Hugging Face's logo

Spaces:
ggml-org
/
gguf-my-repo
Running on A10G

App Files Community
203

GGUF My Repo re-design

#187
by olegshulyakov - opened
base: refs/heads/main
←
from: refs/pr/187
Discussion Files changed
+5894
-2598
olegshulyakov
•
edited Aug 10, 2025
  • Migrate Docker image to official llama.cpp CUDA image.
  • Re-write app.py to OOP to re-design methods signatures.
  • Added additional llama-quantize options: --token-embedding-type, --leave-output-tensor, --output-tensor-type
  • Customizable output options: repo name, file name
  • Upload to different quants to the same repository.
  • Updated imatrix training file to calibration_data_v5_rc.txt.
Migrate Docker to official llama.cpp CUDA imageb7bd975a
Re-write app to OOP with additional optionsa3b65e28
olegshulyakov changed pull request status to open
Fix Dockerfile envs881df402
Fix train_data_file usageefd93397
Cannot merge
This branch has merge conflicts in the following files:
  • app.py

Sign up or log in to comment