metadata
license: apache-2.0
library_name: peft
tags:
- generated_from_trainer
base_model: filipealmeida/Mistral-7B-Instruct-v0.1-sharded
model-index:
- name: mistral7b-finetune-20k-withnoclass
results: []
mistral7b-finetune-20k-withnoclass
This model is a fine-tuned version of filipealmeida/Mistral-7B-Instruct-v0.1-sharded on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.9941
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: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- training_steps: 2500
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.8094 | 0.04 | 100 | 1.4332 |
| 1.274 | 0.08 | 200 | 1.2291 |
| 1.1812 | 0.12 | 300 | 1.1865 |
| 1.1474 | 0.16 | 400 | 1.1542 |
| 1.1421 | 0.2 | 500 | 1.1370 |
| 1.101 | 0.24 | 600 | 1.1154 |
| 1.1111 | 0.29 | 700 | 1.0999 |
| 1.0778 | 0.33 | 800 | 1.0841 |
| 1.0421 | 0.37 | 900 | 1.0784 |
| 1.0468 | 0.41 | 1000 | 1.0644 |
| 1.0389 | 0.45 | 1100 | 1.0540 |
| 1.0023 | 0.49 | 1200 | 1.0434 |
| 1.0396 | 0.53 | 1300 | 1.0342 |
| 0.9991 | 0.57 | 1400 | 1.0260 |
| 1.0304 | 0.61 | 1500 | 1.0238 |
| 1.0033 | 0.65 | 1600 | 1.0159 |
| 1.0065 | 0.69 | 1700 | 1.0109 |
| 0.9587 | 0.73 | 1800 | 1.0072 |
| 0.9725 | 0.78 | 1900 | 1.0025 |
| 0.9738 | 0.82 | 2000 | 0.9997 |
| 0.9816 | 0.86 | 2100 | 0.9972 |
| 0.9858 | 0.9 | 2200 | 0.9956 |
| 0.9477 | 0.94 | 2300 | 0.9946 |
| 0.9834 | 0.98 | 2400 | 0.9941 |
| 0.943 | 1.02 | 2500 | 0.9941 |
Framework versions
- PEFT 0.7.1
- Transformers 4.37.0.dev0
- Pytorch 2.0.0
- Datasets 2.1.0
- Tokenizers 0.15.0