Weightless

Inspect models without loading the weights. Header-only reads open multi-GB .safetensors / .gguf checkpoints instantly — a 500 GB Hub model costs a few hundred KB.

or drag & drop a .safetensors / .gguf file here (drop its config.json alongside for extra detail) — files never leave your machine; only the header is read

Try: gpt2 · Qwen2.5-0.5B · Unlimited-OCR

VS Code extension · Open VSX · GitHub