[GH-ISSUE #9486] ollama create copies the entire Hugging Face safetensors to TEMP before conversion #68233

Closed
opened 2026-05-04 12:57:36 -05:00 by GiteaMirror · 1 comment
Owner

Originally created by @lowlyocean on GitHub (Mar 4, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/9486

ollama create will copy a locally cloned HuggingFace repo to %TEMP% as part of the conversion to GGUF and quantization process - is there any way to avoid this duplication of the Safetensors files (to avoid filling up disk space)?

Originally created by @lowlyocean on GitHub (Mar 4, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/9486 `ollama create` will copy a locally cloned HuggingFace repo to %TEMP% as part of the conversion to GGUF and quantization process - is there any way to avoid this duplication of the Safetensors files (to avoid filling up disk space)?
GiteaMirror added the bug label 2026-05-04 12:57:36 -05:00
Author
Owner

@rick-github commented on GitHub (Mar 4, 2025):

Convert with llama.cpp first, then import the quantized GGUF file.

<!-- gh-comment-id:2695987177 --> @rick-github commented on GitHub (Mar 4, 2025): Convert with llama.cpp first, then import the quantized GGUF file.
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#68233