microsoft/Florence-2-large tuned on Ejafa/ye-pop captioned with CogVLM2

This repository contains a fine-tuned version of the microsoft/Florence-2-large model. The model has been tuned on a 40,000 image subset of the Ejafa/ye-pop dataset, with captions generated using THUDM/cogvlm2-llama3-chat-19B.

Training Details

  • Vision Encoder: The vision encoder was frozen during training.
  • Batch Size: 64
  • Gradient Accumulation Steps: 16
  • Learning Rate: 5.12e-05
  • Optimizer: AdamW
  • Scheduler: polynomial
  • Epochs: 8.36

Dataset

The fine-tuning process utilized a 40,000 image subset from the Ejafa/ye-pop dataset. This dataset contains a wide array of images with varying subjects, providing a robust training ground for improving the model's captioning abilities.

Captioning

The captions were generated using THUDM/cogvlm2-llama3-chat-19B and then post-processed with google/gemma-2-9b to remove vagueness.

Usage

To use this model, you can load it directly from the Hugging Face Model Hub:

from transformers import AutoModelForCausalLM, AutoProcessor, AutoConfig
import torch
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model = AutoModelForCausalLM.from_pretrained("thwri/CogFlorence-2.2-Large", trust_remote_code=True).to(device).eval()
processor = AutoProcessor.from_pretrained("thwri/CogFlorence-2.2-Large", trust_remote_code=True)
# Function to run the model on an example
def run_example(task_prompt, image):
    prompt = task_prompt
    # Ensure the image is in RGB mode
    if image.mode != "RGB":
        image = image.convert("RGB")
    inputs = processor(text=prompt, images=image, return_tensors="pt").to(device)
    generated_ids = model.generate(
        input_ids=inputs["input_ids"],
        pixel_values=inputs["pixel_values"],
        max_new_tokens=1024,
        num_beams=3,
        do_sample=True
    )
    generated_text = processor.batch_decode(generated_ids, skip_special_tokens=False)[0]
    parsed_answer = processor.post_process_generation(generated_text, task=task_prompt, image_size=(image.width, image.height))
    return parsed_answer
from PIL import Image
import requests
import copy
url = "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers/tasks/car.jpg?download=true"
image = Image.open(requests.get(url, stream=True).raw)
result = run_example("<MORE_DETAILED_CAPTION>" , image)
print(result)
# {'<MORE_DETAILED_CAPTION>': 'A vivid portrayal of a classic Volkswagen Beetle parked on a cobblestone street. The car is painted a vibrant turquoise, contrasting with the muted yellow of the building behind it. The building has two wooden doors, one with a white frame and the other with a dark brown finish. The sky is clear, and the sun casts a warm glow on the scene, highlighting the car's details. The image evokes a nostalgic and nostalgic mood, capturing a moment in time without posed elements.'}
Downloads last month
5,868
Safetensors
Model size
823M params
Tensor type
FP16
Β·
Inference Examples
Inference API (serverless) does not yet support model repos that contain custom code.

Model tree for thwri/CogFlorence-2.2-Large

Finetuned
(6)
this model

Dataset used to train thwri/CogFlorence-2.2-Large

Spaces using thwri/CogFlorence-2.2-Large 2