The Vocab size mismatch[BUG]
xwjim opened this issue · 0 comments
xwjim commented
Describe the bug
When I finish training the llama model, I use the llamacpp to quantize the model. There raise a error "Exception: Vocab size mismatch (model has 39424, but ../LMFlow/output_models/test/tokenizer.model combined with ../LMFlow/output_models/test/added_tokens.json has 39410).". Additional 14 tokens are added.
Expected behavior
A clear and concise description of what you expected to happen.
Screenshots
Desktop (please complete the following information):
- OS: Ubuntu
- Browser
- Version .22