uv-scripts/vllm
vLLM Inference Scripts Ready-to-run UV scripts for GPU-accelerated inference using vLLM. These scripts use UV's inline script metadata to automatically manage dependencies - just run with uv run and everything installs automatically! 📋 Available Scripts vlm-classify.py Vision Language Model (VLM) image classification with structured output constraints. Features: 🖼️ Process images through state-of-the-art VLMs (Qwen2-VL) 🎯 Structured… See the full description on the dataset page: https://huggingface.co/datasets/uv-scripts/vllm.
Sync from GitHub via hub-sync
Add hf-jobs tag to README frontmatter
Refactor image_to_data_uri and main functions for improved readability and consistency
Add --max-size flag for image resizing
Add VLM image classification script
pin xet
xet
Add plain text prompt support and sample limiting to generate-responses.py
Remove requirements and performance tips sections from README.md for clarity and conciseness.
secret
Update README.md with enhanced usage instructions for classify-dataset.py and generate-responses.py, including multi-GPU support and environment variable details.
Remove viewer tag from dataset card creation
ruff format
Enhance dataset card creation with filtering statistics and add option to skip long prompts
Add flashinfer-python dependency
Adjust GPU memory utilization default and add max model length argument in generate-responses.py
Update model version and parameters in generate-responses.py
update
README.md
Improve README with UV context and clarify BERT-style models
Add vLLM classification script
initial commit