Model save
Browse files- README.md +13 -9
- all_results.json +7 -7
- test_results.json +7 -7
- trainer_state.json +1377 -70
- val_results.json +7 -7
README.md
CHANGED
|
@@ -18,9 +18,9 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 18 |
|
| 19 |
This model is a fine-tuned version of [MCG-NJU/videomae-base](https://huggingface.co/MCG-NJU/videomae-base) on an unknown dataset.
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
-
- Loss:
|
| 22 |
-
- Accuracy: 0.
|
| 23 |
-
- F1: 0.
|
| 24 |
|
| 25 |
## Model description
|
| 26 |
|
|
@@ -40,19 +40,23 @@ More information needed
|
|
| 40 |
|
| 41 |
The following hyperparameters were used during training:
|
| 42 |
- learning_rate: 5e-05
|
| 43 |
-
- train_batch_size:
|
| 44 |
-
- eval_batch_size:
|
| 45 |
- seed: 42
|
| 46 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 47 |
- lr_scheduler_type: linear
|
| 48 |
- lr_scheduler_warmup_ratio: 0.1
|
| 49 |
-
- training_steps:
|
| 50 |
|
| 51 |
### Training results
|
| 52 |
|
| 53 |
-
| Training Loss | Epoch
|
| 54 |
-
|
| 55 |
-
| 1.
|
|
|
|
|
|
|
|
|
|
|
|
|
| 56 |
|
| 57 |
|
| 58 |
### Framework versions
|
|
|
|
| 18 |
|
| 19 |
This model is a fine-tuned version of [MCG-NJU/videomae-base](https://huggingface.co/MCG-NJU/videomae-base) on an unknown dataset.
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 0.5544
|
| 22 |
+
- Accuracy: 0.8852
|
| 23 |
+
- F1: 0.8840
|
| 24 |
|
| 25 |
## Model description
|
| 26 |
|
|
|
|
| 40 |
|
| 41 |
The following hyperparameters were used during training:
|
| 42 |
- learning_rate: 5e-05
|
| 43 |
+
- train_batch_size: 2
|
| 44 |
+
- eval_batch_size: 2
|
| 45 |
- seed: 42
|
| 46 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 47 |
- lr_scheduler_type: linear
|
| 48 |
- lr_scheduler_warmup_ratio: 0.1
|
| 49 |
+
- training_steps: 1850
|
| 50 |
|
| 51 |
### Training results
|
| 52 |
|
| 53 |
+
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|
| 54 |
+
|:-------------:|:------:|:----:|:---------------:|:--------:|:------:|
|
| 55 |
+
| 1.3096 | 0.2005 | 371 | 1.3267 | 0.4590 | 0.2888 |
|
| 56 |
+
| 1.0586 | 1.2005 | 742 | 1.2866 | 0.5820 | 0.5035 |
|
| 57 |
+
| 0.9781 | 2.2005 | 1113 | 0.7952 | 0.7459 | 0.7466 |
|
| 58 |
+
| 0.0034 | 3.2005 | 1484 | 0.7218 | 0.8361 | 0.8343 |
|
| 59 |
+
| 0.1895 | 4.1978 | 1850 | 0.5544 | 0.8852 | 0.8840 |
|
| 60 |
|
| 61 |
|
| 62 |
### Framework versions
|
all_results.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_f1": 0.
|
| 5 |
-
"eval_loss":
|
| 6 |
-
"eval_runtime":
|
| 7 |
-
"eval_samples_per_second":
|
| 8 |
-
"eval_steps_per_second":
|
| 9 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 4.197837837837838,
|
| 3 |
+
"eval_accuracy": 0.8852459016393442,
|
| 4 |
+
"eval_f1": 0.8839858413628905,
|
| 5 |
+
"eval_loss": 0.5544241070747375,
|
| 6 |
+
"eval_runtime": 784.0148,
|
| 7 |
+
"eval_samples_per_second": 0.156,
|
| 8 |
+
"eval_steps_per_second": 0.078
|
| 9 |
}
|
test_results.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_f1": 0.
|
| 5 |
-
"eval_loss":
|
| 6 |
-
"eval_runtime":
|
| 7 |
-
"eval_samples_per_second":
|
| 8 |
-
"eval_steps_per_second":
|
| 9 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 4.197837837837838,
|
| 3 |
+
"eval_accuracy": 0.851528384279476,
|
| 4 |
+
"eval_f1": 0.8496851913824908,
|
| 5 |
+
"eval_loss": 0.7380829453468323,
|
| 6 |
+
"eval_runtime": 1382.3283,
|
| 7 |
+
"eval_samples_per_second": 0.166,
|
| 8 |
+
"eval_steps_per_second": 0.083
|
| 9 |
}
|
trainer_state.json
CHANGED
|
@@ -1,93 +1,1400 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "videomae-surf-analytics-sans-wandb\\checkpoint-
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 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.
|
| 13 |
-
"grad_norm":
|
| 14 |
-
"learning_rate":
|
| 15 |
-
"loss": 1.
|
| 16 |
"step": 10
|
| 17 |
},
|
| 18 |
{
|
| 19 |
-
"epoch": 0.
|
| 20 |
-
"grad_norm":
|
| 21 |
-
"learning_rate":
|
| 22 |
-
"loss": 1.
|
| 23 |
"step": 20
|
| 24 |
},
|
| 25 |
{
|
| 26 |
-
"epoch": 0.
|
| 27 |
-
"grad_norm":
|
| 28 |
-
"learning_rate":
|
| 29 |
-
"loss": 1.
|
| 30 |
"step": 30
|
| 31 |
},
|
| 32 |
{
|
| 33 |
-
"epoch": 0.
|
| 34 |
-
"grad_norm":
|
| 35 |
-
"learning_rate":
|
| 36 |
-
"loss": 1.
|
| 37 |
"step": 40
|
| 38 |
},
|
| 39 |
{
|
| 40 |
-
"epoch":
|
| 41 |
-
"
|
| 42 |
-
"
|
| 43 |
-
"
|
| 44 |
-
"
|
| 45 |
-
|
| 46 |
-
|
| 47 |
-
"
|
| 48 |
-
|
| 49 |
-
|
| 50 |
-
"
|
| 51 |
-
"step":
|
| 52 |
-
|
| 53 |
-
|
| 54 |
-
"
|
| 55 |
-
"
|
| 56 |
-
"
|
| 57 |
-
|
| 58 |
-
|
| 59 |
-
|
| 60 |
-
|
| 61 |
-
"
|
| 62 |
-
"
|
| 63 |
-
"
|
| 64 |
-
"
|
| 65 |
-
"
|
| 66 |
-
|
| 67 |
-
|
| 68 |
-
|
| 69 |
-
"
|
| 70 |
-
"
|
| 71 |
-
"
|
| 72 |
-
"
|
| 73 |
-
|
| 74 |
-
|
| 75 |
-
"
|
| 76 |
-
"
|
| 77 |
-
|
| 78 |
-
|
| 79 |
-
"
|
| 80 |
-
|
| 81 |
-
|
| 82 |
-
"
|
| 83 |
-
"
|
| 84 |
-
"
|
| 85 |
-
"
|
| 86 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 87 |
}
|
| 88 |
],
|
| 89 |
"logging_steps": 10,
|
| 90 |
-
"max_steps":
|
| 91 |
"num_input_tokens_seen": 0,
|
| 92 |
"num_train_epochs": 9223372036854775807,
|
| 93 |
"save_steps": 500,
|
|
@@ -103,8 +1410,8 @@
|
|
| 103 |
"attributes": {}
|
| 104 |
}
|
| 105 |
},
|
| 106 |
-
"total_flos":
|
| 107 |
-
"train_batch_size":
|
| 108 |
"trial_name": null,
|
| 109 |
"trial_params": null
|
| 110 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.8839858413628905,
|
| 3 |
+
"best_model_checkpoint": "videomae-surf-analytics-sans-wandb\\checkpoint-1850",
|
| 4 |
+
"epoch": 4.197837837837838,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 1850,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
+
"epoch": 0.005405405405405406,
|
| 13 |
+
"grad_norm": 11.531332969665527,
|
| 14 |
+
"learning_rate": 2.702702702702703e-06,
|
| 15 |
+
"loss": 1.4317,
|
| 16 |
"step": 10
|
| 17 |
},
|
| 18 |
{
|
| 19 |
+
"epoch": 0.010810810810810811,
|
| 20 |
+
"grad_norm": 11.917078971862793,
|
| 21 |
+
"learning_rate": 5.405405405405406e-06,
|
| 22 |
+
"loss": 1.3912,
|
| 23 |
"step": 20
|
| 24 |
},
|
| 25 |
{
|
| 26 |
+
"epoch": 0.016216216216216217,
|
| 27 |
+
"grad_norm": 10.456705093383789,
|
| 28 |
+
"learning_rate": 8.108108108108109e-06,
|
| 29 |
+
"loss": 1.3411,
|
| 30 |
"step": 30
|
| 31 |
},
|
| 32 |
{
|
| 33 |
+
"epoch": 0.021621621621621623,
|
| 34 |
+
"grad_norm": 10.632660865783691,
|
| 35 |
+
"learning_rate": 1.0810810810810812e-05,
|
| 36 |
+
"loss": 1.2104,
|
| 37 |
"step": 40
|
| 38 |
},
|
| 39 |
{
|
| 40 |
+
"epoch": 0.02702702702702703,
|
| 41 |
+
"grad_norm": 10.985200881958008,
|
| 42 |
+
"learning_rate": 1.3513513513513515e-05,
|
| 43 |
+
"loss": 1.4311,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.032432432432432434,
|
| 48 |
+
"grad_norm": 12.063570976257324,
|
| 49 |
+
"learning_rate": 1.6216216216216218e-05,
|
| 50 |
+
"loss": 0.9259,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.03783783783783784,
|
| 55 |
+
"grad_norm": 8.3969144821167,
|
| 56 |
+
"learning_rate": 1.891891891891892e-05,
|
| 57 |
+
"loss": 1.2204,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.043243243243243246,
|
| 62 |
+
"grad_norm": 13.087193489074707,
|
| 63 |
+
"learning_rate": 2.1621621621621624e-05,
|
| 64 |
+
"loss": 1.215,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.04864864864864865,
|
| 69 |
+
"grad_norm": 12.476021766662598,
|
| 70 |
+
"learning_rate": 2.4324324324324327e-05,
|
| 71 |
+
"loss": 1.4883,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.05405405405405406,
|
| 76 |
+
"grad_norm": 10.030627250671387,
|
| 77 |
+
"learning_rate": 2.702702702702703e-05,
|
| 78 |
+
"loss": 1.2363,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.05945945945945946,
|
| 83 |
+
"grad_norm": 15.547861099243164,
|
| 84 |
+
"learning_rate": 2.9729729729729733e-05,
|
| 85 |
+
"loss": 1.4148,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.06486486486486487,
|
| 90 |
+
"grad_norm": 11.590038299560547,
|
| 91 |
+
"learning_rate": 3.2432432432432436e-05,
|
| 92 |
+
"loss": 1.3528,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.07027027027027027,
|
| 97 |
+
"grad_norm": 12.326528549194336,
|
| 98 |
+
"learning_rate": 3.513513513513514e-05,
|
| 99 |
+
"loss": 1.2941,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.07567567567567568,
|
| 104 |
+
"grad_norm": 8.991989135742188,
|
| 105 |
+
"learning_rate": 3.783783783783784e-05,
|
| 106 |
+
"loss": 1.1322,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.08108108108108109,
|
| 111 |
+
"grad_norm": 17.619140625,
|
| 112 |
+
"learning_rate": 4.0540540540540545e-05,
|
| 113 |
+
"loss": 1.2129,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.08648648648648649,
|
| 118 |
+
"grad_norm": 11.91755485534668,
|
| 119 |
+
"learning_rate": 4.324324324324325e-05,
|
| 120 |
+
"loss": 1.2466,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.0918918918918919,
|
| 125 |
+
"grad_norm": 9.150936126708984,
|
| 126 |
+
"learning_rate": 4.594594594594595e-05,
|
| 127 |
+
"loss": 1.2624,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.0972972972972973,
|
| 132 |
+
"grad_norm": 12.12965202331543,
|
| 133 |
+
"learning_rate": 4.8648648648648654e-05,
|
| 134 |
+
"loss": 1.2423,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.10270270270270271,
|
| 139 |
+
"grad_norm": 11.131547927856445,
|
| 140 |
+
"learning_rate": 4.984984984984985e-05,
|
| 141 |
+
"loss": 1.0887,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.10810810810810811,
|
| 146 |
+
"grad_norm": 16.159334182739258,
|
| 147 |
+
"learning_rate": 4.954954954954955e-05,
|
| 148 |
+
"loss": 1.2204,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.11351351351351352,
|
| 153 |
+
"grad_norm": 14.13427734375,
|
| 154 |
+
"learning_rate": 4.9249249249249253e-05,
|
| 155 |
+
"loss": 1.2002,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.11891891891891893,
|
| 160 |
+
"grad_norm": 6.860474586486816,
|
| 161 |
+
"learning_rate": 4.8948948948948954e-05,
|
| 162 |
+
"loss": 1.6223,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.12432432432432433,
|
| 167 |
+
"grad_norm": 17.881906509399414,
|
| 168 |
+
"learning_rate": 4.8648648648648654e-05,
|
| 169 |
+
"loss": 1.3443,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.12972972972972974,
|
| 174 |
+
"grad_norm": 9.10927963256836,
|
| 175 |
+
"learning_rate": 4.834834834834835e-05,
|
| 176 |
+
"loss": 1.4485,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.13513513513513514,
|
| 181 |
+
"grad_norm": 6.366971015930176,
|
| 182 |
+
"learning_rate": 4.804804804804805e-05,
|
| 183 |
+
"loss": 0.9077,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.14054054054054055,
|
| 188 |
+
"grad_norm": 22.29749870300293,
|
| 189 |
+
"learning_rate": 4.774774774774775e-05,
|
| 190 |
+
"loss": 1.473,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.14594594594594595,
|
| 195 |
+
"grad_norm": 19.388036727905273,
|
| 196 |
+
"learning_rate": 4.744744744744745e-05,
|
| 197 |
+
"loss": 1.1213,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.15135135135135136,
|
| 202 |
+
"grad_norm": 6.246539115905762,
|
| 203 |
+
"learning_rate": 4.714714714714715e-05,
|
| 204 |
+
"loss": 0.7922,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.15675675675675677,
|
| 209 |
+
"grad_norm": 8.21817684173584,
|
| 210 |
+
"learning_rate": 4.684684684684685e-05,
|
| 211 |
+
"loss": 1.2757,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.16216216216216217,
|
| 216 |
+
"grad_norm": 18.414051055908203,
|
| 217 |
+
"learning_rate": 4.654654654654655e-05,
|
| 218 |
+
"loss": 1.3382,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.16756756756756758,
|
| 223 |
+
"grad_norm": 11.922811508178711,
|
| 224 |
+
"learning_rate": 4.624624624624625e-05,
|
| 225 |
+
"loss": 0.9546,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.17297297297297298,
|
| 230 |
+
"grad_norm": 15.078409194946289,
|
| 231 |
+
"learning_rate": 4.594594594594595e-05,
|
| 232 |
+
"loss": 1.7355,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.1783783783783784,
|
| 237 |
+
"grad_norm": 10.318249702453613,
|
| 238 |
+
"learning_rate": 4.5645645645645645e-05,
|
| 239 |
+
"loss": 1.2742,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.1837837837837838,
|
| 244 |
+
"grad_norm": 12.838833808898926,
|
| 245 |
+
"learning_rate": 4.5345345345345345e-05,
|
| 246 |
+
"loss": 1.5675,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.1891891891891892,
|
| 251 |
+
"grad_norm": 12.560096740722656,
|
| 252 |
+
"learning_rate": 4.5045045045045046e-05,
|
| 253 |
+
"loss": 1.3238,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.1945945945945946,
|
| 258 |
+
"grad_norm": 3.9458367824554443,
|
| 259 |
+
"learning_rate": 4.4744744744744746e-05,
|
| 260 |
+
"loss": 1.4337,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.2,
|
| 265 |
+
"grad_norm": 14.40661334991455,
|
| 266 |
+
"learning_rate": 4.4444444444444447e-05,
|
| 267 |
+
"loss": 1.3096,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.20054054054054055,
|
| 272 |
+
"eval_accuracy": 0.45901639344262296,
|
| 273 |
+
"eval_f1": 0.28881930373917847,
|
| 274 |
+
"eval_loss": 1.3266738653182983,
|
| 275 |
+
"eval_runtime": 589.7403,
|
| 276 |
+
"eval_samples_per_second": 0.207,
|
| 277 |
+
"eval_steps_per_second": 0.103,
|
| 278 |
+
"step": 371
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 1.0048648648648648,
|
| 282 |
+
"grad_norm": 11.33437728881836,
|
| 283 |
+
"learning_rate": 4.414414414414415e-05,
|
| 284 |
+
"loss": 1.1584,
|
| 285 |
+
"step": 380
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 1.0102702702702702,
|
| 289 |
+
"grad_norm": 12.032054901123047,
|
| 290 |
+
"learning_rate": 4.384384384384385e-05,
|
| 291 |
+
"loss": 1.3792,
|
| 292 |
+
"step": 390
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 1.0156756756756757,
|
| 296 |
+
"grad_norm": 10.13355541229248,
|
| 297 |
+
"learning_rate": 4.354354354354355e-05,
|
| 298 |
+
"loss": 1.3946,
|
| 299 |
+
"step": 400
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 1.021081081081081,
|
| 303 |
+
"grad_norm": 8.573601722717285,
|
| 304 |
+
"learning_rate": 4.324324324324325e-05,
|
| 305 |
+
"loss": 1.4136,
|
| 306 |
+
"step": 410
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 1.0264864864864864,
|
| 310 |
+
"grad_norm": 6.312205791473389,
|
| 311 |
+
"learning_rate": 4.294294294294294e-05,
|
| 312 |
+
"loss": 1.3542,
|
| 313 |
+
"step": 420
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 1.0318918918918918,
|
| 317 |
+
"grad_norm": 9.014967918395996,
|
| 318 |
+
"learning_rate": 4.264264264264264e-05,
|
| 319 |
+
"loss": 1.3646,
|
| 320 |
+
"step": 430
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 1.0372972972972974,
|
| 324 |
+
"grad_norm": 8.805109977722168,
|
| 325 |
+
"learning_rate": 4.234234234234234e-05,
|
| 326 |
+
"loss": 1.2612,
|
| 327 |
+
"step": 440
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 1.0427027027027027,
|
| 331 |
+
"grad_norm": 5.080739974975586,
|
| 332 |
+
"learning_rate": 4.204204204204204e-05,
|
| 333 |
+
"loss": 1.1235,
|
| 334 |
+
"step": 450
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 1.048108108108108,
|
| 338 |
+
"grad_norm": 9.063738822937012,
|
| 339 |
+
"learning_rate": 4.1741741741741744e-05,
|
| 340 |
+
"loss": 1.1287,
|
| 341 |
+
"step": 460
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 1.0535135135135134,
|
| 345 |
+
"grad_norm": 10.316679954528809,
|
| 346 |
+
"learning_rate": 4.1441441441441444e-05,
|
| 347 |
+
"loss": 1.043,
|
| 348 |
+
"step": 470
|
| 349 |
+
},
|
| 350 |
+
{
|
| 351 |
+
"epoch": 1.058918918918919,
|
| 352 |
+
"grad_norm": 7.451129913330078,
|
| 353 |
+
"learning_rate": 4.1141141141141144e-05,
|
| 354 |
+
"loss": 1.1908,
|
| 355 |
+
"step": 480
|
| 356 |
+
},
|
| 357 |
+
{
|
| 358 |
+
"epoch": 1.0643243243243243,
|
| 359 |
+
"grad_norm": 4.216222286224365,
|
| 360 |
+
"learning_rate": 4.0840840840840845e-05,
|
| 361 |
+
"loss": 1.132,
|
| 362 |
+
"step": 490
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 1.0697297297297297,
|
| 366 |
+
"grad_norm": 9.235713005065918,
|
| 367 |
+
"learning_rate": 4.0540540540540545e-05,
|
| 368 |
+
"loss": 1.7454,
|
| 369 |
+
"step": 500
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 1.075135135135135,
|
| 373 |
+
"grad_norm": 12.876686096191406,
|
| 374 |
+
"learning_rate": 4.024024024024024e-05,
|
| 375 |
+
"loss": 1.0389,
|
| 376 |
+
"step": 510
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 1.0805405405405406,
|
| 380 |
+
"grad_norm": 9.994712829589844,
|
| 381 |
+
"learning_rate": 3.993993993993994e-05,
|
| 382 |
+
"loss": 1.0837,
|
| 383 |
+
"step": 520
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 1.085945945945946,
|
| 387 |
+
"grad_norm": 15.395247459411621,
|
| 388 |
+
"learning_rate": 3.963963963963964e-05,
|
| 389 |
+
"loss": 1.4464,
|
| 390 |
+
"step": 530
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 1.0913513513513513,
|
| 394 |
+
"grad_norm": 8.740015029907227,
|
| 395 |
+
"learning_rate": 3.933933933933934e-05,
|
| 396 |
+
"loss": 1.0569,
|
| 397 |
+
"step": 540
|
| 398 |
+
},
|
| 399 |
+
{
|
| 400 |
+
"epoch": 1.0967567567567567,
|
| 401 |
+
"grad_norm": 6.516251087188721,
|
| 402 |
+
"learning_rate": 3.903903903903904e-05,
|
| 403 |
+
"loss": 1.2919,
|
| 404 |
+
"step": 550
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 1.1021621621621622,
|
| 408 |
+
"grad_norm": 9.043025970458984,
|
| 409 |
+
"learning_rate": 3.873873873873874e-05,
|
| 410 |
+
"loss": 1.2064,
|
| 411 |
+
"step": 560
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 1.1075675675675676,
|
| 415 |
+
"grad_norm": 7.937928199768066,
|
| 416 |
+
"learning_rate": 3.843843843843844e-05,
|
| 417 |
+
"loss": 1.1334,
|
| 418 |
+
"step": 570
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 1.112972972972973,
|
| 422 |
+
"grad_norm": 7.148234844207764,
|
| 423 |
+
"learning_rate": 3.813813813813814e-05,
|
| 424 |
+
"loss": 1.221,
|
| 425 |
+
"step": 580
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 1.1183783783783783,
|
| 429 |
+
"grad_norm": 18.380756378173828,
|
| 430 |
+
"learning_rate": 3.783783783783784e-05,
|
| 431 |
+
"loss": 1.0975,
|
| 432 |
+
"step": 590
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 1.1237837837837839,
|
| 436 |
+
"grad_norm": 14.88819408416748,
|
| 437 |
+
"learning_rate": 3.7537537537537536e-05,
|
| 438 |
+
"loss": 0.7921,
|
| 439 |
+
"step": 600
|
| 440 |
+
},
|
| 441 |
+
{
|
| 442 |
+
"epoch": 1.1291891891891892,
|
| 443 |
+
"grad_norm": 29.024171829223633,
|
| 444 |
+
"learning_rate": 3.7237237237237236e-05,
|
| 445 |
+
"loss": 1.1379,
|
| 446 |
+
"step": 610
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 1.1345945945945946,
|
| 450 |
+
"grad_norm": 17.080385208129883,
|
| 451 |
+
"learning_rate": 3.693693693693694e-05,
|
| 452 |
+
"loss": 0.7285,
|
| 453 |
+
"step": 620
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 1.1400000000000001,
|
| 457 |
+
"grad_norm": 5.5788187980651855,
|
| 458 |
+
"learning_rate": 3.663663663663664e-05,
|
| 459 |
+
"loss": 1.2987,
|
| 460 |
+
"step": 630
|
| 461 |
+
},
|
| 462 |
+
{
|
| 463 |
+
"epoch": 1.1454054054054055,
|
| 464 |
+
"grad_norm": 8.62403392791748,
|
| 465 |
+
"learning_rate": 3.633633633633634e-05,
|
| 466 |
+
"loss": 1.162,
|
| 467 |
+
"step": 640
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 1.1508108108108108,
|
| 471 |
+
"grad_norm": 13.667816162109375,
|
| 472 |
+
"learning_rate": 3.603603603603604e-05,
|
| 473 |
+
"loss": 1.1039,
|
| 474 |
+
"step": 650
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 1.1562162162162162,
|
| 478 |
+
"grad_norm": 15.54757022857666,
|
| 479 |
+
"learning_rate": 3.573573573573574e-05,
|
| 480 |
+
"loss": 1.1578,
|
| 481 |
+
"step": 660
|
| 482 |
+
},
|
| 483 |
+
{
|
| 484 |
+
"epoch": 1.1616216216216215,
|
| 485 |
+
"grad_norm": 21.381877899169922,
|
| 486 |
+
"learning_rate": 3.543543543543544e-05,
|
| 487 |
+
"loss": 1.1655,
|
| 488 |
+
"step": 670
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 1.167027027027027,
|
| 492 |
+
"grad_norm": 16.7823429107666,
|
| 493 |
+
"learning_rate": 3.513513513513514e-05,
|
| 494 |
+
"loss": 1.2388,
|
| 495 |
+
"step": 680
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 1.1724324324324324,
|
| 499 |
+
"grad_norm": 0.6257322430610657,
|
| 500 |
+
"learning_rate": 3.483483483483483e-05,
|
| 501 |
+
"loss": 0.4094,
|
| 502 |
+
"step": 690
|
| 503 |
+
},
|
| 504 |
+
{
|
| 505 |
+
"epoch": 1.1778378378378378,
|
| 506 |
+
"grad_norm": 6.534117221832275,
|
| 507 |
+
"learning_rate": 3.453453453453453e-05,
|
| 508 |
+
"loss": 1.1549,
|
| 509 |
+
"step": 700
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 1.1832432432432431,
|
| 513 |
+
"grad_norm": 13.602584838867188,
|
| 514 |
+
"learning_rate": 3.4234234234234234e-05,
|
| 515 |
+
"loss": 1.077,
|
| 516 |
+
"step": 710
|
| 517 |
+
},
|
| 518 |
+
{
|
| 519 |
+
"epoch": 1.1886486486486487,
|
| 520 |
+
"grad_norm": 18.255807876586914,
|
| 521 |
+
"learning_rate": 3.3933933933933934e-05,
|
| 522 |
+
"loss": 1.1547,
|
| 523 |
+
"step": 720
|
| 524 |
+
},
|
| 525 |
+
{
|
| 526 |
+
"epoch": 1.194054054054054,
|
| 527 |
+
"grad_norm": 0.5059702396392822,
|
| 528 |
+
"learning_rate": 3.3633633633633635e-05,
|
| 529 |
+
"loss": 1.1022,
|
| 530 |
+
"step": 730
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 1.1994594594594594,
|
| 534 |
+
"grad_norm": 0.5154173970222473,
|
| 535 |
+
"learning_rate": 3.3333333333333335e-05,
|
| 536 |
+
"loss": 1.0586,
|
| 537 |
+
"step": 740
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 1.2005405405405405,
|
| 541 |
+
"eval_accuracy": 0.5819672131147541,
|
| 542 |
+
"eval_f1": 0.5034862905182809,
|
| 543 |
+
"eval_loss": 1.2865582704544067,
|
| 544 |
+
"eval_runtime": 398.9746,
|
| 545 |
+
"eval_samples_per_second": 0.306,
|
| 546 |
+
"eval_steps_per_second": 0.153,
|
| 547 |
+
"step": 742
|
| 548 |
+
},
|
| 549 |
+
{
|
| 550 |
+
"epoch": 2.0043243243243243,
|
| 551 |
+
"grad_norm": 1.7890368700027466,
|
| 552 |
+
"learning_rate": 3.3033033033033035e-05,
|
| 553 |
+
"loss": 0.5098,
|
| 554 |
+
"step": 750
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 2.0097297297297296,
|
| 558 |
+
"grad_norm": 0.3416980504989624,
|
| 559 |
+
"learning_rate": 3.2732732732732736e-05,
|
| 560 |
+
"loss": 0.4915,
|
| 561 |
+
"step": 760
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 2.015135135135135,
|
| 565 |
+
"grad_norm": 38.26604461669922,
|
| 566 |
+
"learning_rate": 3.2432432432432436e-05,
|
| 567 |
+
"loss": 1.6502,
|
| 568 |
+
"step": 770
|
| 569 |
+
},
|
| 570 |
+
{
|
| 571 |
+
"epoch": 2.0205405405405403,
|
| 572 |
+
"grad_norm": 0.615081787109375,
|
| 573 |
+
"learning_rate": 3.213213213213213e-05,
|
| 574 |
+
"loss": 0.5299,
|
| 575 |
+
"step": 780
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 2.025945945945946,
|
| 579 |
+
"grad_norm": 0.5157684683799744,
|
| 580 |
+
"learning_rate": 3.183183183183183e-05,
|
| 581 |
+
"loss": 0.9929,
|
| 582 |
+
"step": 790
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 2.0313513513513515,
|
| 586 |
+
"grad_norm": 7.0944318771362305,
|
| 587 |
+
"learning_rate": 3.153153153153153e-05,
|
| 588 |
+
"loss": 0.836,
|
| 589 |
+
"step": 800
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 2.036756756756757,
|
| 593 |
+
"grad_norm": 35.1708984375,
|
| 594 |
+
"learning_rate": 3.123123123123123e-05,
|
| 595 |
+
"loss": 0.6723,
|
| 596 |
+
"step": 810
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 2.042162162162162,
|
| 600 |
+
"grad_norm": 21.07259750366211,
|
| 601 |
+
"learning_rate": 3.093093093093093e-05,
|
| 602 |
+
"loss": 1.307,
|
| 603 |
+
"step": 820
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 2.0475675675675675,
|
| 607 |
+
"grad_norm": 8.620600700378418,
|
| 608 |
+
"learning_rate": 3.063063063063063e-05,
|
| 609 |
+
"loss": 0.7787,
|
| 610 |
+
"step": 830
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"epoch": 2.052972972972973,
|
| 614 |
+
"grad_norm": 5.618999004364014,
|
| 615 |
+
"learning_rate": 3.0330330330330332e-05,
|
| 616 |
+
"loss": 0.8469,
|
| 617 |
+
"step": 840
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 2.0583783783783782,
|
| 621 |
+
"grad_norm": 23.006450653076172,
|
| 622 |
+
"learning_rate": 3.0030030030030033e-05,
|
| 623 |
+
"loss": 1.4332,
|
| 624 |
+
"step": 850
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 2.0637837837837836,
|
| 628 |
+
"grad_norm": 28.410768508911133,
|
| 629 |
+
"learning_rate": 2.9729729729729733e-05,
|
| 630 |
+
"loss": 1.1213,
|
| 631 |
+
"step": 860
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 2.0691891891891894,
|
| 635 |
+
"grad_norm": 35.26121139526367,
|
| 636 |
+
"learning_rate": 2.9429429429429427e-05,
|
| 637 |
+
"loss": 1.1048,
|
| 638 |
+
"step": 870
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 2.0745945945945947,
|
| 642 |
+
"grad_norm": 10.972575187683105,
|
| 643 |
+
"learning_rate": 2.912912912912913e-05,
|
| 644 |
+
"loss": 0.7834,
|
| 645 |
+
"step": 880
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 2.08,
|
| 649 |
+
"grad_norm": 4.512319087982178,
|
| 650 |
+
"learning_rate": 2.882882882882883e-05,
|
| 651 |
+
"loss": 0.3903,
|
| 652 |
+
"step": 890
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 2.0854054054054054,
|
| 656 |
+
"grad_norm": 14.238494873046875,
|
| 657 |
+
"learning_rate": 2.852852852852853e-05,
|
| 658 |
+
"loss": 1.626,
|
| 659 |
+
"step": 900
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 2.0908108108108108,
|
| 663 |
+
"grad_norm": 15.402420997619629,
|
| 664 |
+
"learning_rate": 2.8228228228228232e-05,
|
| 665 |
+
"loss": 1.1367,
|
| 666 |
+
"step": 910
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 2.096216216216216,
|
| 670 |
+
"grad_norm": 1.233629822731018,
|
| 671 |
+
"learning_rate": 2.7927927927927926e-05,
|
| 672 |
+
"loss": 0.844,
|
| 673 |
+
"step": 920
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 2.1016216216216215,
|
| 677 |
+
"grad_norm": 25.64818572998047,
|
| 678 |
+
"learning_rate": 2.762762762762763e-05,
|
| 679 |
+
"loss": 0.8664,
|
| 680 |
+
"step": 930
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 2.107027027027027,
|
| 684 |
+
"grad_norm": 0.6849324703216553,
|
| 685 |
+
"learning_rate": 2.732732732732733e-05,
|
| 686 |
+
"loss": 0.5829,
|
| 687 |
+
"step": 940
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 2.1124324324324326,
|
| 691 |
+
"grad_norm": 2.735356092453003,
|
| 692 |
+
"learning_rate": 2.702702702702703e-05,
|
| 693 |
+
"loss": 0.4426,
|
| 694 |
+
"step": 950
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 2.117837837837838,
|
| 698 |
+
"grad_norm": 40.63231658935547,
|
| 699 |
+
"learning_rate": 2.672672672672673e-05,
|
| 700 |
+
"loss": 0.4518,
|
| 701 |
+
"step": 960
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"epoch": 2.1232432432432433,
|
| 705 |
+
"grad_norm": 5.400685787200928,
|
| 706 |
+
"learning_rate": 2.6426426426426428e-05,
|
| 707 |
+
"loss": 0.4123,
|
| 708 |
+
"step": 970
|
| 709 |
+
},
|
| 710 |
+
{
|
| 711 |
+
"epoch": 2.1286486486486487,
|
| 712 |
+
"grad_norm": 39.79267883300781,
|
| 713 |
+
"learning_rate": 2.6126126126126128e-05,
|
| 714 |
+
"loss": 0.8665,
|
| 715 |
+
"step": 980
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"epoch": 2.134054054054054,
|
| 719 |
+
"grad_norm": 26.877548217773438,
|
| 720 |
+
"learning_rate": 2.582582582582583e-05,
|
| 721 |
+
"loss": 0.9006,
|
| 722 |
+
"step": 990
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 2.1394594594594594,
|
| 726 |
+
"grad_norm": 17.957853317260742,
|
| 727 |
+
"learning_rate": 2.552552552552553e-05,
|
| 728 |
+
"loss": 0.1425,
|
| 729 |
+
"step": 1000
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 2.1448648648648647,
|
| 733 |
+
"grad_norm": 1.0850050449371338,
|
| 734 |
+
"learning_rate": 2.5225225225225222e-05,
|
| 735 |
+
"loss": 0.6361,
|
| 736 |
+
"step": 1010
|
| 737 |
+
},
|
| 738 |
+
{
|
| 739 |
+
"epoch": 2.15027027027027,
|
| 740 |
+
"grad_norm": 65.59326171875,
|
| 741 |
+
"learning_rate": 2.4924924924924926e-05,
|
| 742 |
+
"loss": 0.5456,
|
| 743 |
+
"step": 1020
|
| 744 |
+
},
|
| 745 |
+
{
|
| 746 |
+
"epoch": 2.155675675675676,
|
| 747 |
+
"grad_norm": 75.51756286621094,
|
| 748 |
+
"learning_rate": 2.4624624624624627e-05,
|
| 749 |
+
"loss": 0.5168,
|
| 750 |
+
"step": 1030
|
| 751 |
+
},
|
| 752 |
+
{
|
| 753 |
+
"epoch": 2.161081081081081,
|
| 754 |
+
"grad_norm": 0.284772127866745,
|
| 755 |
+
"learning_rate": 2.4324324324324327e-05,
|
| 756 |
+
"loss": 0.5227,
|
| 757 |
+
"step": 1040
|
| 758 |
+
},
|
| 759 |
+
{
|
| 760 |
+
"epoch": 2.1664864864864866,
|
| 761 |
+
"grad_norm": 57.656734466552734,
|
| 762 |
+
"learning_rate": 2.4024024024024024e-05,
|
| 763 |
+
"loss": 0.978,
|
| 764 |
+
"step": 1050
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 2.171891891891892,
|
| 768 |
+
"grad_norm": 5.7262468338012695,
|
| 769 |
+
"learning_rate": 2.3723723723723725e-05,
|
| 770 |
+
"loss": 0.4327,
|
| 771 |
+
"step": 1060
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 2.1772972972972973,
|
| 775 |
+
"grad_norm": 0.4865647256374359,
|
| 776 |
+
"learning_rate": 2.3423423423423425e-05,
|
| 777 |
+
"loss": 0.5062,
|
| 778 |
+
"step": 1070
|
| 779 |
+
},
|
| 780 |
+
{
|
| 781 |
+
"epoch": 2.1827027027027026,
|
| 782 |
+
"grad_norm": 30.51907730102539,
|
| 783 |
+
"learning_rate": 2.3123123123123125e-05,
|
| 784 |
+
"loss": 0.3505,
|
| 785 |
+
"step": 1080
|
| 786 |
+
},
|
| 787 |
+
{
|
| 788 |
+
"epoch": 2.188108108108108,
|
| 789 |
+
"grad_norm": 0.11615642160177231,
|
| 790 |
+
"learning_rate": 2.2822822822822822e-05,
|
| 791 |
+
"loss": 0.7081,
|
| 792 |
+
"step": 1090
|
| 793 |
+
},
|
| 794 |
+
{
|
| 795 |
+
"epoch": 2.1935135135135133,
|
| 796 |
+
"grad_norm": 30.363637924194336,
|
| 797 |
+
"learning_rate": 2.2522522522522523e-05,
|
| 798 |
+
"loss": 0.0567,
|
| 799 |
+
"step": 1100
|
| 800 |
+
},
|
| 801 |
+
{
|
| 802 |
+
"epoch": 2.198918918918919,
|
| 803 |
+
"grad_norm": 0.29021716117858887,
|
| 804 |
+
"learning_rate": 2.2222222222222223e-05,
|
| 805 |
+
"loss": 0.9781,
|
| 806 |
+
"step": 1110
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 2.2005405405405405,
|
| 810 |
+
"eval_accuracy": 0.7459016393442623,
|
| 811 |
+
"eval_f1": 0.7466341940069591,
|
| 812 |
+
"eval_loss": 0.7952272891998291,
|
| 813 |
+
"eval_runtime": 615.8439,
|
| 814 |
+
"eval_samples_per_second": 0.198,
|
| 815 |
+
"eval_steps_per_second": 0.099,
|
| 816 |
+
"step": 1113
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 3.003783783783784,
|
| 820 |
+
"grad_norm": 0.5570007562637329,
|
| 821 |
+
"learning_rate": 2.1921921921921924e-05,
|
| 822 |
+
"loss": 0.8402,
|
| 823 |
+
"step": 1120
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 3.0091891891891893,
|
| 827 |
+
"grad_norm": 0.30576571822166443,
|
| 828 |
+
"learning_rate": 2.1621621621621624e-05,
|
| 829 |
+
"loss": 0.5837,
|
| 830 |
+
"step": 1130
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 3.0145945945945947,
|
| 834 |
+
"grad_norm": 0.4280945956707001,
|
| 835 |
+
"learning_rate": 2.132132132132132e-05,
|
| 836 |
+
"loss": 0.609,
|
| 837 |
+
"step": 1140
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 3.02,
|
| 841 |
+
"grad_norm": 0.037259869277477264,
|
| 842 |
+
"learning_rate": 2.102102102102102e-05,
|
| 843 |
+
"loss": 0.1856,
|
| 844 |
+
"step": 1150
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 3.0254054054054054,
|
| 848 |
+
"grad_norm": 0.031064892187714577,
|
| 849 |
+
"learning_rate": 2.0720720720720722e-05,
|
| 850 |
+
"loss": 0.1205,
|
| 851 |
+
"step": 1160
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 3.0308108108108107,
|
| 855 |
+
"grad_norm": 0.029491625726222992,
|
| 856 |
+
"learning_rate": 2.0420420420420422e-05,
|
| 857 |
+
"loss": 0.4757,
|
| 858 |
+
"step": 1170
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 3.036216216216216,
|
| 862 |
+
"grad_norm": 0.06313157826662064,
|
| 863 |
+
"learning_rate": 2.012012012012012e-05,
|
| 864 |
+
"loss": 0.54,
|
| 865 |
+
"step": 1180
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 3.0416216216216214,
|
| 869 |
+
"grad_norm": 44.53404235839844,
|
| 870 |
+
"learning_rate": 1.981981981981982e-05,
|
| 871 |
+
"loss": 0.7178,
|
| 872 |
+
"step": 1190
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 3.047027027027027,
|
| 876 |
+
"grad_norm": 28.65042495727539,
|
| 877 |
+
"learning_rate": 1.951951951951952e-05,
|
| 878 |
+
"loss": 0.2756,
|
| 879 |
+
"step": 1200
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 3.0524324324324326,
|
| 883 |
+
"grad_norm": 3.5685808658599854,
|
| 884 |
+
"learning_rate": 1.921921921921922e-05,
|
| 885 |
+
"loss": 0.2041,
|
| 886 |
+
"step": 1210
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 3.057837837837838,
|
| 890 |
+
"grad_norm": 100.94510650634766,
|
| 891 |
+
"learning_rate": 1.891891891891892e-05,
|
| 892 |
+
"loss": 0.1725,
|
| 893 |
+
"step": 1220
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 3.0632432432432433,
|
| 897 |
+
"grad_norm": 69.8983383178711,
|
| 898 |
+
"learning_rate": 1.8618618618618618e-05,
|
| 899 |
+
"loss": 0.8736,
|
| 900 |
+
"step": 1230
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 3.0686486486486486,
|
| 904 |
+
"grad_norm": 15.938907623291016,
|
| 905 |
+
"learning_rate": 1.831831831831832e-05,
|
| 906 |
+
"loss": 0.3815,
|
| 907 |
+
"step": 1240
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 3.074054054054054,
|
| 911 |
+
"grad_norm": 14.810787200927734,
|
| 912 |
+
"learning_rate": 1.801801801801802e-05,
|
| 913 |
+
"loss": 0.3635,
|
| 914 |
+
"step": 1250
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 3.0794594594594593,
|
| 918 |
+
"grad_norm": 18.970157623291016,
|
| 919 |
+
"learning_rate": 1.771771771771772e-05,
|
| 920 |
+
"loss": 0.3695,
|
| 921 |
+
"step": 1260
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 3.0848648648648647,
|
| 925 |
+
"grad_norm": 0.7603211998939514,
|
| 926 |
+
"learning_rate": 1.7417417417417416e-05,
|
| 927 |
+
"loss": 0.4226,
|
| 928 |
+
"step": 1270
|
| 929 |
+
},
|
| 930 |
+
{
|
| 931 |
+
"epoch": 3.0902702702702705,
|
| 932 |
+
"grad_norm": 0.04897380620241165,
|
| 933 |
+
"learning_rate": 1.7117117117117117e-05,
|
| 934 |
+
"loss": 0.0055,
|
| 935 |
+
"step": 1280
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 3.095675675675676,
|
| 939 |
+
"grad_norm": 0.04457269236445427,
|
| 940 |
+
"learning_rate": 1.6816816816816817e-05,
|
| 941 |
+
"loss": 0.2734,
|
| 942 |
+
"step": 1290
|
| 943 |
+
},
|
| 944 |
+
{
|
| 945 |
+
"epoch": 3.101081081081081,
|
| 946 |
+
"grad_norm": 11.417162895202637,
|
| 947 |
+
"learning_rate": 1.6516516516516518e-05,
|
| 948 |
+
"loss": 0.0319,
|
| 949 |
+
"step": 1300
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 3.1064864864864865,
|
| 953 |
+
"grad_norm": 73.96791076660156,
|
| 954 |
+
"learning_rate": 1.6216216216216218e-05,
|
| 955 |
+
"loss": 0.3966,
|
| 956 |
+
"step": 1310
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 3.111891891891892,
|
| 960 |
+
"grad_norm": 0.051069825887680054,
|
| 961 |
+
"learning_rate": 1.5915915915915915e-05,
|
| 962 |
+
"loss": 0.5875,
|
| 963 |
+
"step": 1320
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 3.117297297297297,
|
| 967 |
+
"grad_norm": 0.6849485635757446,
|
| 968 |
+
"learning_rate": 1.5615615615615616e-05,
|
| 969 |
+
"loss": 0.233,
|
| 970 |
+
"step": 1330
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 3.1227027027027026,
|
| 974 |
+
"grad_norm": 0.09105366468429565,
|
| 975 |
+
"learning_rate": 1.5315315315315316e-05,
|
| 976 |
+
"loss": 0.4759,
|
| 977 |
+
"step": 1340
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 3.128108108108108,
|
| 981 |
+
"grad_norm": 0.02636696584522724,
|
| 982 |
+
"learning_rate": 1.5015015015015016e-05,
|
| 983 |
+
"loss": 0.1464,
|
| 984 |
+
"step": 1350
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 3.1335135135135137,
|
| 988 |
+
"grad_norm": 9.379318237304688,
|
| 989 |
+
"learning_rate": 1.4714714714714713e-05,
|
| 990 |
+
"loss": 0.2798,
|
| 991 |
+
"step": 1360
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 3.138918918918919,
|
| 995 |
+
"grad_norm": 0.11463417857885361,
|
| 996 |
+
"learning_rate": 1.4414414414414416e-05,
|
| 997 |
+
"loss": 0.3684,
|
| 998 |
+
"step": 1370
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 3.1443243243243244,
|
| 1002 |
+
"grad_norm": 28.1702823638916,
|
| 1003 |
+
"learning_rate": 1.4114114114114116e-05,
|
| 1004 |
+
"loss": 0.3324,
|
| 1005 |
+
"step": 1380
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 3.1497297297297298,
|
| 1009 |
+
"grad_norm": 0.024587994441390038,
|
| 1010 |
+
"learning_rate": 1.3813813813813815e-05,
|
| 1011 |
+
"loss": 0.3369,
|
| 1012 |
+
"step": 1390
|
| 1013 |
+
},
|
| 1014 |
+
{
|
| 1015 |
+
"epoch": 3.155135135135135,
|
| 1016 |
+
"grad_norm": 0.03437427058815956,
|
| 1017 |
+
"learning_rate": 1.3513513513513515e-05,
|
| 1018 |
+
"loss": 0.2755,
|
| 1019 |
+
"step": 1400
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 3.1605405405405405,
|
| 1023 |
+
"grad_norm": 0.034997034817934036,
|
| 1024 |
+
"learning_rate": 1.3213213213213214e-05,
|
| 1025 |
+
"loss": 0.0174,
|
| 1026 |
+
"step": 1410
|
| 1027 |
+
},
|
| 1028 |
+
{
|
| 1029 |
+
"epoch": 3.165945945945946,
|
| 1030 |
+
"grad_norm": 50.559940338134766,
|
| 1031 |
+
"learning_rate": 1.2912912912912914e-05,
|
| 1032 |
+
"loss": 0.3477,
|
| 1033 |
+
"step": 1420
|
| 1034 |
+
},
|
| 1035 |
+
{
|
| 1036 |
+
"epoch": 3.171351351351351,
|
| 1037 |
+
"grad_norm": 1.0858763456344604,
|
| 1038 |
+
"learning_rate": 1.2612612612612611e-05,
|
| 1039 |
+
"loss": 0.1806,
|
| 1040 |
+
"step": 1430
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 3.176756756756757,
|
| 1044 |
+
"grad_norm": 95.9212417602539,
|
| 1045 |
+
"learning_rate": 1.2312312312312313e-05,
|
| 1046 |
+
"loss": 0.3487,
|
| 1047 |
+
"step": 1440
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 3.1821621621621623,
|
| 1051 |
+
"grad_norm": 35.4121208190918,
|
| 1052 |
+
"learning_rate": 1.2012012012012012e-05,
|
| 1053 |
+
"loss": 0.3323,
|
| 1054 |
+
"step": 1450
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"epoch": 3.1875675675675677,
|
| 1058 |
+
"grad_norm": 0.050625383853912354,
|
| 1059 |
+
"learning_rate": 1.1711711711711713e-05,
|
| 1060 |
+
"loss": 0.227,
|
| 1061 |
+
"step": 1460
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"epoch": 3.192972972972973,
|
| 1065 |
+
"grad_norm": 1.5548391342163086,
|
| 1066 |
+
"learning_rate": 1.1411411411411411e-05,
|
| 1067 |
+
"loss": 0.797,
|
| 1068 |
+
"step": 1470
|
| 1069 |
+
},
|
| 1070 |
+
{
|
| 1071 |
+
"epoch": 3.1983783783783784,
|
| 1072 |
+
"grad_norm": 0.6227735877037048,
|
| 1073 |
+
"learning_rate": 1.1111111111111112e-05,
|
| 1074 |
+
"loss": 0.0034,
|
| 1075 |
+
"step": 1480
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"epoch": 3.2005405405405405,
|
| 1079 |
+
"eval_accuracy": 0.8360655737704918,
|
| 1080 |
+
"eval_f1": 0.8343035855751879,
|
| 1081 |
+
"eval_loss": 0.7218076586723328,
|
| 1082 |
+
"eval_runtime": 507.9346,
|
| 1083 |
+
"eval_samples_per_second": 0.24,
|
| 1084 |
+
"eval_steps_per_second": 0.12,
|
| 1085 |
+
"step": 1484
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 4.003243243243244,
|
| 1089 |
+
"grad_norm": 0.12078996747732162,
|
| 1090 |
+
"learning_rate": 1.0810810810810812e-05,
|
| 1091 |
+
"loss": 0.8434,
|
| 1092 |
+
"step": 1490
|
| 1093 |
+
},
|
| 1094 |
+
{
|
| 1095 |
+
"epoch": 4.008648648648649,
|
| 1096 |
+
"grad_norm": 0.06610610336065292,
|
| 1097 |
+
"learning_rate": 1.051051051051051e-05,
|
| 1098 |
+
"loss": 0.4016,
|
| 1099 |
+
"step": 1500
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 4.014054054054054,
|
| 1103 |
+
"grad_norm": 0.03932103142142296,
|
| 1104 |
+
"learning_rate": 1.0210210210210211e-05,
|
| 1105 |
+
"loss": 0.1955,
|
| 1106 |
+
"step": 1510
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 4.019459459459459,
|
| 1110 |
+
"grad_norm": 0.05058516561985016,
|
| 1111 |
+
"learning_rate": 9.90990990990991e-06,
|
| 1112 |
+
"loss": 0.063,
|
| 1113 |
+
"step": 1520
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 4.024864864864865,
|
| 1117 |
+
"grad_norm": 15.977706909179688,
|
| 1118 |
+
"learning_rate": 9.60960960960961e-06,
|
| 1119 |
+
"loss": 0.1466,
|
| 1120 |
+
"step": 1530
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"epoch": 4.03027027027027,
|
| 1124 |
+
"grad_norm": 12.75985050201416,
|
| 1125 |
+
"learning_rate": 9.309309309309309e-06,
|
| 1126 |
+
"loss": 0.0269,
|
| 1127 |
+
"step": 1540
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 4.035675675675676,
|
| 1131 |
+
"grad_norm": 65.94677734375,
|
| 1132 |
+
"learning_rate": 9.00900900900901e-06,
|
| 1133 |
+
"loss": 0.3077,
|
| 1134 |
+
"step": 1550
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 4.041081081081081,
|
| 1138 |
+
"grad_norm": 0.05467180907726288,
|
| 1139 |
+
"learning_rate": 8.708708708708708e-06,
|
| 1140 |
+
"loss": 0.4931,
|
| 1141 |
+
"step": 1560
|
| 1142 |
+
},
|
| 1143 |
+
{
|
| 1144 |
+
"epoch": 4.0464864864864865,
|
| 1145 |
+
"grad_norm": 0.03017774224281311,
|
| 1146 |
+
"learning_rate": 8.408408408408409e-06,
|
| 1147 |
+
"loss": 0.0026,
|
| 1148 |
+
"step": 1570
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 4.051891891891892,
|
| 1152 |
+
"grad_norm": 0.03868546336889267,
|
| 1153 |
+
"learning_rate": 8.108108108108109e-06,
|
| 1154 |
+
"loss": 0.6954,
|
| 1155 |
+
"step": 1580
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 4.057297297297297,
|
| 1159 |
+
"grad_norm": 0.025062745437026024,
|
| 1160 |
+
"learning_rate": 7.807807807807808e-06,
|
| 1161 |
+
"loss": 0.2377,
|
| 1162 |
+
"step": 1590
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 4.062702702702703,
|
| 1166 |
+
"grad_norm": 6.950279712677002,
|
| 1167 |
+
"learning_rate": 7.507507507507508e-06,
|
| 1168 |
+
"loss": 0.006,
|
| 1169 |
+
"step": 1600
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 4.068108108108108,
|
| 1173 |
+
"grad_norm": 0.03524166718125343,
|
| 1174 |
+
"learning_rate": 7.207207207207208e-06,
|
| 1175 |
+
"loss": 0.0245,
|
| 1176 |
+
"step": 1610
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 4.073513513513514,
|
| 1180 |
+
"grad_norm": 0.02506636269390583,
|
| 1181 |
+
"learning_rate": 6.906906906906907e-06,
|
| 1182 |
+
"loss": 0.003,
|
| 1183 |
+
"step": 1620
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 4.078918918918919,
|
| 1187 |
+
"grad_norm": 0.0216878280043602,
|
| 1188 |
+
"learning_rate": 6.606606606606607e-06,
|
| 1189 |
+
"loss": 0.0095,
|
| 1190 |
+
"step": 1630
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 4.084324324324324,
|
| 1194 |
+
"grad_norm": 0.01947568915784359,
|
| 1195 |
+
"learning_rate": 6.306306306306306e-06,
|
| 1196 |
+
"loss": 0.0029,
|
| 1197 |
+
"step": 1640
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 4.08972972972973,
|
| 1201 |
+
"grad_norm": 0.17050786316394806,
|
| 1202 |
+
"learning_rate": 6.006006006006006e-06,
|
| 1203 |
+
"loss": 0.32,
|
| 1204 |
+
"step": 1650
|
| 1205 |
+
},
|
| 1206 |
+
{
|
| 1207 |
+
"epoch": 4.095135135135135,
|
| 1208 |
+
"grad_norm": 0.025170955806970596,
|
| 1209 |
+
"learning_rate": 5.705705705705706e-06,
|
| 1210 |
+
"loss": 0.1245,
|
| 1211 |
+
"step": 1660
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 4.100540540540541,
|
| 1215 |
+
"grad_norm": 0.024301931262016296,
|
| 1216 |
+
"learning_rate": 5.405405405405406e-06,
|
| 1217 |
+
"loss": 0.1942,
|
| 1218 |
+
"step": 1670
|
| 1219 |
+
},
|
| 1220 |
+
{
|
| 1221 |
+
"epoch": 4.105945945945946,
|
| 1222 |
+
"grad_norm": 0.03195532411336899,
|
| 1223 |
+
"learning_rate": 5.105105105105106e-06,
|
| 1224 |
+
"loss": 0.0019,
|
| 1225 |
+
"step": 1680
|
| 1226 |
+
},
|
| 1227 |
+
{
|
| 1228 |
+
"epoch": 4.1113513513513515,
|
| 1229 |
+
"grad_norm": 0.09079870581626892,
|
| 1230 |
+
"learning_rate": 4.804804804804805e-06,
|
| 1231 |
+
"loss": 0.0139,
|
| 1232 |
+
"step": 1690
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 4.1167567567567565,
|
| 1236 |
+
"grad_norm": 0.02011438086628914,
|
| 1237 |
+
"learning_rate": 4.504504504504505e-06,
|
| 1238 |
+
"loss": 0.0021,
|
| 1239 |
+
"step": 1700
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 4.122162162162162,
|
| 1243 |
+
"grad_norm": 0.016972195357084274,
|
| 1244 |
+
"learning_rate": 4.204204204204204e-06,
|
| 1245 |
+
"loss": 0.0021,
|
| 1246 |
+
"step": 1710
|
| 1247 |
+
},
|
| 1248 |
+
{
|
| 1249 |
+
"epoch": 4.127567567567567,
|
| 1250 |
+
"grad_norm": 17.876731872558594,
|
| 1251 |
+
"learning_rate": 3.903903903903904e-06,
|
| 1252 |
+
"loss": 0.4547,
|
| 1253 |
+
"step": 1720
|
| 1254 |
+
},
|
| 1255 |
+
{
|
| 1256 |
+
"epoch": 4.132972972972973,
|
| 1257 |
+
"grad_norm": 0.1327604353427887,
|
| 1258 |
+
"learning_rate": 3.603603603603604e-06,
|
| 1259 |
+
"loss": 0.0154,
|
| 1260 |
+
"step": 1730
|
| 1261 |
+
},
|
| 1262 |
+
{
|
| 1263 |
+
"epoch": 4.138378378378379,
|
| 1264 |
+
"grad_norm": 0.0263642817735672,
|
| 1265 |
+
"learning_rate": 3.3033033033033035e-06,
|
| 1266 |
+
"loss": 0.3324,
|
| 1267 |
+
"step": 1740
|
| 1268 |
+
},
|
| 1269 |
+
{
|
| 1270 |
+
"epoch": 4.143783783783784,
|
| 1271 |
+
"grad_norm": 56.30789566040039,
|
| 1272 |
+
"learning_rate": 3.003003003003003e-06,
|
| 1273 |
+
"loss": 0.3968,
|
| 1274 |
+
"step": 1750
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 4.149189189189189,
|
| 1278 |
+
"grad_norm": 0.2022549957036972,
|
| 1279 |
+
"learning_rate": 2.702702702702703e-06,
|
| 1280 |
+
"loss": 0.3955,
|
| 1281 |
+
"step": 1760
|
| 1282 |
+
},
|
| 1283 |
+
{
|
| 1284 |
+
"epoch": 4.154594594594594,
|
| 1285 |
+
"grad_norm": 0.02462898939847946,
|
| 1286 |
+
"learning_rate": 2.4024024024024026e-06,
|
| 1287 |
+
"loss": 0.0026,
|
| 1288 |
+
"step": 1770
|
| 1289 |
+
},
|
| 1290 |
+
{
|
| 1291 |
+
"epoch": 4.16,
|
| 1292 |
+
"grad_norm": 0.030352883040905,
|
| 1293 |
+
"learning_rate": 2.102102102102102e-06,
|
| 1294 |
+
"loss": 0.0999,
|
| 1295 |
+
"step": 1780
|
| 1296 |
+
},
|
| 1297 |
+
{
|
| 1298 |
+
"epoch": 4.165405405405405,
|
| 1299 |
+
"grad_norm": 85.04549407958984,
|
| 1300 |
+
"learning_rate": 1.801801801801802e-06,
|
| 1301 |
+
"loss": 0.1927,
|
| 1302 |
+
"step": 1790
|
| 1303 |
+
},
|
| 1304 |
+
{
|
| 1305 |
+
"epoch": 4.170810810810811,
|
| 1306 |
+
"grad_norm": 1.2970046997070312,
|
| 1307 |
+
"learning_rate": 1.5015015015015015e-06,
|
| 1308 |
+
"loss": 0.0034,
|
| 1309 |
+
"step": 1800
|
| 1310 |
+
},
|
| 1311 |
+
{
|
| 1312 |
+
"epoch": 4.176216216216217,
|
| 1313 |
+
"grad_norm": 0.02083350159227848,
|
| 1314 |
+
"learning_rate": 1.2012012012012013e-06,
|
| 1315 |
+
"loss": 0.014,
|
| 1316 |
+
"step": 1810
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 4.1816216216216215,
|
| 1320 |
+
"grad_norm": 0.4275698959827423,
|
| 1321 |
+
"learning_rate": 9.00900900900901e-07,
|
| 1322 |
+
"loss": 0.2991,
|
| 1323 |
+
"step": 1820
|
| 1324 |
+
},
|
| 1325 |
+
{
|
| 1326 |
+
"epoch": 4.187027027027027,
|
| 1327 |
+
"grad_norm": 0.022126413881778717,
|
| 1328 |
+
"learning_rate": 6.006006006006006e-07,
|
| 1329 |
+
"loss": 0.2785,
|
| 1330 |
+
"step": 1830
|
| 1331 |
+
},
|
| 1332 |
+
{
|
| 1333 |
+
"epoch": 4.192432432432432,
|
| 1334 |
+
"grad_norm": 0.020599039271473885,
|
| 1335 |
+
"learning_rate": 3.003003003003003e-07,
|
| 1336 |
+
"loss": 0.0024,
|
| 1337 |
+
"step": 1840
|
| 1338 |
+
},
|
| 1339 |
+
{
|
| 1340 |
+
"epoch": 4.197837837837838,
|
| 1341 |
+
"grad_norm": 0.04799671843647957,
|
| 1342 |
+
"learning_rate": 0.0,
|
| 1343 |
+
"loss": 0.1895,
|
| 1344 |
+
"step": 1850
|
| 1345 |
+
},
|
| 1346 |
+
{
|
| 1347 |
+
"epoch": 4.197837837837838,
|
| 1348 |
+
"eval_accuracy": 0.8852459016393442,
|
| 1349 |
+
"eval_f1": 0.8839858413628905,
|
| 1350 |
+
"eval_loss": 0.5544242262840271,
|
| 1351 |
+
"eval_runtime": 710.05,
|
| 1352 |
+
"eval_samples_per_second": 0.172,
|
| 1353 |
+
"eval_steps_per_second": 0.086,
|
| 1354 |
+
"step": 1850
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 4.197837837837838,
|
| 1358 |
+
"step": 1850,
|
| 1359 |
+
"total_flos": 4.605536866872066e+18,
|
| 1360 |
+
"train_loss": 0.7479323436269486,
|
| 1361 |
+
"train_runtime": 45711.7743,
|
| 1362 |
+
"train_samples_per_second": 0.081,
|
| 1363 |
+
"train_steps_per_second": 0.04
|
| 1364 |
+
},
|
| 1365 |
+
{
|
| 1366 |
+
"epoch": 4.197837837837838,
|
| 1367 |
+
"eval_accuracy": 0.9784075573549258,
|
| 1368 |
+
"eval_f1": 0.9783336380382718,
|
| 1369 |
+
"eval_loss": 0.09701067209243774,
|
| 1370 |
+
"eval_runtime": 4652.924,
|
| 1371 |
+
"eval_samples_per_second": 0.159,
|
| 1372 |
+
"eval_steps_per_second": 0.08,
|
| 1373 |
+
"step": 1850
|
| 1374 |
+
},
|
| 1375 |
+
{
|
| 1376 |
+
"epoch": 4.197837837837838,
|
| 1377 |
+
"eval_accuracy": 0.851528384279476,
|
| 1378 |
+
"eval_f1": 0.8496851913824908,
|
| 1379 |
+
"eval_loss": 0.7380829453468323,
|
| 1380 |
+
"eval_runtime": 1382.3283,
|
| 1381 |
+
"eval_samples_per_second": 0.166,
|
| 1382 |
+
"eval_steps_per_second": 0.083,
|
| 1383 |
+
"step": 1850
|
| 1384 |
+
},
|
| 1385 |
+
{
|
| 1386 |
+
"epoch": 4.197837837837838,
|
| 1387 |
+
"eval_accuracy": 0.8852459016393442,
|
| 1388 |
+
"eval_f1": 0.8839858413628905,
|
| 1389 |
+
"eval_loss": 0.5544241070747375,
|
| 1390 |
+
"eval_runtime": 784.0148,
|
| 1391 |
+
"eval_samples_per_second": 0.156,
|
| 1392 |
+
"eval_steps_per_second": 0.078,
|
| 1393 |
+
"step": 1850
|
| 1394 |
}
|
| 1395 |
],
|
| 1396 |
"logging_steps": 10,
|
| 1397 |
+
"max_steps": 1850,
|
| 1398 |
"num_input_tokens_seen": 0,
|
| 1399 |
"num_train_epochs": 9223372036854775807,
|
| 1400 |
"save_steps": 500,
|
|
|
|
| 1410 |
"attributes": {}
|
| 1411 |
}
|
| 1412 |
},
|
| 1413 |
+
"total_flos": 4.605536866872066e+18,
|
| 1414 |
+
"train_batch_size": 2,
|
| 1415 |
"trial_name": null,
|
| 1416 |
"trial_params": null
|
| 1417 |
}
|
val_results.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_f1": 0.
|
| 5 |
-
"eval_loss":
|
| 6 |
-
"eval_runtime":
|
| 7 |
-
"eval_samples_per_second":
|
| 8 |
-
"eval_steps_per_second":
|
| 9 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 4.197837837837838,
|
| 3 |
+
"eval_accuracy": 0.8852459016393442,
|
| 4 |
+
"eval_f1": 0.8839858413628905,
|
| 5 |
+
"eval_loss": 0.5544241070747375,
|
| 6 |
+
"eval_runtime": 784.0148,
|
| 7 |
+
"eval_samples_per_second": 0.156,
|
| 8 |
+
"eval_steps_per_second": 0.078
|
| 9 |
}
|