |
+
"step": 600
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.271186440677965,
|
| 529 |
+
"grad_norm": 0.8091392517089844,
|
| 530 |
+
"learning_rate": 0.0008356164383561645,
|
| 531 |
+
"loss": 0.3834,
|
| 532 |
+
"step": 610
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.40677966101695,
|
| 536 |
+
"grad_norm": 0.47697487473487854,
|
| 537 |
+
"learning_rate": 0.000767123287671233,
|
| 538 |
+
"loss": 0.3501,
|
| 539 |
+
"step": 620
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.542372881355933,
|
| 543 |
+
"grad_norm": 0.419321209192276,
|
| 544 |
+
"learning_rate": 0.0006986301369863014,
|
| 545 |
+
"loss": 0.349,
|
| 546 |
+
"step": 630
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.677966101694915,
|
| 550 |
+
"grad_norm": 2.6489877700805664,
|
| 551 |
+
"learning_rate": 0.0006301369863013699,
|
| 552 |
+
"loss": 0.3791,
|
| 553 |
+
"step": 640
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 8.813559322033898,
|
| 557 |
+
"grad_norm": 0.7624053359031677,
|
| 558 |
+
"learning_rate": 0.0005616438356164384,
|
| 559 |
+
"loss": 0.3655,
|
| 560 |
+
"step": 650
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 8.94915254237288,
|
| 564 |
+
"grad_norm": 0.44931384921073914,
|
| 565 |
+
"learning_rate": 0.0004931506849315068,
|
| 566 |
+
"loss": 0.3964,
|
| 567 |
+
"step": 660
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 8.989830508474576,
|
| 571 |
+
"eval_accuracy": 0.8587786259541985,
|
| 572 |
+
"eval_f1": 0.830509659935309,
|
| 573 |
+
"eval_loss": 0.3624710738658905,
|
| 574 |
+
"eval_precision": 0.8127875562832583,
|
| 575 |
+
"eval_recall": 0.8637722555460345,
|
| 576 |
+
"eval_runtime": 3.2987,
|
| 577 |
+
"eval_samples_per_second": 158.852,
|
| 578 |
+
"eval_steps_per_second": 10.004,
|
| 579 |
+
"step": 663
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 9.084745762711865,
|
| 583 |
+
"grad_norm": 0.7742342948913574,
|
| 584 |
+
"learning_rate": 0.00042465753424657536,
|
| 585 |
+
"loss": 0.3609,
|
| 586 |
+
"step": 670
|
| 587 |
+
},
|
| 588 |
+
{
|
| 589 |
+
"epoch": 9.220338983050848,
|
| 590 |
+
"grad_norm": 0.2361624389886856,
|
| 591 |
+
"learning_rate": 0.00035616438356164383,
|
| 592 |
+
"loss": 0.3231,
|
| 593 |
+
"step": 680
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 9.35593220338983,
|
| 597 |
+
"grad_norm": 0.3743881583213806,
|
| 598 |
+
"learning_rate": 0.00028767123287671236,
|
| 599 |
+
"loss": 0.3835,
|
| 600 |
+
"step": 690
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 9.491525423728813,
|
| 604 |
+
"grad_norm": 0.3935371935367584,
|
| 605 |
+
"learning_rate": 0.00021917808219178083,
|
| 606 |
+
"loss": 0.3503,
|
| 607 |
+
"step": 700
|
| 608 |
+
},
|
| 609 |
+
{
|
| 610 |
+
"epoch": 9.627118644067796,
|
| 611 |
+
"grad_norm": 0.6624675393104553,
|
| 612 |
+
"learning_rate": 0.00015068493150684933,
|
| 613 |
+
"loss": 0.3023,
|
| 614 |
+
"step": 710
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 9.76271186440678,
|
| 618 |
+
"grad_norm": 1.4423840045928955,
|
| 619 |
+
"learning_rate": 8.21917808219178e-05,
|
| 620 |
+
"loss": 0.404,
|
| 621 |
+
"step": 720
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 9.898305084745763,
|
| 625 |
+
"grad_norm": 0.639860212802887,
|
| 626 |
+
"learning_rate": 1.3698630136986302e-05,
|
| 627 |
+
"loss": 0.3845,
|
| 628 |
+
"step": 730
|
| 629 |
+
},
|
| 630 |
+
{
|
| 631 |
+
"epoch": 9.898305084745763,
|
| 632 |
+
"eval_accuracy": 0.8568702290076335,
|
| 633 |
+
"eval_f1": 0.8291712851274695,
|
| 634 |
+
"eval_loss": 0.3568853735923767,
|
| 635 |
+
"eval_precision": 0.8111482756381113,
|
| 636 |
+
"eval_recall": 0.8649052651623346,
|
| 637 |
+
"eval_runtime": 3.0816,
|
| 638 |
+
"eval_samples_per_second": 170.043,
|
| 639 |
+
"eval_steps_per_second": 10.709,
|
| 640 |
+
"step": 730
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 9.898305084745763,
|
| 644 |
+
"step": 730,
|
| 645 |
+
"total_flos": 3.6354082184735293e+18,
|
| 646 |
+
"train_loss": 0.42946596570210915,
|
| 647 |
+
"train_runtime": 617.6458,
|
| 648 |
+
"train_samples_per_second": 76.225,
|
| 649 |
+
"train_steps_per_second": 1.182
|
| 650 |
+
}
|
| 651 |
+
],
|
| 652 |
+
"logging_steps": 10,
|
| 653 |
+
"max_steps": 730,
|
| 654 |
+
"num_input_tokens_seen": 0,
|
| 655 |
+
"num_train_epochs": 10,
|
| 656 |
+
"save_steps": 500,
|
| 657 |
+
"stateful_callbacks": {
|
| 658 |
+
"TrainerControl": {
|
| 659 |
+
"args": {
|
| 660 |
+
"should_epoch_stop": false,
|
| 661 |
+
"should_evaluate": false,
|
| 662 |
+
"should_log": false,
|
| 663 |
+
"should_save": true,
|
| 664 |
+
"should_training_stop": true
|
| 665 |
+
},
|
| 666 |
+
"attributes": {}
|
| 667 |
+
}
|
| 668 |
+
},
|
| 669 |
+
"total_flos": 3.6354082184735293e+18,
|
| 670 |
+
"train_batch_size": 16,
|
| 671 |
+
"trial_name": null,
|
| 672 |
+
"trial_params": null
|
| 673 |
+
}
|