]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
convert : ability to lazy-load safetensors remotely without downloading to disk ...
authorXuan-Son Nguyen <redacted>
Thu, 10 Apr 2025 15:24:44 +0000 (17:24 +0200)
committerGitHub <redacted>
Thu, 10 Apr 2025 15:24:44 +0000 (17:24 +0200)
commit64eda5deb9859e87a020e56bab5d2f9ca956f1de
tree167816702ec3aba3ce68089c17207d221672a82e
parentfe5b78c89670b2f37ecb216306bed3e677b49d9f
convert : ability to lazy-load safetensors remotely without downloading to disk (#12820)

* gguf util : add SafetensorRemote

* fix style

* convert: add --remote option

* convert : allow using lazy remote tensors

It's a bit slow for now since everything is blocking and single-threaded.

* correct metadata.name

* small style fix

* support HF_TOKEN

* convert : use writeable buffer for remote lazy tensors

* convert : fix flake8 lint regarding lamdba assigment

* multithreaded download

* multithread: print debug

* fix style

* Revert "multithreaded download"

This reverts commit 42fc895ace385edc972ad819c76c704aeea61791.

* bring back _get_request_headers

---------

Co-authored-by: Francis Couture-Harpin <redacted>
convert_hf_to_gguf.py
gguf-py/gguf/utility.py