技能库 37 个

Optimizes transformer attention with Flash Attention for 2-4x speedup and 10-20x memory reduction. Use when training/run

Facebook's library for efficient similarity search and clustering of dense vectors. Supports billions of vectors, GPU ac

OpenAI's model connecting vision and language. Enables zero-shot image classification, image-text matching, and cross-mo

Open-source embedding database for AI applications. Store embeddings and metadata, perform vector and full-text search,

Simplest distributed training API. 4 lines to add distributed support to any PyTorch script. Unified API for DeepSpeed/F

SAM: zero-shot image segmentation via points, boxes, masks.

AudioCraft: MusicGen text-to-music, AudioGen text-to-sound.

vLLM: high-throughput LLM serving, OpenAI API, quantization.

OBLITERATUS: abliterate LLM refusals (diff-in-means).

llama.cpp local GGUF inference + HF Hub model discovery.

HuggingFace hf CLI: search/download/upload models, datasets.

W&B: log ML experiments, sweeps, model registry, dashboards.

lm-eval-harness: benchmark LLMs (MMLU, GSM8K, etc.).