Mistral-Nemo-Instruct-Baseline-SFT
This model is a fine-tuned version of mistralai/Mistral-Nemo-Base-2407 on the pbevan11/ultrafeedback_binarized_multilingual dataset. It achieves the following results on the evaluation set:
- Loss: 1.1750
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: 8
- eval_batch_size: 8
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- gradient_accumulation_steps: 4
- total_train_batch_size: 256
- total_eval_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.6158 | 0.9455 | 13 | 1.3800 |
| 1.1061 | 1.9636 | 27 | 1.1854 |
| 0.9071 | 2.8364 | 39 | 1.1750 |
Framework versions
- Transformers 4.45.1
- Pytorch 2.4.1+cu121
- Datasets 3.0.1
- Tokenizers 0.20.0