Commit
•
60c4c2d
1
Parent(s):
6afe404
🍻 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 +611 -0
README.md
CHANGED
|
@@ -1,7 +1,11 @@
|
|
| 1 |
---
|
|
|
|
|
|
|
| 2 |
license: apache-2.0
|
| 3 |
base_model: facebook/dinov2-base
|
| 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-base-2023_11_24-unfreeze
|
| 17 |
|
| 18 |
-
This model is a fine-tuned version of [facebook/dinov2-base](https://huggingface.co/facebook/dinov2-base) 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-base
|
| 6 |
tags:
|
| 7 |
+
- image-classification
|
| 8 |
+
- multilabel
|
| 9 |
- generated_from_trainer
|
| 10 |
metrics:
|
| 11 |
- accuracy
|
|
|
|
| 19 |
|
| 20 |
# dino-base-2023_11_24-unfreeze
|
| 21 |
|
| 22 |
+
This model is a fine-tuned version of [facebook/dinov2-base](https://huggingface.co/facebook/dinov2-base) on the multilabel_complete_dataset dataset.
|
| 23 |
It achieves the following results on the evaluation set:
|
| 24 |
+
- Loss: 0.2045
|
| 25 |
+
- F1 Micro: 0.6595
|
| 26 |
+
- F1 Macro: 0.5161
|
| 27 |
+
- Roc Auc: 0.7681
|
| 28 |
+
- Accuracy: 0.2735
|
| 29 |
- Learning Rate: 0.001
|
| 30 |
|
| 31 |
## Model description
|
all_results.json
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 30.0,
|
| 3 |
+
"eval_accuracy": 0.2734952481520591,
|
| 4 |
+
"eval_f1_macro": 0.5160923574023799,
|
| 5 |
+
"eval_f1_micro": 0.6595394736842105,
|
| 6 |
+
"eval_loss": 0.2044779360294342,
|
| 7 |
+
"eval_roc_auc": 0.768097908147388,
|
| 8 |
+
"eval_runtime": 673.2705,
|
| 9 |
+
"eval_samples_per_second": 4.22,
|
| 10 |
+
"eval_steps_per_second": 0.264,
|
| 11 |
+
"learning_rate": 0.001,
|
| 12 |
+
"train_loss": 0.2889823566028728,
|
| 13 |
+
"train_runtime": 84191.1285,
|
| 14 |
+
"train_samples_per_second": 3.053,
|
| 15 |
+
"train_steps_per_second": 0.191
|
| 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-base2023_11_24",
|
| 3 |
"apply_layernorm": true,
|
| 4 |
"architectures": [
|
| 5 |
"NewheadDinov2ForImageClassification"
|
eval_results.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 30.0,
|
| 3 |
+
"eval_accuracy": 0.2734952481520591,
|
| 4 |
+
"eval_f1_macro": 0.5160923574023799,
|
| 5 |
+
"eval_f1_micro": 0.6595394736842105,
|
| 6 |
+
"eval_loss": 0.2044779360294342,
|
| 7 |
+
"eval_roc_auc": 0.768097908147388,
|
| 8 |
+
"eval_runtime": 673.2705,
|
| 9 |
+
"eval_samples_per_second": 4.22,
|
| 10 |
+
"eval_steps_per_second": 0.264,
|
| 11 |
+
"learning_rate": 0.001
|
| 12 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 30.0,
|
| 3 |
+
"learning_rate": 0.001,
|
| 4 |
+
"train_loss": 0.2889823566028728,
|
| 5 |
+
"train_runtime": 84191.1285,
|
| 6 |
+
"train_samples_per_second": 3.053,
|
| 7 |
+
"train_steps_per_second": 0.191
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,611 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.20685863494873047,
|
| 3 |
+
"best_model_checkpoint": "./dino-base-2023_11_24-unfreeze/checkpoint-16080",
|
| 4 |
+
"epoch": 30.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 16080,
|
| 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": 0.3961,
|
| 15 |
+
"step": 500
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 1.0,
|
| 19 |
+
"eval_accuracy": 0.0735977134690961,
|
| 20 |
+
"eval_f1_macro": 0.09660814200449142,
|
| 21 |
+
"eval_f1_micro": 0.3295561046263951,
|
| 22 |
+
"eval_loss": 0.32934051752090454,
|
| 23 |
+
"eval_roc_auc": 0.6004986038194676,
|
| 24 |
+
"eval_runtime": 666.7423,
|
| 25 |
+
"eval_samples_per_second": 4.198,
|
| 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": 0.3418,
|
| 34 |
+
"step": 1000
|
| 35 |
+
},
|
| 36 |
+
{
|
| 37 |
+
"epoch": 2.0,
|
| 38 |
+
"eval_accuracy": 0.04394426580921758,
|
| 39 |
+
"eval_f1_macro": 0.12832379577401481,
|
| 40 |
+
"eval_f1_micro": 0.33792181231809953,
|
| 41 |
+
"eval_loss": 0.3381994962692261,
|
| 42 |
+
"eval_roc_auc": 0.6053684490893504,
|
| 43 |
+
"eval_runtime": 663.9506,
|
| 44 |
+
"eval_samples_per_second": 4.216,
|
| 45 |
+
"eval_steps_per_second": 0.264,
|
| 46 |
+
"learning_rate": 0.01,
|
| 47 |
+
"step": 1072
|
| 48 |
+
},
|
| 49 |
+
{
|
| 50 |
+
"epoch": 2.8,
|
| 51 |
+
"learning_rate": 0.01,
|
| 52 |
+
"loss": 0.3334,
|
| 53 |
+
"step": 1500
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 3.0,
|
| 57 |
+
"eval_accuracy": 0.07859949982136477,
|
| 58 |
+
"eval_f1_macro": 0.2420082514645649,
|
| 59 |
+
"eval_f1_micro": 0.39049407685639986,
|
| 60 |
+
"eval_loss": 0.3550550937652588,
|
| 61 |
+
"eval_roc_auc": 0.6319467749490031,
|
| 62 |
+
"eval_runtime": 662.2838,
|
| 63 |
+
"eval_samples_per_second": 4.226,
|
| 64 |
+
"eval_steps_per_second": 0.264,
|
| 65 |
+
"learning_rate": 0.01,
|
| 66 |
+
"step": 1608
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 3.73,
|
| 70 |
+
"learning_rate": 0.01,
|
| 71 |
+
"loss": 0.3323,
|
| 72 |
+
"step": 2000
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 4.0,
|
| 76 |
+
"eval_accuracy": 0.06252232940335835,
|
| 77 |
+
"eval_f1_macro": 0.10992415783030689,
|
| 78 |
+
"eval_f1_micro": 0.2554886521537749,
|
| 79 |
+
"eval_loss": 0.32132452726364136,
|
| 80 |
+
"eval_roc_auc": 0.5720375077972393,
|
| 81 |
+
"eval_runtime": 688.3751,
|
| 82 |
+
"eval_samples_per_second": 4.066,
|
| 83 |
+
"eval_steps_per_second": 0.254,
|
| 84 |
+
"learning_rate": 0.01,
|
| 85 |
+
"step": 2144
|
| 86 |
+
},
|
| 87 |
+
{
|
| 88 |
+
"epoch": 4.66,
|
| 89 |
+
"learning_rate": 0.01,
|
| 90 |
+
"loss": 0.3248,
|
| 91 |
+
"step": 2500
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 5.0,
|
| 95 |
+
"eval_accuracy": 0.04501607717041801,
|
| 96 |
+
"eval_f1_macro": 0.12977783250072622,
|
| 97 |
+
"eval_f1_micro": 0.33549517966695885,
|
| 98 |
+
"eval_loss": 0.3164198696613312,
|
| 99 |
+
"eval_roc_auc": 0.6023923090497274,
|
| 100 |
+
"eval_runtime": 684.0693,
|
| 101 |
+
"eval_samples_per_second": 4.092,
|
| 102 |
+
"eval_steps_per_second": 0.256,
|
| 103 |
+
"learning_rate": 0.01,
|
| 104 |
+
"step": 2680
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 5.6,
|
| 108 |
+
"learning_rate": 0.01,
|
| 109 |
+
"loss": 0.3235,
|
| 110 |
+
"step": 3000
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 6.0,
|
| 114 |
+
"eval_accuracy": 0.023937120400142908,
|
| 115 |
+
"eval_f1_macro": 0.08064530717139505,
|
| 116 |
+
"eval_f1_micro": 0.28637243158573666,
|
| 117 |
+
"eval_loss": 0.33460524678230286,
|
| 118 |
+
"eval_roc_auc": 0.5833928540072826,
|
| 119 |
+
"eval_runtime": 694.014,
|
| 120 |
+
"eval_samples_per_second": 4.033,
|
| 121 |
+
"eval_steps_per_second": 0.252,
|
| 122 |
+
"learning_rate": 0.01,
|
| 123 |
+
"step": 3216
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 6.53,
|
| 127 |
+
"learning_rate": 0.01,
|
| 128 |
+
"loss": 0.32,
|
| 129 |
+
"step": 3500
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 7.0,
|
| 133 |
+
"eval_accuracy": 0.06823865666309396,
|
| 134 |
+
"eval_f1_macro": 0.19677941241236616,
|
| 135 |
+
"eval_f1_micro": 0.4593774264134962,
|
| 136 |
+
"eval_loss": 0.3029455542564392,
|
| 137 |
+
"eval_roc_auc": 0.6663014350966586,
|
| 138 |
+
"eval_runtime": 675.5021,
|
| 139 |
+
"eval_samples_per_second": 4.144,
|
| 140 |
+
"eval_steps_per_second": 0.259,
|
| 141 |
+
"learning_rate": 0.01,
|
| 142 |
+
"step": 3752
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 7.46,
|
| 146 |
+
"learning_rate": 0.01,
|
| 147 |
+
"loss": 0.3138,
|
| 148 |
+
"step": 4000
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 8.0,
|
| 152 |
+
"eval_accuracy": 0.05787781350482315,
|
| 153 |
+
"eval_f1_macro": 0.29399778254935266,
|
| 154 |
+
"eval_f1_micro": 0.5467535723238907,
|
| 155 |
+
"eval_loss": 0.2865561246871948,
|
| 156 |
+
"eval_roc_auc": 0.7240336144955091,
|
| 157 |
+
"eval_runtime": 678.713,
|
| 158 |
+
"eval_samples_per_second": 4.124,
|
| 159 |
+
"eval_steps_per_second": 0.258,
|
| 160 |
+
"learning_rate": 0.01,
|
| 161 |
+
"step": 4288
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 8.4,
|
| 165 |
+
"learning_rate": 0.01,
|
| 166 |
+
"loss": 0.3052,
|
| 167 |
+
"step": 4500
|
| 168 |
+
},
|
| 169 |
+
{
|
| 170 |
+
"epoch": 9.0,
|
| 171 |
+
"eval_accuracy": 0.12254376563058235,
|
| 172 |
+
"eval_f1_macro": 0.29927318211914594,
|
| 173 |
+
"eval_f1_micro": 0.4766527970411466,
|
| 174 |
+
"eval_loss": 0.2806786894798279,
|
| 175 |
+
"eval_roc_auc": 0.6671674748237758,
|
| 176 |
+
"eval_runtime": 665.488,
|
| 177 |
+
"eval_samples_per_second": 4.206,
|
| 178 |
+
"eval_steps_per_second": 0.263,
|
| 179 |
+
"learning_rate": 0.01,
|
| 180 |
+
"step": 4824
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 9.33,
|
| 184 |
+
"learning_rate": 0.01,
|
| 185 |
+
"loss": 0.3157,
|
| 186 |
+
"step": 5000
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 10.0,
|
| 190 |
+
"eval_accuracy": 0.0707395498392283,
|
| 191 |
+
"eval_f1_macro": 0.20914365928952397,
|
| 192 |
+
"eval_f1_micro": 0.47518133978097,
|
| 193 |
+
"eval_loss": 0.29552116990089417,
|
| 194 |
+
"eval_roc_auc": 0.6732798876351146,
|
| 195 |
+
"eval_runtime": 670.1708,
|
| 196 |
+
"eval_samples_per_second": 4.177,
|
| 197 |
+
"eval_steps_per_second": 0.261,
|
| 198 |
+
"learning_rate": 0.01,
|
| 199 |
+
"step": 5360
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 10.26,
|
| 203 |
+
"learning_rate": 0.01,
|
| 204 |
+
"loss": 0.3119,
|
| 205 |
+
"step": 5500
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 11.0,
|
| 209 |
+
"eval_accuracy": 0.13612004287245444,
|
| 210 |
+
"eval_f1_macro": 0.215957700272248,
|
| 211 |
+
"eval_f1_micro": 0.40282685512367494,
|
| 212 |
+
"eval_loss": 0.340472012758255,
|
| 213 |
+
"eval_roc_auc": 0.6335842288286788,
|
| 214 |
+
"eval_runtime": 655.9163,
|
| 215 |
+
"eval_samples_per_second": 4.267,
|
| 216 |
+
"eval_steps_per_second": 0.267,
|
| 217 |
+
"learning_rate": 0.01,
|
| 218 |
+
"step": 5896
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 11.19,
|
| 222 |
+
"learning_rate": 0.01,
|
| 223 |
+
"loss": 0.3162,
|
| 224 |
+
"step": 6000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 12.0,
|
| 228 |
+
"eval_accuracy": 0.05323329760628796,
|
| 229 |
+
"eval_f1_macro": 0.2964933264503757,
|
| 230 |
+
"eval_f1_micro": 0.4898614735996788,
|
| 231 |
+
"eval_loss": 0.41625335812568665,
|
| 232 |
+
"eval_roc_auc": 0.6862559753775964,
|
| 233 |
+
"eval_runtime": 663.2841,
|
| 234 |
+
"eval_samples_per_second": 4.22,
|
| 235 |
+
"eval_steps_per_second": 0.264,
|
| 236 |
+
"learning_rate": 0.01,
|
| 237 |
+
"step": 6432
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 12.13,
|
| 241 |
+
"learning_rate": 0.01,
|
| 242 |
+
"loss": 0.3184,
|
| 243 |
+
"step": 6500
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 13.0,
|
| 247 |
+
"eval_accuracy": 0.10468024294390854,
|
| 248 |
+
"eval_f1_macro": 0.32985179099978684,
|
| 249 |
+
"eval_f1_micro": 0.5429481760842708,
|
| 250 |
+
"eval_loss": 0.29642441868782043,
|
| 251 |
+
"eval_roc_auc": 0.7170301433203381,
|
| 252 |
+
"eval_runtime": 670.4311,
|
| 253 |
+
"eval_samples_per_second": 4.175,
|
| 254 |
+
"eval_steps_per_second": 0.261,
|
| 255 |
+
"learning_rate": 0.01,
|
| 256 |
+
"step": 6968
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 13.06,
|
| 260 |
+
"learning_rate": 0.01,
|
| 261 |
+
"loss": 0.3131,
|
| 262 |
+
"step": 7000
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 13.99,
|
| 266 |
+
"learning_rate": 0.01,
|
| 267 |
+
"loss": 0.3142,
|
| 268 |
+
"step": 7500
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 14.0,
|
| 272 |
+
"eval_accuracy": 0.08324401571989996,
|
| 273 |
+
"eval_f1_macro": 0.3153914854710691,
|
| 274 |
+
"eval_f1_micro": 0.5253185233583796,
|
| 275 |
+
"eval_loss": 0.30047285556793213,
|
| 276 |
+
"eval_roc_auc": 0.7071979710639686,
|
| 277 |
+
"eval_runtime": 655.5852,
|
| 278 |
+
"eval_samples_per_second": 4.269,
|
| 279 |
+
"eval_steps_per_second": 0.267,
|
| 280 |
+
"learning_rate": 0.01,
|
| 281 |
+
"step": 7504
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 14.93,
|
| 285 |
+
"learning_rate": 0.01,
|
| 286 |
+
"loss": 0.3104,
|
| 287 |
+
"step": 8000
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 15.0,
|
| 291 |
+
"eval_accuracy": 0.0,
|
| 292 |
+
"eval_f1_macro": 0.06737263276203229,
|
| 293 |
+
"eval_f1_micro": 0.16727748691099475,
|
| 294 |
+
"eval_loss": 3.1991209983825684,
|
| 295 |
+
"eval_roc_auc": 0.48793717022406635,
|
| 296 |
+
"eval_runtime": 660.6942,
|
| 297 |
+
"eval_samples_per_second": 4.236,
|
| 298 |
+
"eval_steps_per_second": 0.265,
|
| 299 |
+
"learning_rate": 0.01,
|
| 300 |
+
"step": 8040
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"epoch": 15.86,
|
| 304 |
+
"learning_rate": 0.001,
|
| 305 |
+
"loss": 0.3042,
|
| 306 |
+
"step": 8500
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 16.0,
|
| 310 |
+
"eval_accuracy": 0.15827081100392998,
|
| 311 |
+
"eval_f1_macro": 0.2746423799349764,
|
| 312 |
+
"eval_f1_micro": 0.45441654850279417,
|
| 313 |
+
"eval_loss": 0.2820233702659607,
|
| 314 |
+
"eval_roc_auc": 0.6518577907371156,
|
| 315 |
+
"eval_runtime": 657.3215,
|
| 316 |
+
"eval_samples_per_second": 4.258,
|
| 317 |
+
"eval_steps_per_second": 0.266,
|
| 318 |
+
"learning_rate": 0.001,
|
| 319 |
+
"step": 8576
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 16.79,
|
| 323 |
+
"learning_rate": 0.001,
|
| 324 |
+
"loss": 0.2788,
|
| 325 |
+
"step": 9000
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 17.0,
|
| 329 |
+
"eval_accuracy": 0.1639871382636656,
|
| 330 |
+
"eval_f1_macro": 0.3842369260673032,
|
| 331 |
+
"eval_f1_micro": 0.5743618993298161,
|
| 332 |
+
"eval_loss": 0.2741135358810425,
|
| 333 |
+
"eval_roc_auc": 0.7204845597401146,
|
| 334 |
+
"eval_runtime": 676.3087,
|
| 335 |
+
"eval_samples_per_second": 4.139,
|
| 336 |
+
"eval_steps_per_second": 0.259,
|
| 337 |
+
"learning_rate": 0.001,
|
| 338 |
+
"step": 9112
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 17.72,
|
| 342 |
+
"learning_rate": 0.001,
|
| 343 |
+
"loss": 0.2724,
|
| 344 |
+
"step": 9500
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 18.0,
|
| 348 |
+
"eval_accuracy": 0.2072168631654162,
|
| 349 |
+
"eval_f1_macro": 0.39364174946303265,
|
| 350 |
+
"eval_f1_micro": 0.5903019744483159,
|
| 351 |
+
"eval_loss": 0.2424442023038864,
|
| 352 |
+
"eval_roc_auc": 0.725566189869564,
|
| 353 |
+
"eval_runtime": 676.7446,
|
| 354 |
+
"eval_samples_per_second": 4.136,
|
| 355 |
+
"eval_steps_per_second": 0.259,
|
| 356 |
+
"learning_rate": 0.001,
|
| 357 |
+
"step": 9648
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 18.66,
|
| 361 |
+
"learning_rate": 0.001,
|
| 362 |
+
"loss": 0.2642,
|
| 363 |
+
"step": 10000
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 19.0,
|
| 367 |
+
"eval_accuracy": 0.21864951768488747,
|
| 368 |
+
"eval_f1_macro": 0.4095283624260902,
|
| 369 |
+
"eval_f1_micro": 0.6020972084455152,
|
| 370 |
+
"eval_loss": 0.24139995872974396,
|
| 371 |
+
"eval_roc_auc": 0.7346777613512268,
|
| 372 |
+
"eval_runtime": 679.8261,
|
| 373 |
+
"eval_samples_per_second": 4.117,
|
| 374 |
+
"eval_steps_per_second": 0.257,
|
| 375 |
+
"learning_rate": 0.001,
|
| 376 |
+
"step": 10184
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 19.59,
|
| 380 |
+
"learning_rate": 0.001,
|
| 381 |
+
"loss": 0.2597,
|
| 382 |
+
"step": 10500
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 20.0,
|
| 386 |
+
"eval_accuracy": 0.22508038585209003,
|
| 387 |
+
"eval_f1_macro": 0.4156043928581225,
|
| 388 |
+
"eval_f1_micro": 0.6078926484380637,
|
| 389 |
+
"eval_loss": 0.22694532573223114,
|
| 390 |
+
"eval_roc_auc": 0.734743268508761,
|
| 391 |
+
"eval_runtime": 658.6647,
|
| 392 |
+
"eval_samples_per_second": 4.25,
|
| 393 |
+
"eval_steps_per_second": 0.266,
|
| 394 |
+
"learning_rate": 0.001,
|
| 395 |
+
"step": 10720
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 20.52,
|
| 399 |
+
"learning_rate": 0.001,
|
| 400 |
+
"loss": 0.2575,
|
| 401 |
+
"step": 11000
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 21.0,
|
| 405 |
+
"eval_accuracy": 0.23401214719542693,
|
| 406 |
+
"eval_f1_macro": 0.42525523109884183,
|
| 407 |
+
"eval_f1_micro": 0.6231395675372087,
|
| 408 |
+
"eval_loss": 0.22489750385284424,
|
| 409 |
+
"eval_roc_auc": 0.7462705119874329,
|
| 410 |
+
"eval_runtime": 659.2031,
|
| 411 |
+
"eval_samples_per_second": 4.246,
|
| 412 |
+
"eval_steps_per_second": 0.265,
|
| 413 |
+
"learning_rate": 0.001,
|
| 414 |
+
"step": 11256
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 21.46,
|
| 418 |
+
"learning_rate": 0.001,
|
| 419 |
+
"loss": 0.253,
|
| 420 |
+
"step": 11500
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 22.0,
|
| 424 |
+
"eval_accuracy": 0.24294390853876385,
|
| 425 |
+
"eval_f1_macro": 0.4638642250813612,
|
| 426 |
+
"eval_f1_micro": 0.6291112183686134,
|
| 427 |
+
"eval_loss": 0.2261122316122055,
|
| 428 |
+
"eval_roc_auc": 0.7521458927635197,
|
| 429 |
+
"eval_runtime": 657.2132,
|
| 430 |
+
"eval_samples_per_second": 4.259,
|
| 431 |
+
"eval_steps_per_second": 0.266,
|
| 432 |
+
"learning_rate": 0.001,
|
| 433 |
+
"step": 11792
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 22.39,
|
| 437 |
+
"learning_rate": 0.001,
|
| 438 |
+
"loss": 0.2491,
|
| 439 |
+
"step": 12000
|
| 440 |
+
},
|
| 441 |
+
{
|
| 442 |
+
"epoch": 23.0,
|
| 443 |
+
"eval_accuracy": 0.2536620221507681,
|
| 444 |
+
"eval_f1_macro": 0.4855675958097963,
|
| 445 |
+
"eval_f1_micro": 0.6453674121405751,
|
| 446 |
+
"eval_loss": 0.21625448763370514,
|
| 447 |
+
"eval_roc_auc": 0.7626888699414743,
|
| 448 |
+
"eval_runtime": 684.5667,
|
| 449 |
+
"eval_samples_per_second": 4.089,
|
| 450 |
+
"eval_steps_per_second": 0.256,
|
| 451 |
+
"learning_rate": 0.001,
|
| 452 |
+
"step": 12328
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 23.32,
|
| 456 |
+
"learning_rate": 0.001,
|
| 457 |
+
"loss": 0.2484,
|
| 458 |
+
"step": 12500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 24.0,
|
| 462 |
+
"eval_accuracy": 0.2568774562343694,
|
| 463 |
+
"eval_f1_macro": 0.4635023998508455,
|
| 464 |
+
"eval_f1_micro": 0.6261708770933863,
|
| 465 |
+
"eval_loss": 0.22124029695987701,
|
| 466 |
+
"eval_roc_auc": 0.746011775710583,
|
| 467 |
+
"eval_runtime": 686.8149,
|
| 468 |
+
"eval_samples_per_second": 4.075,
|
| 469 |
+
"eval_steps_per_second": 0.255,
|
| 470 |
+
"learning_rate": 0.001,
|
| 471 |
+
"step": 12864
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 24.25,
|
| 475 |
+
"learning_rate": 0.001,
|
| 476 |
+
"loss": 0.2465,
|
| 477 |
+
"step": 13000
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 25.0,
|
| 481 |
+
"eval_accuracy": 0.27724187209717754,
|
| 482 |
+
"eval_f1_macro": 0.4780125036155298,
|
| 483 |
+
"eval_f1_micro": 0.6486374896779521,
|
| 484 |
+
"eval_loss": 0.21179212629795074,
|
| 485 |
+
"eval_roc_auc": 0.7621535864293458,
|
| 486 |
+
"eval_runtime": 682.8306,
|
| 487 |
+
"eval_samples_per_second": 4.099,
|
| 488 |
+
"eval_steps_per_second": 0.256,
|
| 489 |
+
"learning_rate": 0.001,
|
| 490 |
+
"step": 13400
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 25.19,
|
| 494 |
+
"learning_rate": 0.001,
|
| 495 |
+
"loss": 0.241,
|
| 496 |
+
"step": 13500
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 26.0,
|
| 500 |
+
"eval_accuracy": 0.255805644873169,
|
| 501 |
+
"eval_f1_macro": 0.5159201891126339,
|
| 502 |
+
"eval_f1_micro": 0.6601850610164945,
|
| 503 |
+
"eval_loss": 0.21056728065013885,
|
| 504 |
+
"eval_roc_auc": 0.7726679521523573,
|
| 505 |
+
"eval_runtime": 690.8376,
|
| 506 |
+
"eval_samples_per_second": 4.052,
|
| 507 |
+
"eval_steps_per_second": 0.253,
|
| 508 |
+
"learning_rate": 0.001,
|
| 509 |
+
"step": 13936
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 26.12,
|
| 513 |
+
"learning_rate": 0.001,
|
| 514 |
+
"loss": 0.2413,
|
| 515 |
+
"step": 14000
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 27.0,
|
| 519 |
+
"eval_accuracy": 0.2722400857449089,
|
| 520 |
+
"eval_f1_macro": 0.4978739205590764,
|
| 521 |
+
"eval_f1_micro": 0.6389924716808555,
|
| 522 |
+
"eval_loss": 0.21349740028381348,
|
| 523 |
+
"eval_roc_auc": 0.753602824058608,
|
| 524 |
+
"eval_runtime": 690.4835,
|
| 525 |
+
"eval_samples_per_second": 4.054,
|
| 526 |
+
"eval_steps_per_second": 0.253,
|
| 527 |
+
"learning_rate": 0.001,
|
| 528 |
+
"step": 14472
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 27.05,
|
| 532 |
+
"learning_rate": 0.001,
|
| 533 |
+
"loss": 0.2401,
|
| 534 |
+
"step": 14500
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 27.99,
|
| 538 |
+
"learning_rate": 0.001,
|
| 539 |
+
"loss": 0.2385,
|
| 540 |
+
"step": 15000
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 28.0,
|
| 544 |
+
"eval_accuracy": 0.27724187209717754,
|
| 545 |
+
"eval_f1_macro": 0.4596468154208962,
|
| 546 |
+
"eval_f1_micro": 0.6102875778238955,
|
| 547 |
+
"eval_loss": 0.2182074636220932,
|
| 548 |
+
"eval_roc_auc": 0.7318729802955509,
|
| 549 |
+
"eval_runtime": 665.6085,
|
| 550 |
+
"eval_samples_per_second": 4.205,
|
| 551 |
+
"eval_steps_per_second": 0.263,
|
| 552 |
+
"learning_rate": 0.001,
|
| 553 |
+
"step": 15008
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 28.92,
|
| 557 |
+
"learning_rate": 0.001,
|
| 558 |
+
"loss": 0.2366,
|
| 559 |
+
"step": 15500
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 29.0,
|
| 563 |
+
"eval_accuracy": 0.270811003929975,
|
| 564 |
+
"eval_f1_macro": 0.5354446223521995,
|
| 565 |
+
"eval_f1_micro": 0.6614600569197168,
|
| 566 |
+
"eval_loss": 0.2131979614496231,
|
| 567 |
+
"eval_roc_auc": 0.7757714197210208,
|
| 568 |
+
"eval_runtime": 663.3734,
|
| 569 |
+
"eval_samples_per_second": 4.219,
|
| 570 |
+
"eval_steps_per_second": 0.264,
|
| 571 |
+
"learning_rate": 0.001,
|
| 572 |
+
"step": 15544
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 29.85,
|
| 576 |
+
"learning_rate": 0.001,
|
| 577 |
+
"loss": 0.2345,
|
| 578 |
+
"step": 16000
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 30.0,
|
| 582 |
+
"eval_accuracy": 0.2747409789210432,
|
| 583 |
+
"eval_f1_macro": 0.5121854108049624,
|
| 584 |
+
"eval_f1_micro": 0.6565579984836998,
|
| 585 |
+
"eval_loss": 0.20685863494873047,
|
| 586 |
+
"eval_roc_auc": 0.7657921716163675,
|
| 587 |
+
"eval_runtime": 663.8358,
|
| 588 |
+
"eval_samples_per_second": 4.216,
|
| 589 |
+
"eval_steps_per_second": 0.264,
|
| 590 |
+
"learning_rate": 0.001,
|
| 591 |
+
"step": 16080
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 30.0,
|
| 595 |
+
"learning_rate": 0.001,
|
| 596 |
+
"step": 16080,
|
| 597 |
+
"total_flos": 2.04446834386654e+19,
|
| 598 |
+
"train_loss": 0.2889823566028728,
|
| 599 |
+
"train_runtime": 84191.1285,
|
| 600 |
+
"train_samples_per_second": 3.053,
|
| 601 |
+
"train_steps_per_second": 0.191
|
| 602 |
+
}
|
| 603 |
+
],
|
| 604 |
+
"logging_steps": 500,
|
| 605 |
+
"max_steps": 16080,
|
| 606 |
+
"num_train_epochs": 30,
|
| 607 |
+
"save_steps": 500,
|
| 608 |
+
"total_flos": 2.04446834386654e+19,
|
| 609 |
+
"trial_name": null,
|
| 610 |
+
"trial_params": null
|
| 611 |
+
}
|