---
library_name: transformers
license: mit
base_model: camembert-base
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: camembert-math-classification
  results: []
---

<!-- 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. -->

# camembert-math-classification

This model is a fine-tuned version of [camembert-base](https://huggingface.co/camembert-base) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 1.6519
- F1 Macro: 0.6973
- F1 Weighted: 0.7076
- F1 Correct: 0.8571
- F1 Partiel: 0.8
- F1 Incorrect: 0.4348
- Accuracy: 0.75

## 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: 32
- eval_batch_size: 64
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 0.1
- num_epochs: 4
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step | Validation Loss | F1 Macro | F1 Weighted | F1 Correct | F1 Partiel | F1 Incorrect | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:-----------:|:----------:|:----------:|:------------:|:--------:|
| No log        | 1.0   | 15   | 2.1103          | 0.1905   | 0.2286      | 0.0        | 0.5714     | 0.0          | 0.4      |
| 2.1496        | 2.0   | 30   | 1.9187          | 0.2928   | 0.3228      | 0.2857     | 0.5926     | 0.0          | 0.45     |
| 2.0111        | 3.0   | 45   | 1.7323          | 0.7855   | 0.7870      | 0.9412     | 0.8        | 0.6154       | 0.8      |
| 1.8019        | 4.0   | 60   | 1.6519          | 0.6973   | 0.7076      | 0.8571     | 0.8        | 0.4348       | 0.75     |


### Framework versions

- Transformers 5.0.0
- Pytorch 2.10.0+cu128
- Datasets 5.0.0
- Tokenizers 0.22.2