Skip to content

Conversation

kustaaya
Copy link
Contributor

@kustaaya kustaaya commented Jun 26, 2024

Implemented suggested changes from discussion in pull request #7328 by @akx . Added pre-tokenizer support for Viking-models.

  • Added tokenizer type for Viking-7B in convert-hf-to-gguf-update.py
  • Added the chkhsh for Viking-7B in convert-hf-to-gguf.py
  • Added LLAMA_VOCAB_PRE_TYPE_VIKING enum to llama.h
  • Added pre-tokenizer regex for LLAMA_VOCAB_PRE_TYPE_VIKING to llama.cpp
  • Ran ./tests/test-tokenizer-0 ./models/ggml-vocab-viking.gguf. Tests passed.

Tokenizer is the same for Viking-7B and its 13B and 33B siblings. Fixes issue #7309

@github-actions github-actions bot added the python python script changes label Jun 26, 2024
@slaren
Copy link
Member

slaren commented Jun 27, 2024

The changes look correct me, but there are conflicts that need to be resolved before this can be merged. I think it is just because llama.h and llama.cpp were moved to a different directory.

@kustaaya
Copy link
Contributor Author

The changes look correct me, but there are conflicts that need to be resolved before this can be merged. I think it is just because llama.h and llama.cpp were moved to a different directory.

Thanks! I updated the pull request to fix the conflicts.

@slaren slaren merged commit f675b20 into ggml-org:master Jun 27, 2024
Nexesenex pushed a commit to Nexesenex/croco.cpp that referenced this pull request Jun 28, 2024
Nexesenex pushed a commit to Nexesenex/croco.cpp that referenced this pull request Jun 28, 2024
arthw pushed a commit to arthw/llama.cpp that referenced this pull request Jun 30, 2024
@akx akx mentioned this pull request Jul 1, 2024
MagnusS0 pushed a commit to MagnusS0/llama.cpp-normistral-tokenizer that referenced this pull request Jul 1, 2024
Nexesenex pushed a commit to Nexesenex/croco.cpp that referenced this pull request Feb 25, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

python python script changes

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants