metadata
license: apache-2.0
base_model: mistralai/Mistral-7B-Instruct-v0.1
tags:
- generated_from_trainer
model-index:
- name: Mistral-7B-Instruct-v0.1-LC-noSW
results: []
Mistral-7B-Instruct-v0.1-LC-noSW
This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.1 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.8686
Model description
This model is a fine-tuning of Mistral-7B-Instruct-v0.1. This FT was donne with full attention (removing the 4k SWA).
Intended uses & limitations
More information needed
Training and evaluation data
Data is a 9k sample from the RedPajama datset. The context is <=7200 with a decreasing exponential distribution of scale 1500.
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 3e-05
- train_batch_size: 1
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 32
- 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: 20
- training_steps: 300
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 2.0233 | 0.18 | 50 | 1.8792 |
| 2.0805 | 0.36 | 100 | 1.8732 |
| 1.9414 | 0.55 | 150 | 1.8705 |
| 1.8302 | 0.73 | 200 | 1.8692 |
| 2.015 | 0.91 | 250 | 1.8686 |
| 1.894 | 1.09 | 300 | 1.8686 |
Framework versions
- Transformers 4.34.1
- Pytorch 2.0.0+cu117
- Datasets 2.14.6
- Tokenizers 0.14.1