metadata
license: apache-2.0
library_name: peft
tags:
- trl
- sft
- generated_from_trainer
base_model: mistralai/Mistral-7B-Instruct-v0.3
datasets:
- generator
model-index:
- name: mad_sci_mistral_instruct
results: []
mad_sci_mistral_instruct
This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.3 on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 1.2648
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: 0.0002
- train_batch_size: 4
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant
- lr_scheduler_warmup_steps: 0.03
- training_steps: 100
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.5582 | 0.8333 | 20 | 1.4274 |
| 1.3494 | 1.6667 | 40 | 1.2948 |
| 1.2893 | 2.5 | 60 | 1.2649 |
| 1.2327 | 3.3333 | 80 | 1.2638 |
| 1.1812 | 4.1667 | 100 | 1.2648 |
Framework versions
- PEFT 0.11.1
- Transformers 4.41.1
- Pytorch 2.3.0+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1