tonyassi / git-base-luc1

git-base-lucy1

This model is a fine-tuned version of microsoft/git-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 5.9368
• Wer Score: 3.1310

Model description

Fine-tuned captioning model on Lucy in the Sky images.

Dataset: tonyassi/lucy-caption-2

Usage

import torch
from PIL import Image
from transformers import AutoProcessor, AutoModelForCausalLM
import requests

# Load model directly
processor = AutoProcessor.from_pretrained("tonyassi/git-base-lucy1")
model = AutoModelForCausalLM.from_pretrained("tonyassi/git-base-lucy1")

# Load image
url = "https://datasets-server.huggingface.co/cached-assets/tonyassi/lucy-caption-2/--/94d2ffc965a7a0a50beebbeb60d04fa38a24ff66/--/default/train/6/image/image.jpg?Expires=1727109954&Signature=IMpoIBQV-ICPaC8V4NF2SUn0OQE7cOtIJZIeGSpOQLifkjlXl3rx6CAukc2Ax3Gkl4eQ3BfcDrnV2HNzE-c3c5WC5lcj30PWTkSczcqN7YtkKGFHOxlS6-Gv8zotQw8NJPn0d-LoCERHlxA75Sbh8vF8X6DE1SCRJIitT3DFcObTdKpZpHYDv21BYq4-A4EN04wX6aKHWyz8xR0NorlOtcB8dzPdsSpRGy3gcgLU9kHeBNWpa22KsMDJmDP8QferzrnG5bnb5fi9RxrMCTURCPUB8AyNJ1mVwuAorW4GJIdm40UEoqanQzrST3hIp-dTEH47w5-GY5PnOrWUcaxYGQ__&Key-Pair-Id=K3EI6M078Z3AC3"
image = Image.open(requests.get(url, stream=True).raw)

# GPU or CPU
device = "cuda" if torch.cuda.is_available() else "cpu"

# Inference
inputs = processor(images=image, return_tensors="pt").to(device)
pixel_values = inputs.pixel_values
generated_ids = model.generate(pixel_values=pixel_values, max_length=50)
generated_caption = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
print(generated_caption)

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: 5e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 50
• mixed_precision_training: Native AMP

Training results

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1