Git 20

🚀 git_20

This model is fine - tuned based on Microsoft GIT, aiming to create a machine - learning pipeline for providing automatic diagnostic feedback on students' mathematical reasoning using multi - modal and multi - task deep - learning models.

🚀 Quick Start

This model is fine - tuned with Microsoft GIT using 1 Nvidia A100 - 80G GPU. We selected 100,000 student assignments with teacher feedback from 3 million student assignments as training data. The training data is split into the image part of student assignments and the text part of teacher feedback. git_20 has 18 layers and over 170 million parameters, taking up to 0.7 gigabytes of disk space. The project's goal is to utilize multi - modal and multi - task deep learning models to build a machine learning pipeline that offers automatic diagnostic feedback for students' mathematical reasoning. Researchers can experiment with and fine - tune the model to assist in constructing multi - models capable of effectively providing such feedback.

💻 Usage Examples

Basic Usage

from transformers import AutoModelForCausalLM
from transformers import AutoProcessor
from PIL import Image
model = AutoModelForCausalLM.from_pretrained("Fan21/git_20")
processor = AutoProcessor.from_pretrained("Fan21/git_20")

image_path ='Please enter the image address here'
image = Image.open(image_path)
width, height = image.size
display(image.resize((int(1 * width), int(1 * height))))
pixel_values = processor(images=image, return_tensors="pt").pixel_values
with torch.no_grad():
    outputs = model.generate(pixel_values=pixel_values, max_length=50)

answer = tokenizer.decode(outputs[0], skip_special_tokens=True)

📄 License

This project is licensed under the MIT license.