10-convnext-base-384-finetuned-spiderTraining20-500
This model is a fine-tuned version of facebook/convnext-base-384 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.1900
- Accuracy: 0.9510
- Precision: 0.9493
- Recall: 0.9488
- F1: 0.9482
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0005
- train_batch_size: 25
- eval_batch_size: 25
- seed: 42
- distributed_type: multi-GPU
- gradient_accumulation_steps: 4
- total_train_batch_size: 100
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 |
|---|---|---|---|---|---|---|---|
| 0.8521 | 1.0 | 80 | 0.6379 | 0.7838 | 0.8075 | 0.7774 | 0.7542 |
| 0.5214 | 2.0 | 160 | 0.3445 | 0.8909 | 0.8935 | 0.8833 | 0.8847 |
| 0.4013 | 3.0 | 240 | 0.2821 | 0.9119 | 0.9205 | 0.9048 | 0.9091 |
| 0.3152 | 4.0 | 320 | 0.2633 | 0.9249 | 0.9264 | 0.9234 | 0.9225 |
| 0.2552 | 5.0 | 400 | 0.2837 | 0.9229 | 0.9246 | 0.9179 | 0.9194 |
| 0.236 | 6.0 | 480 | 0.2367 | 0.9329 | 0.9311 | 0.9309 | 0.9301 |
| 0.2178 | 7.0 | 560 | 0.2161 | 0.9389 | 0.9384 | 0.9354 | 0.9360 |
| 0.1712 | 8.0 | 640 | 0.1985 | 0.9459 | 0.9461 | 0.9434 | 0.9439 |
| 0.1607 | 9.0 | 720 | 0.2024 | 0.9489 | 0.9463 | 0.9473 | 0.9454 |
| 0.1592 | 10.0 | 800 | 0.1900 | 0.9510 | 0.9493 | 0.9488 | 0.9482 |
Framework versions
- Transformers 4.33.3
- Pytorch 2.0.1+cu117
- Datasets 2.14.5
- Tokenizers 0.13.3