---
library_name: peft
license: other
base_model: mistralai/Ministral-8B-Instruct-2410
tags:
- generated_from_trainer
model-index:
- name: Mistral-8B-Instruct-2410-009
  results: []
datasets:
- jdavit/colombian-conflict-SQA
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# Mistral-8B-Instruct-2410-009

Este modelo es un afnamiento del modelo [mistralai/Ministral-8B-Instruct-2410](https://huggingface.co/mistralai/Ministral-8B-Instruct-2410) sobre
un dataset del acuerdo 009 de la univalle con 8 mil millones de parámetros a 10 épocas, con un batch_size de 1 y ocupando por completo
toda la memoria VRAM de la GPU de 24 Gz, logrando una función de pérdida de:
- Loss: 0.3784

## 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: 1
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 4
- optimizer: Use paged_adamw_8bit with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 10
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch  | Step | Validation Loss |
|:-------------:|:------:|:----:|:---------------:|
| 1.4242        | 0.2694 | 100  | 1.3903          |
| 1.2006        | 0.5387 | 200  | 1.2413          |
| 1.1104        | 0.8081 | 300  | 1.1213          |
| 0.9163        | 1.0754 | 400  | 1.0207          |
| 0.8715        | 1.3448 | 500  | 0.9261          |
| 0.7834        | 1.6141 | 600  | 0.8410          |
| 0.8636        | 1.8835 | 700  | 0.7739          |
| 0.7017        | 2.1508 | 800  | 0.7145          |
| 0.6127        | 2.4202 | 900  | 0.6664          |
| 0.6409        | 2.6896 | 1000 | 0.6310          |
| 0.584         | 2.9589 | 1100 | 0.5932          |
| 0.5592        | 3.2263 | 1200 | 0.5683          |
| 0.4336        | 3.4956 | 1300 | 0.5447          |
| 0.5287        | 3.7650 | 1400 | 0.5292          |
| 0.4449        | 4.0323 | 1500 | 0.5119          |
| 0.4894        | 4.3017 | 1600 | 0.4994          |
| 0.436         | 4.5710 | 1700 | 0.4799          |
| 0.3756        | 4.8404 | 1800 | 0.4676          |
| 0.3174        | 5.1077 | 1900 | 0.4545          |
| 0.3721        | 5.3771 | 2000 | 0.4475          |
| 0.3813        | 5.6465 | 2100 | 0.4367          |
| 0.3972        | 5.9158 | 2200 | 0.4281          |
| 0.354         | 6.1832 | 2300 | 0.4244          |
| 0.3299        | 6.4525 | 2400 | 0.4206          |
| 0.4017        | 6.7219 | 2500 | 0.4112          |
| 0.3103        | 6.9912 | 2600 | 0.4060          |
| 0.3299        | 7.2586 | 2700 | 0.4060          |
| 0.3874        | 7.5279 | 2800 | 0.3989          |
| 0.3838        | 7.7973 | 2900 | 0.3940          |
| 0.3446        | 8.0646 | 3000 | 0.3907          |
| 0.2674        | 8.3340 | 3100 | 0.3887          |
| 0.2681        | 8.6034 | 3200 | 0.3839          |
| 0.2922        | 8.8727 | 3300 | 0.3806          |
| 0.3125        | 9.1401 | 3400 | 0.3820          |
| 0.3042        | 9.4094 | 3500 | 0.3802          |
| 0.2623        | 9.6788 | 3600 | 0.3790          |
| 0.3382        | 9.9481 | 3700 | 0.3784          |


### Framework versions

- PEFT 0.15.2
- Transformers 4.51.3
- Pytorch 2.6.0+cu126
- Datasets 3.5.0
- Tokenizers 0.21.1