cdp-fp-paf-weighted
This model is a fine-tuned version of alex-miller/ODABert on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.5030
- Accuracy: 0.7719
- F1: 0.8143
- Precision: 0.7917
- Recall: 0.8382
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: 1e-06
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 15
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Precision | Recall |
---|---|---|---|---|---|---|---|
0.6795 | 1.0 | 64 | 0.6463 | 0.7456 | 0.8079 | 0.7349 | 0.8971 |
0.6551 | 2.0 | 128 | 0.6245 | 0.7456 | 0.8105 | 0.7294 | 0.9118 |
0.6387 | 3.0 | 192 | 0.6081 | 0.7456 | 0.8079 | 0.7349 | 0.8971 |
0.6218 | 4.0 | 256 | 0.5948 | 0.7456 | 0.8079 | 0.7349 | 0.8971 |
0.6061 | 5.0 | 320 | 0.5802 | 0.7368 | 0.8 | 0.7317 | 0.8824 |
0.5962 | 6.0 | 384 | 0.5687 | 0.7456 | 0.8054 | 0.7407 | 0.8824 |
0.5828 | 7.0 | 448 | 0.5555 | 0.7544 | 0.8108 | 0.75 | 0.8824 |
0.5633 | 8.0 | 512 | 0.5410 | 0.7719 | 0.8194 | 0.7763 | 0.8676 |
0.5502 | 9.0 | 576 | 0.5317 | 0.7719 | 0.8219 | 0.7692 | 0.8824 |
0.5363 | 10.0 | 640 | 0.5235 | 0.7719 | 0.8219 | 0.7692 | 0.8824 |
0.5311 | 11.0 | 704 | 0.5152 | 0.7719 | 0.8169 | 0.7838 | 0.8529 |
0.5137 | 12.0 | 768 | 0.5092 | 0.7807 | 0.8227 | 0.7945 | 0.8529 |
0.5058 | 13.0 | 832 | 0.5060 | 0.7807 | 0.8227 | 0.7945 | 0.8529 |
0.5003 | 14.0 | 896 | 0.5037 | 0.7719 | 0.8143 | 0.7917 | 0.8382 |
0.4943 | 15.0 | 960 | 0.5030 | 0.7719 | 0.8143 | 0.7917 | 0.8382 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.3.0+cu121
- Datasets 2.19.2
- Tokenizers 0.19.1
- Downloads last month
- 1
This model does not have enough activity to be deployed to Inference API (serverless) yet.
Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.