Package: wnpp
Severity: wishlist
Owner: Juan Manuel Méndez Rey <[email protected]>
X-Debbugs-Cc: [email protected], [email protected]

* Package name    : golang-github-gpustack-gguf-parser-go
  Version         : 0.26.3
  Upstream Contact: https://github.com/gpustack/gguf-parser-go/issues
* URL             : https://github.com/gpustack/gguf-parser-go
* License         : Expat
  Programming Lang: Go
  Description     : parse GGUF model files and estimate memory usage

A Go library for parsing GGUF files, the binary model format used by
 GGML and llama.cpp. It reads a model's metadata and tensor layout, estimates
 the memory required to run it, and can predict maximum tokens per second from
 supplied device metrics.
 .
 Because it reads remote files in chunks, it can inspect a model hosted on
 Hugging Face, ModelScope or an Ollama registry without downloading it in full.
 It also classifies what a GGUF file is for, distinguishing embedding,
 reranking, LoRA and diffusion models.
 .
 This package is a build dependency of syft (ITP 
https://bugs.debian.org/1124819).

I intend to maintain this package within the Debian Go Packaging Team.

Reply via email to