e5_General_2026_V2
This model is a fine-tuned version of intfloat/multilingual-e5-base on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.2693
- F1 Weighted: 0.8980
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: 2e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- 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 | F1 Weighted |
|---|---|---|---|---|
| 0.9259 | 1.0 | 1047 | 0.2613 | 0.8027 |
| 0.4565 | 2.0 | 2094 | 0.2130 | 0.8427 |
| 0.3400 | 3.0 | 3141 | 0.2124 | 0.8579 |
| 0.2721 | 4.0 | 4188 | 0.2109 | 0.8694 |
| 0.2249 | 5.0 | 5235 | 0.2290 | 0.8845 |
| 0.1894 | 6.0 | 6282 | 0.2344 | 0.8825 |
| 0.1568 | 7.0 | 7329 | 0.2507 | 0.8949 |
| 0.1381 | 8.0 | 8376 | 0.2556 | 0.8937 |
| 0.1176 | 9.0 | 9423 | 0.2743 | 0.8974 |
| 0.1108 | 10.0 | 10470 | 0.2693 | 0.8980 |
Framework versions
- Transformers 5.13.1
- Pytorch 2.11.0+cu128
- Datasets 4.0.0
- Tokenizers 0.22.2