Commit
•
ef30713
1
Parent(s):
ab12dea
🍻 cheers
Browse files- README.md +10 -6
- all_results.json +16 -0
- config.json +1 -1
- eval_results.json +12 -0
- train_results.json +8 -0
- trainer_state.json +415 -0
README.md
CHANGED
|
@@ -1,7 +1,11 @@
|
|
| 1 |
---
|
|
|
|
|
|
|
| 2 |
license: apache-2.0
|
| 3 |
base_model: facebook/dinov2-large
|
| 4 |
tags:
|
|
|
|
|
|
|
| 5 |
- generated_from_trainer
|
| 6 |
metrics:
|
| 7 |
- accuracy
|
|
@@ -15,13 +19,13 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 15 |
|
| 16 |
# dino-large-2023_12_19-kornia_img-size518_batch-size16_epochs20
|
| 17 |
|
| 18 |
-
This model is a fine-tuned version of [facebook/dinov2-large](https://huggingface.co/facebook/dinov2-large) on the
|
| 19 |
It achieves the following results on the evaluation set:
|
| 20 |
-
- Loss: 0.
|
| 21 |
-
- F1 Micro: 0.
|
| 22 |
-
- F1 Macro: 0.
|
| 23 |
-
- Roc Auc: 0.
|
| 24 |
-
- Accuracy: 0.
|
| 25 |
- Learning Rate: 0.001
|
| 26 |
|
| 27 |
## Model description
|
|
|
|
| 1 |
---
|
| 2 |
+
language:
|
| 3 |
+
- eng
|
| 4 |
license: apache-2.0
|
| 5 |
base_model: facebook/dinov2-large
|
| 6 |
tags:
|
| 7 |
+
- multilabel-image-classification
|
| 8 |
+
- multilabel
|
| 9 |
- generated_from_trainer
|
| 10 |
metrics:
|
| 11 |
- accuracy
|
|
|
|
| 19 |
|
| 20 |
# dino-large-2023_12_19-kornia_img-size518_batch-size16_epochs20
|
| 21 |
|
| 22 |
+
This model is a fine-tuned version of [facebook/dinov2-large](https://huggingface.co/facebook/dinov2-large) on the multilabel_complete_dataset dataset.
|
| 23 |
It achieves the following results on the evaluation set:
|
| 24 |
+
- Loss: 0.1199
|
| 25 |
+
- F1 Micro: 0.8367
|
| 26 |
+
- F1 Macro: 0.8026
|
| 27 |
+
- Roc Auc: 0.9072
|
| 28 |
+
- Accuracy: 0.5354
|
| 29 |
- Learning Rate: 0.001
|
| 30 |
|
| 31 |
## Model description
|
all_results.json
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"eval_accuracy": 0.5353748680042238,
|
| 4 |
+
"eval_f1_macro": 0.8025605807344918,
|
| 5 |
+
"eval_f1_micro": 0.8367241081205695,
|
| 6 |
+
"eval_loss": 0.11987907439470291,
|
| 7 |
+
"eval_roc_auc": 0.9072272159350702,
|
| 8 |
+
"eval_runtime": 691.1321,
|
| 9 |
+
"eval_samples_per_second": 4.111,
|
| 10 |
+
"eval_steps_per_second": 0.258,
|
| 11 |
+
"learning_rate": 0.001,
|
| 12 |
+
"train_loss": 0.18060117693089728,
|
| 13 |
+
"train_runtime": 57213.4657,
|
| 14 |
+
"train_samples_per_second": 2.995,
|
| 15 |
+
"train_steps_per_second": 0.187
|
| 16 |
+
}
|
config.json
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
{
|
| 2 |
-
"_name_or_path": "facebook/dinov2-
|
| 3 |
"apply_layernorm": true,
|
| 4 |
"architectures": [
|
| 5 |
"NewheadDinov2ForImageClassification"
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "facebook/dinov2-large2023_12_19",
|
| 3 |
"apply_layernorm": true,
|
| 4 |
"architectures": [
|
| 5 |
"NewheadDinov2ForImageClassification"
|
eval_results.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"eval_accuracy": 0.5353748680042238,
|
| 4 |
+
"eval_f1_macro": 0.8025605807344918,
|
| 5 |
+
"eval_f1_micro": 0.8367241081205695,
|
| 6 |
+
"eval_loss": 0.11987907439470291,
|
| 7 |
+
"eval_roc_auc": 0.9072272159350702,
|
| 8 |
+
"eval_runtime": 691.1321,
|
| 9 |
+
"eval_samples_per_second": 4.111,
|
| 10 |
+
"eval_steps_per_second": 0.258,
|
| 11 |
+
"learning_rate": 0.001
|
| 12 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"learning_rate": 0.001,
|
| 4 |
+
"train_loss": 0.18060117693089728,
|
| 5 |
+
"train_runtime": 57213.4657,
|
| 6 |
+
"train_samples_per_second": 2.995,
|
| 7 |
+
"train_steps_per_second": 0.187
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,415 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.11592021584510803,
|
| 3 |
+
"best_model_checkpoint": "/home1/datawork/mcontini/models/multilabel/huggingface/dino-large-2023_12_19-kornia_img-size518_batch-size16_epochs20/checkpoint-9112",
|
| 4 |
+
"epoch": 20.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 10720,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.93,
|
| 13 |
+
"learning_rate": 0.01,
|
| 14 |
+
"loss": 0.2545,
|
| 15 |
+
"step": 500
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 1.0,
|
| 19 |
+
"eval_accuracy": 0.41336191496963204,
|
| 20 |
+
"eval_f1_macro": 0.5552795713944485,
|
| 21 |
+
"eval_f1_micro": 0.7016183870529035,
|
| 22 |
+
"eval_loss": 0.19357216358184814,
|
| 23 |
+
"eval_roc_auc": 0.7908326703460342,
|
| 24 |
+
"eval_runtime": 669.1996,
|
| 25 |
+
"eval_samples_per_second": 4.183,
|
| 26 |
+
"eval_steps_per_second": 0.262,
|
| 27 |
+
"learning_rate": 0.01,
|
| 28 |
+
"step": 536
|
| 29 |
+
},
|
| 30 |
+
{
|
| 31 |
+
"epoch": 1.87,
|
| 32 |
+
"learning_rate": 0.01,
|
| 33 |
+
"loss": 0.2163,
|
| 34 |
+
"step": 1000
|
| 35 |
+
},
|
| 36 |
+
{
|
| 37 |
+
"epoch": 2.0,
|
| 38 |
+
"eval_accuracy": 0.4240800285816363,
|
| 39 |
+
"eval_f1_macro": 0.6939559668202788,
|
| 40 |
+
"eval_f1_micro": 0.7671674437345306,
|
| 41 |
+
"eval_loss": 0.16442517936229706,
|
| 42 |
+
"eval_roc_auc": 0.8668895881413139,
|
| 43 |
+
"eval_runtime": 667.4861,
|
| 44 |
+
"eval_samples_per_second": 4.193,
|
| 45 |
+
"eval_steps_per_second": 0.262,
|
| 46 |
+
"learning_rate": 0.01,
|
| 47 |
+
"step": 1072
|
| 48 |
+
},
|
| 49 |
+
{
|
| 50 |
+
"epoch": 2.8,
|
| 51 |
+
"learning_rate": 0.01,
|
| 52 |
+
"loss": 0.2142,
|
| 53 |
+
"step": 1500
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 3.0,
|
| 57 |
+
"eval_accuracy": 0.4258663808503037,
|
| 58 |
+
"eval_f1_macro": 0.6226385110394398,
|
| 59 |
+
"eval_f1_micro": 0.7264495032071437,
|
| 60 |
+
"eval_loss": 0.17197823524475098,
|
| 61 |
+
"eval_roc_auc": 0.8209912192682602,
|
| 62 |
+
"eval_runtime": 671.4241,
|
| 63 |
+
"eval_samples_per_second": 4.169,
|
| 64 |
+
"eval_steps_per_second": 0.261,
|
| 65 |
+
"learning_rate": 0.01,
|
| 66 |
+
"step": 1608
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 3.73,
|
| 70 |
+
"learning_rate": 0.01,
|
| 71 |
+
"loss": 0.2107,
|
| 72 |
+
"step": 2000
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 4.0,
|
| 76 |
+
"eval_accuracy": 0.40192926045016075,
|
| 77 |
+
"eval_f1_macro": 0.60556232845899,
|
| 78 |
+
"eval_f1_micro": 0.7311342592592591,
|
| 79 |
+
"eval_loss": 0.17794817686080933,
|
| 80 |
+
"eval_roc_auc": 0.844154343443223,
|
| 81 |
+
"eval_runtime": 669.3476,
|
| 82 |
+
"eval_samples_per_second": 4.182,
|
| 83 |
+
"eval_steps_per_second": 0.261,
|
| 84 |
+
"learning_rate": 0.01,
|
| 85 |
+
"step": 2144
|
| 86 |
+
},
|
| 87 |
+
{
|
| 88 |
+
"epoch": 4.66,
|
| 89 |
+
"learning_rate": 0.01,
|
| 90 |
+
"loss": 0.2117,
|
| 91 |
+
"step": 2500
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 5.0,
|
| 95 |
+
"eval_accuracy": 0.38335119685602,
|
| 96 |
+
"eval_f1_macro": 0.674484878286151,
|
| 97 |
+
"eval_f1_micro": 0.754215411654858,
|
| 98 |
+
"eval_loss": 0.1834840625524521,
|
| 99 |
+
"eval_roc_auc": 0.8723557298678095,
|
| 100 |
+
"eval_runtime": 667.4497,
|
| 101 |
+
"eval_samples_per_second": 4.194,
|
| 102 |
+
"eval_steps_per_second": 0.262,
|
| 103 |
+
"learning_rate": 0.01,
|
| 104 |
+
"step": 2680
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 5.6,
|
| 108 |
+
"learning_rate": 0.01,
|
| 109 |
+
"loss": 0.2171,
|
| 110 |
+
"step": 3000
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 6.0,
|
| 114 |
+
"eval_accuracy": 0.42086459449803504,
|
| 115 |
+
"eval_f1_macro": 0.595946612091723,
|
| 116 |
+
"eval_f1_micro": 0.7347275031685679,
|
| 117 |
+
"eval_loss": 0.1732301414012909,
|
| 118 |
+
"eval_roc_auc": 0.8236385276926494,
|
| 119 |
+
"eval_runtime": 665.3111,
|
| 120 |
+
"eval_samples_per_second": 4.207,
|
| 121 |
+
"eval_steps_per_second": 0.263,
|
| 122 |
+
"learning_rate": 0.01,
|
| 123 |
+
"step": 3216
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 6.53,
|
| 127 |
+
"learning_rate": 0.01,
|
| 128 |
+
"loss": 0.2178,
|
| 129 |
+
"step": 3500
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 7.0,
|
| 133 |
+
"eval_accuracy": 0.3904966059306895,
|
| 134 |
+
"eval_f1_macro": 0.5932113008464752,
|
| 135 |
+
"eval_f1_micro": 0.7253084446717382,
|
| 136 |
+
"eval_loss": 0.2698315978050232,
|
| 137 |
+
"eval_roc_auc": 0.8165478137957464,
|
| 138 |
+
"eval_runtime": 666.3656,
|
| 139 |
+
"eval_samples_per_second": 4.2,
|
| 140 |
+
"eval_steps_per_second": 0.263,
|
| 141 |
+
"learning_rate": 0.01,
|
| 142 |
+
"step": 3752
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 7.46,
|
| 146 |
+
"learning_rate": 0.01,
|
| 147 |
+
"loss": 0.2177,
|
| 148 |
+
"step": 4000
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 8.0,
|
| 152 |
+
"eval_accuracy": 0.4119328331546981,
|
| 153 |
+
"eval_f1_macro": 0.6280317383192718,
|
| 154 |
+
"eval_f1_micro": 0.7360358677377172,
|
| 155 |
+
"eval_loss": 0.19401037693023682,
|
| 156 |
+
"eval_roc_auc": 0.828564591229026,
|
| 157 |
+
"eval_runtime": 671.4879,
|
| 158 |
+
"eval_samples_per_second": 4.168,
|
| 159 |
+
"eval_steps_per_second": 0.261,
|
| 160 |
+
"learning_rate": 0.01,
|
| 161 |
+
"step": 4288
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 8.4,
|
| 165 |
+
"learning_rate": 0.001,
|
| 166 |
+
"loss": 0.212,
|
| 167 |
+
"step": 4500
|
| 168 |
+
},
|
| 169 |
+
{
|
| 170 |
+
"epoch": 9.0,
|
| 171 |
+
"eval_accuracy": 0.48981779206859594,
|
| 172 |
+
"eval_f1_macro": 0.7491476401858057,
|
| 173 |
+
"eval_f1_micro": 0.7992611571232794,
|
| 174 |
+
"eval_loss": 0.1454659402370453,
|
| 175 |
+
"eval_roc_auc": 0.8756991587148717,
|
| 176 |
+
"eval_runtime": 673.1035,
|
| 177 |
+
"eval_samples_per_second": 4.158,
|
| 178 |
+
"eval_steps_per_second": 0.26,
|
| 179 |
+
"learning_rate": 0.001,
|
| 180 |
+
"step": 4824
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 9.33,
|
| 184 |
+
"learning_rate": 0.001,
|
| 185 |
+
"loss": 0.1761,
|
| 186 |
+
"step": 5000
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 10.0,
|
| 190 |
+
"eval_accuracy": 0.5123258306538049,
|
| 191 |
+
"eval_f1_macro": 0.7661404438716006,
|
| 192 |
+
"eval_f1_micro": 0.8116191765647116,
|
| 193 |
+
"eval_loss": 0.13573050498962402,
|
| 194 |
+
"eval_roc_auc": 0.8733133406359384,
|
| 195 |
+
"eval_runtime": 691.4938,
|
| 196 |
+
"eval_samples_per_second": 4.048,
|
| 197 |
+
"eval_steps_per_second": 0.253,
|
| 198 |
+
"learning_rate": 0.001,
|
| 199 |
+
"step": 5360
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 10.26,
|
| 203 |
+
"learning_rate": 0.001,
|
| 204 |
+
"loss": 0.1681,
|
| 205 |
+
"step": 5500
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 11.0,
|
| 209 |
+
"eval_accuracy": 0.5166130760986066,
|
| 210 |
+
"eval_f1_macro": 0.7752662433951194,
|
| 211 |
+
"eval_f1_micro": 0.8151612707282998,
|
| 212 |
+
"eval_loss": 0.1386074423789978,
|
| 213 |
+
"eval_roc_auc": 0.8790914856208012,
|
| 214 |
+
"eval_runtime": 689.8939,
|
| 215 |
+
"eval_samples_per_second": 4.057,
|
| 216 |
+
"eval_steps_per_second": 0.254,
|
| 217 |
+
"learning_rate": 0.001,
|
| 218 |
+
"step": 5896
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 11.19,
|
| 222 |
+
"learning_rate": 0.001,
|
| 223 |
+
"loss": 0.1579,
|
| 224 |
+
"step": 6000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 12.0,
|
| 228 |
+
"eval_accuracy": 0.5162558056448732,
|
| 229 |
+
"eval_f1_macro": 0.7826923679681995,
|
| 230 |
+
"eval_f1_micro": 0.8219824282092104,
|
| 231 |
+
"eval_loss": 0.1820007711648941,
|
| 232 |
+
"eval_roc_auc": 0.8918796772066396,
|
| 233 |
+
"eval_runtime": 692.3943,
|
| 234 |
+
"eval_samples_per_second": 4.042,
|
| 235 |
+
"eval_steps_per_second": 0.253,
|
| 236 |
+
"learning_rate": 0.001,
|
| 237 |
+
"step": 6432
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 12.13,
|
| 241 |
+
"learning_rate": 0.001,
|
| 242 |
+
"loss": 0.1553,
|
| 243 |
+
"step": 6500
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 13.0,
|
| 247 |
+
"eval_accuracy": 0.5326902465166131,
|
| 248 |
+
"eval_f1_macro": 0.7907678649401602,
|
| 249 |
+
"eval_f1_micro": 0.8296795952782462,
|
| 250 |
+
"eval_loss": 0.1227828711271286,
|
| 251 |
+
"eval_roc_auc": 0.8898245747393245,
|
| 252 |
+
"eval_runtime": 691.2613,
|
| 253 |
+
"eval_samples_per_second": 4.049,
|
| 254 |
+
"eval_steps_per_second": 0.253,
|
| 255 |
+
"learning_rate": 0.001,
|
| 256 |
+
"step": 6968
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 13.06,
|
| 260 |
+
"learning_rate": 0.001,
|
| 261 |
+
"loss": 0.1517,
|
| 262 |
+
"step": 7000
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 13.99,
|
| 266 |
+
"learning_rate": 0.001,
|
| 267 |
+
"loss": 0.1512,
|
| 268 |
+
"step": 7500
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 14.0,
|
| 272 |
+
"eval_accuracy": 0.5301893533404788,
|
| 273 |
+
"eval_f1_macro": 0.7815063263088662,
|
| 274 |
+
"eval_f1_micro": 0.8257773991358852,
|
| 275 |
+
"eval_loss": 0.12334247678518295,
|
| 276 |
+
"eval_roc_auc": 0.8845045355403204,
|
| 277 |
+
"eval_runtime": 684.2138,
|
| 278 |
+
"eval_samples_per_second": 4.091,
|
| 279 |
+
"eval_steps_per_second": 0.256,
|
| 280 |
+
"learning_rate": 0.001,
|
| 281 |
+
"step": 7504
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 14.93,
|
| 285 |
+
"learning_rate": 0.001,
|
| 286 |
+
"loss": 0.1508,
|
| 287 |
+
"step": 8000
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 15.0,
|
| 291 |
+
"eval_accuracy": 0.5305466237942122,
|
| 292 |
+
"eval_f1_macro": 0.7681866841890026,
|
| 293 |
+
"eval_f1_micro": 0.817921948794432,
|
| 294 |
+
"eval_loss": 0.12476785480976105,
|
| 295 |
+
"eval_roc_auc": 0.8739908739957604,
|
| 296 |
+
"eval_runtime": 690.4773,
|
| 297 |
+
"eval_samples_per_second": 4.054,
|
| 298 |
+
"eval_steps_per_second": 0.253,
|
| 299 |
+
"learning_rate": 0.001,
|
| 300 |
+
"step": 8040
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"epoch": 15.86,
|
| 304 |
+
"learning_rate": 0.001,
|
| 305 |
+
"loss": 0.1499,
|
| 306 |
+
"step": 8500
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 16.0,
|
| 310 |
+
"eval_accuracy": 0.5394783851375491,
|
| 311 |
+
"eval_f1_macro": 0.7902552675483185,
|
| 312 |
+
"eval_f1_micro": 0.8276715549780336,
|
| 313 |
+
"eval_loss": 0.11932893097400665,
|
| 314 |
+
"eval_roc_auc": 0.8805997685784225,
|
| 315 |
+
"eval_runtime": 690.469,
|
| 316 |
+
"eval_samples_per_second": 4.054,
|
| 317 |
+
"eval_steps_per_second": 0.253,
|
| 318 |
+
"learning_rate": 0.001,
|
| 319 |
+
"step": 8576
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 16.79,
|
| 323 |
+
"learning_rate": 0.001,
|
| 324 |
+
"loss": 0.1435,
|
| 325 |
+
"step": 9000
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 17.0,
|
| 329 |
+
"eval_accuracy": 0.5380493033226152,
|
| 330 |
+
"eval_f1_macro": 0.7995743847858905,
|
| 331 |
+
"eval_f1_micro": 0.8381208681916574,
|
| 332 |
+
"eval_loss": 0.11592021584510803,
|
| 333 |
+
"eval_roc_auc": 0.9037053380559585,
|
| 334 |
+
"eval_runtime": 685.0697,
|
| 335 |
+
"eval_samples_per_second": 4.086,
|
| 336 |
+
"eval_steps_per_second": 0.255,
|
| 337 |
+
"learning_rate": 0.001,
|
| 338 |
+
"step": 9112
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 17.72,
|
| 342 |
+
"learning_rate": 0.001,
|
| 343 |
+
"loss": 0.1463,
|
| 344 |
+
"step": 9500
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 18.0,
|
| 348 |
+
"eval_accuracy": 0.5480528760271526,
|
| 349 |
+
"eval_f1_macro": 0.803284244956481,
|
| 350 |
+
"eval_f1_micro": 0.8392878623842175,
|
| 351 |
+
"eval_loss": 0.11661119759082794,
|
| 352 |
+
"eval_roc_auc": 0.8956742640228449,
|
| 353 |
+
"eval_runtime": 684.4351,
|
| 354 |
+
"eval_samples_per_second": 4.09,
|
| 355 |
+
"eval_steps_per_second": 0.256,
|
| 356 |
+
"learning_rate": 0.001,
|
| 357 |
+
"step": 9648
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 18.66,
|
| 361 |
+
"learning_rate": 0.001,
|
| 362 |
+
"loss": 0.1423,
|
| 363 |
+
"step": 10000
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 19.0,
|
| 367 |
+
"eval_accuracy": 0.5459092533047517,
|
| 368 |
+
"eval_f1_macro": 0.8008505005735617,
|
| 369 |
+
"eval_f1_micro": 0.8327005511328843,
|
| 370 |
+
"eval_loss": 0.1216365247964859,
|
| 371 |
+
"eval_roc_auc": 0.8864852456725716,
|
| 372 |
+
"eval_runtime": 683.4387,
|
| 373 |
+
"eval_samples_per_second": 4.095,
|
| 374 |
+
"eval_steps_per_second": 0.256,
|
| 375 |
+
"learning_rate": 0.001,
|
| 376 |
+
"step": 10184
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 19.59,
|
| 380 |
+
"learning_rate": 0.001,
|
| 381 |
+
"loss": 0.1444,
|
| 382 |
+
"step": 10500
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 20.0,
|
| 386 |
+
"eval_accuracy": 0.5509110396570204,
|
| 387 |
+
"eval_f1_macro": 0.801973744395412,
|
| 388 |
+
"eval_f1_micro": 0.8382846336851095,
|
| 389 |
+
"eval_loss": 0.1171422228217125,
|
| 390 |
+
"eval_roc_auc": 0.8907672852306303,
|
| 391 |
+
"eval_runtime": 675.9942,
|
| 392 |
+
"eval_samples_per_second": 4.141,
|
| 393 |
+
"eval_steps_per_second": 0.259,
|
| 394 |
+
"learning_rate": 0.001,
|
| 395 |
+
"step": 10720
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 20.0,
|
| 399 |
+
"learning_rate": 0.001,
|
| 400 |
+
"step": 10720,
|
| 401 |
+
"total_flos": 2.5409000327874334e+20,
|
| 402 |
+
"train_loss": 0.18060117693089728,
|
| 403 |
+
"train_runtime": 57213.4657,
|
| 404 |
+
"train_samples_per_second": 2.995,
|
| 405 |
+
"train_steps_per_second": 0.187
|
| 406 |
+
}
|
| 407 |
+
],
|
| 408 |
+
"logging_steps": 500,
|
| 409 |
+
"max_steps": 10720,
|
| 410 |
+
"num_train_epochs": 20,
|
| 411 |
+
"save_steps": 500,
|
| 412 |
+
"total_flos": 2.5409000327874334e+20,
|
| 413 |
+
"trial_name": null,
|
| 414 |
+
"trial_params": null
|
| 415 |
+
}
|