Instructions to use neoALI/ARC-Hunyuan-Video-7B-Emotion with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use neoALI/ARC-Hunyuan-Video-7B-Emotion with PEFT:

from peft import PeftModel
from transformers import AutoModelForCausalLM

base_model = AutoModelForCausalLM.from_pretrained("TencentARC/ARC-Hunyuan-Video-7B")
model = PeftModel.from_pretrained(base_model, "neoALI/ARC-Hunyuan-Video-7B-Emotion")

Transformers

How to use neoALI/ARC-Hunyuan-Video-7B-Emotion with Transformers:

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

pipe = pipeline("text-generation", model="neoALI/ARC-Hunyuan-Video-7B-Emotion")

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("neoALI/ARC-Hunyuan-Video-7B-Emotion", dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use neoALI/ARC-Hunyuan-Video-7B-Emotion with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "neoALI/ARC-Hunyuan-Video-7B-Emotion"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "neoALI/ARC-Hunyuan-Video-7B-Emotion",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/neoALI/ARC-Hunyuan-Video-7B-Emotion

SGLang

How to use neoALI/ARC-Hunyuan-Video-7B-Emotion 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 "neoALI/ARC-Hunyuan-Video-7B-Emotion" \
    --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": "neoALI/ARC-Hunyuan-Video-7B-Emotion",
		"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 "neoALI/ARC-Hunyuan-Video-7B-Emotion" \
        --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": "neoALI/ARC-Hunyuan-Video-7B-Emotion",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use neoALI/ARC-Hunyuan-Video-7B-Emotion with Docker Model Runner:
```
docker model run hf.co/neoALI/ARC-Hunyuan-Video-7B-Emotion
```

ARC-Hunyuan-Video-7B-Emotion

A fine-tuned version of TencentARC/ARC-Hunyuan-Video-7B specialized for emotion classification in videos.

Model Description

This model is a LoRA adapter fine-tuned on the ARC-Hunyuan-Video-7B base model for emotion classification tasks.

Key Features:

Task: Video emotion classification
Base Model: ARC-Hunyuan-Video-7B (7B parameters)
Training Method: LoRA (Low-Rank Adaptation)
Special Feature: Trained using LLM-generated feature descriptions of videos, enabling better understanding of emotional content

Model Details

Developed by: NEOALI
Model type: Video-language model with LoRA adapter
Language(s): English and Chinese
License: MIT
Fine-tuned from: TencentARC/ARC-Hunyuan-Video-7B

Training Details

Training regime: LoRA fine-tuning
LoRA rank: 8
LoRA alpha: 8
Training data: Videos with LLM-generated emotional feature descriptions

Usage

Requirements

pip install torch transformers peft

Loading the Model

from transformers import AutoModel, AutoTokenizer
from peft import PeftModel

# Load base model
base_model = AutoModel.from_pretrained("TencentARC/ARC-Hunyuan-Video-7B")

# Load LoRA adapter
model = PeftModel.from_pretrained(base_model, "neoALI/ARC-Hunyuan-Video-7B-Emotion")

# Load tokenizer
tokenizer = AutoTokenizer.from_pretrained("TencentARC/ARC-Hunyuan-Video-7B")

Intended Use

This model is designed for:

Emotion classification in short videos (up to 5 minutes)
Understanding emotional content in user-generated videos
Video content analysis requiring emotional intelligence

Limitations

Inherits limitations from the base ARC-Hunyuan-Video-7B model
Best performance on videos up to 5 minutes in length
Optimized for emotion classification; may require additional fine-tuning for other tasks

Acknowledgements

This model is built upon ARC-Hunyuan-Video-7B by TencentARC. We thank the original authors for their excellent work.

Downloads last month: 4

Model tree for neoALI/ARC-Hunyuan-Video-7B-Emotion

Base model

TencentARC/ARC-Hunyuan-Video-7B

Adapter

(1)

this model