im2latex_model

This model is a VisionEncoderDecoderModel trained on a dataset for generating LaTeX formulas from images. This is part of a project that reproduces the following paper: https://arxiv.org/html/2408.04015v1. NOTE: In the paper, the model is finetuned on handwritten data after training. This is the model before finetuning.

Model Details

Encoder: Swin Transformer
Decoder: GPT-2
Framework: PyTorch

Training Data

The data is taken from OleehyO/latex-formulas. The data was divided into 80:10:10 for train, val and test. The splits were made as follows:

dataset = load_dataset(OleehyO/latex-formulas, cleaned_formulas)
train_val_split = dataset["train"].train_test_split(test_size=0.2, seed=42)
train_ds = train_val_split["train"]
val_test_split = train_val_split["test"].train_test_split(test_size=0.5, seed=42)
val_ds = val_test_split["train"]
test_ds = val_test_split["test"]

Evaluation Metrics

The model was evaluated on a test set with the following results:

Test Loss: TBA
Test BLEU Score: ~0.7

Usage

You can use the model directly with the transformers library:

from transformers import VisionEncoderDecoderModel, AutoTokenizer, AutoFeatureExtractor
import torch
from PIL import Image

# Load model, tokenizer, and feature extractor
model = VisionEncoderDecoderModel.from_pretrained("your-username/your-model-name")
tokenizer = AutoTokenizer.from_pretrained("your-username/your-model-name")
feature_extractor = AutoFeatureExtractor.from_pretrained("your-username/your-model-name")

# Prepare an image
image = Image.open("path/to/your/image.png")
pixel_values = feature_extractor(images=image, return_tensors="pt").pixel_values

# Generate LaTeX formula
generated_ids = model.generate(pixel_values)
generated_texts = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)

print("Generated LaTeX formula:", generated_texts[0])

Training Script

The training script for this model can be found in the following repository: GitHub

Matthijs0
/

im2latex_base