Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

k-quants : remove unnecessary tensor shape restrictions #2811

Merged
merged 1 commit into from
Aug 26, 2023

Conversation

ggerganov
Copy link
Owner

This is needed to improve the Code Llama 7B and 13B quantum mixtures as they have tensor rows of 32016 for the token_embd.weight and output.weight tensors

@ggerganov ggerganov merged commit 04f4b1e into master Aug 26, 2023
25 checks passed
@ggerganov ggerganov deleted the fix-code-llama-quantum-mixtures branch August 26, 2023 14:37
@slaren
Copy link
Collaborator

slaren commented Aug 26, 2023

@TheBloke this PR changes the quantization for all the models with an extended vocab.

@TheBloke
Copy link
Contributor

Thanks for the ping!

I have to re-do all the GGUFs soon anyway, for the uint64 thing. So I'll do them all as soon as that's out.

@ggerganov
Copy link
Owner Author

Ah yes, I'll fix that tomorrow likely

mattgauf added a commit to mattgauf/llama.cpp that referenced this pull request Aug 26, 2023
* master: (773 commits)
  server : add `/detokenize` endpoint (ggerganov#2802)
  convert.py : advanced option (ggerganov#2753)
  llama : use Unicode Escape Sequence to replace encoded characters (ggerganov#2814)
  flake.nix : add rocm support and cleanup (ggerganov#2808)
  llama : move #includes out of _GNU_SOURCE conditional (ggerganov#2817)
  main : fix bug (penalize_nl=false doesn't work) + suppress warning on mingw (ggerganov#1528)
  llama : use std::abs in llama_sample_tail_free (ggerganov#2800)
  k-quants : remove unnecessary tensor shape restrictions (ggerganov#2811)
  Better perplexity for 2- and 3-bit quantization for LLaMA-v2-70B (ggerganov#2807)
  Fix HellaSwag (ggerganov#2805)
  flake : build llama.cpp on Intel with nix (ggerganov#2795)
  Handle null rope scaling value (ggerganov#2793)
  Fix spm whitespaces (ggerganov#2806)
  examples : skip unnecessary external lib in server README.md how-to (ggerganov#2804)
  llama : fix struct decl (ggerganov#2790)
  Faster perplexity computation (ggerganov#2786)
  llama : add llama_beam_search() (ggerganov#2267)
  convert.py : Get rope scale from HuggingFace models (ggerganov#2772)
  llama-bench : add model sizes (ggerganov#2771)
  convert.py : export rope freq_base when converting CodeLlama from an HF model (ggerganov#2773)
  ...
akawrykow pushed a commit to akawrykow/llama.cpp that referenced this pull request Aug 29, 2023
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants