USS-reward-model-grl_source-WRS
This model is a fine-tuned version of answerdotai/ModernBERT-large on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 17.9649
- Mse: 0.2146
- Mae: 0.3752
- R2: -0.1125
- Spearman Correlation: 0.2297
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: 5e-05
- train_batch_size: 2
- eval_batch_size: 1
- seed: 42
- distributed_type: multi-GPU
- gradient_accumulation_steps: 10
- total_train_batch_size: 20
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Mse | Mae | R2 | Spearman Correlation |
|---|---|---|---|---|---|---|---|
| 152.7645 | 1.0 | 97 | 1.4359 | 0.8449 | 0.7494 | -3.3804 | 0.2728 |
| 22.0290 | 2.0 | 194 | 1.4892 | 0.2564 | 0.3991 | -0.3291 | 0.2993 |
| 48.0805 | 3.0 | 291 | 17.9649 | 0.2146 | 0.3752 | -0.1125 | 0.2297 |
| 364.7763 | 4.0 | 388 | 36.3906 | 0.4170 | 0.5303 | -1.1617 | 0.3048 |
| 405.2551 | 5.0 | 485 | 38.5040 | 1.3588 | 1.0247 | -6.0447 | 0.1875 |
| 424.8833 | 6.0 | 582 | 40.7644 | 1.1246 | 0.8422 | -4.8304 | 0.0984 |
| 418.9320 | 7.0 | 679 | 38.2721 | 0.5541 | 0.6169 | -1.8725 | 0.2504 |
| 394.4251 | 8.0 | 776 | 36.5107 | 0.3334 | 0.4657 | -0.7284 | 0.2747 |
| 496.0636 | 9.0 | 873 | 55.5876 | 0.4464 | 0.5373 | -1.3142 | 0.2367 |
| 568.5653 | 10.0 | 970 | 56.9526 | 0.2958 | 0.4334 | -0.5335 | 0.2825 |
Framework versions
- Transformers 5.9.0
- Pytorch 2.12.0+cu130
- Datasets 4.8.5
- Tokenizers 0.22.2
- Downloads last month
- -
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Model tree for athirorg/USS-reward-model-grl_source-WRS
Base model
answerdotai/ModernBERT-large