Instructions to use HuggingFaceM4/idefics2-8b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use HuggingFaceM4/idefics2-8b with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("image-text-to-text", model="HuggingFaceM4/idefics2-8b")

# Load model directly
from transformers import AutoProcessor, AutoModelForImageTextToText

processor = AutoProcessor.from_pretrained("HuggingFaceM4/idefics2-8b")
model = AutoModelForImageTextToText.from_pretrained("HuggingFaceM4/idefics2-8b")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use HuggingFaceM4/idefics2-8b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "HuggingFaceM4/idefics2-8b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "HuggingFaceM4/idefics2-8b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/HuggingFaceM4/idefics2-8b

SGLang

How to use HuggingFaceM4/idefics2-8b with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "HuggingFaceM4/idefics2-8b" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "HuggingFaceM4/idefics2-8b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "HuggingFaceM4/idefics2-8b" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "HuggingFaceM4/idefics2-8b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use HuggingFaceM4/idefics2-8b with Docker Model Runner:
```
docker model run hf.co/HuggingFaceM4/idefics2-8b
```

After fine tuning, there is a problem for using it.

#50

by SalmanFaroz - opened May 15, 2024

Discussion

SalmanFaroz

May 15, 2024

I've fine tuned the model with DocVQA Dataset, after finetuning pushed to hub

https://huggingface.co/SalmanFaroz/idefics2-8b-DocVQA-SP

transformers version == 4.40.0

When I use it in private, it works as expected.

however, if I make it public, I receive the error message "config file not found."

let me know what i'm missing , and how can i load the model ?

thanks in advance

HugoLaurencon

May 15, 2024

You don't seem to have this file https://huggingface.co/HuggingFaceM4/idefics2-8b/blob/main/config.json on your repo, did you delete it?

SalmanFaroz

May 16, 2024

•

edited May 16, 2024

@HugoLaurencon No, I haven't removed it. After fine-tuning, there wasn't a file named config.json, but when I use it in private with the hugging face access token, the adaptor_config is downloaded first, followed by config.json. It works properly

but i tried manually copied the config.json from idefics2 to my repo , but still it wasn't working so i've deleted it

Update : @HugoLaurencon now again i tried adding that config.json and then i get

OSError: SalmanFaroz/idefics2-8b-DocVQA-SP does not appear to have a file named pytorch_model.bin, tf_model.h5, model.ckpt or flax_model.msgpack.

HugoLaurencon

May 17, 2024

•

edited May 17, 2024

Yes it's because you don't have any model uploaded on your repo, you currently only have the adapters weights (that I suppose you trained with Lora)!
You need to upload the original weights of Idefics2 too

VictorSanh

May 17, 2024

@SalmanFaroz , you should not need the og idefics2 weights, just the adapter weights are sufficient because in the adapter_config.json file, the base model to which the dapter weights is specified.
i am also able to load your (public) adapted model: from transformers import Idefics2ForConditionalGeneration; Idefics2ForConditionalGeneration.from_pretrained("SalmanFaroz/idefics2-8b-DocVQA-SP").
can you try again from scratch? delete your model, and re-upload it to the hub.
for instance, see my debugging model: https://huggingface.co/VictorSanh/idefics2-8b-docvqa-finetuned-tutorial

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment