metadata
license: apache-2.0
base_model: mistralai/Mistral-7B-v0.1
tags:
- generated_from_trainer
model-index:
- name: qlora-out
results: []
qlora-out
This model is a fine-tuned version of mistralai/Mistral-7B-v0.1 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.5840
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: 0.0003
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 300
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.767 | 0.24 | 20 | 0.6343 |
| 0.6849 | 0.48 | 40 | 0.5669 |
| 0.6761 | 0.72 | 60 | 0.5247 |
| 0.5534 | 0.96 | 80 | 0.5044 |
| 0.4757 | 1.2 | 100 | 0.5023 |
| 0.5158 | 1.44 | 120 | 0.4883 |
| 0.5414 | 1.68 | 140 | 0.4809 |
| 0.4715 | 1.92 | 160 | 0.4748 |
| 0.4037 | 2.16 | 180 | 0.4873 |
| 0.4213 | 2.4 | 200 | 0.5194 |
| 0.2988 | 2.64 | 220 | 0.6278 |
| 0.3477 | 2.88 | 240 | 0.5840 |
Framework versions
- Transformers 4.34.1
- Pytorch 2.0.1+cu118
- Datasets 2.14.6
- Tokenizers 0.14.1