---
library_name: transformers
base_model: uitnlp/visobert
tags:
- generated_from_trainer
metrics:
- accuracy
- precision
- recall
- f1
model-index:
- name: visobert-human-tl-seg-seed-42
  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. -->

# visobert-human-tl-seg-seed-42

This model is a fine-tuned version of [uitnlp/visobert](https://huggingface.co/uitnlp/visobert) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3753
- Accuracy: 0.8615
- Precision: 0.6981
- Recall: 0.5887
- F1: 0.6154

## 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: 0.0001
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH 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: 500
- num_epochs: 50

### Training results

| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1     |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|
| No log        | 1.0   | 346  | 0.4133          | 0.8477   | 0.6961    | 0.4837 | 0.5065 |
| 0.4914        | 2.0   | 692  | 0.3963          | 0.8570   | 0.7033    | 0.5283 | 0.5612 |
| 0.3761        | 3.0   | 1038 | 0.3948          | 0.8578   | 0.7199    | 0.5285 | 0.5656 |
| 0.3761        | 4.0   | 1384 | 0.3859          | 0.8582   | 0.7133    | 0.5452 | 0.5769 |
| 0.3709        | 5.0   | 1730 | 0.3851          | 0.8608   | 0.7156    | 0.5520 | 0.5850 |
| 0.3629        | 6.0   | 2076 | 0.3886          | 0.8574   | 0.7250    | 0.5445 | 0.5762 |
| 0.3629        | 7.0   | 2422 | 0.3781          | 0.8612   | 0.7065    | 0.5498 | 0.5921 |
| 0.3634        | 8.0   | 2768 | 0.3760          | 0.8619   | 0.7147    | 0.5534 | 0.5921 |
| 0.3549        | 9.0   | 3114 | 0.3724          | 0.8608   | 0.6930    | 0.5689 | 0.6070 |
| 0.3549        | 10.0  | 3460 | 0.3875          | 0.8585   | 0.7057    | 0.5612 | 0.5835 |
| 0.3614        | 11.0  | 3806 | 0.3738          | 0.8604   | 0.7037    | 0.5570 | 0.5975 |
| 0.3513        | 12.0  | 4152 | 0.3753          | 0.8615   | 0.6981    | 0.5887 | 0.6154 |
| 0.3513        | 13.0  | 4498 | 0.3736          | 0.8619   | 0.7121    | 0.5563 | 0.5926 |
| 0.3533        | 14.0  | 4844 | 0.3705          | 0.8604   | 0.6942    | 0.5570 | 0.5963 |
| 0.3476        | 15.0  | 5190 | 0.3711          | 0.8619   | 0.7094    | 0.5619 | 0.6016 |
| 0.351         | 16.0  | 5536 | 0.3724          | 0.8600   | 0.7204    | 0.5401 | 0.5816 |
| 0.351         | 17.0  | 5882 | 0.3706          | 0.8582   | 0.6790    | 0.5826 | 0.6105 |


### Framework versions

- Transformers 4.51.1
- Pytorch 2.5.1+cu124
- Datasets 3.5.0
- Tokenizers 0.21.0