GGUF Defect? | unknown pre-tokenizer type: 'smaug-bpe'
Hi, I tried two different Quantizations with the same error:
\L3-SnowStorm-v1.15-4x8B-A¼zO¤rllama_model_load: error loading model: error loading model vocabulary: unknown pre-tokenizer type: 'smaug-bpe'
llama_load_model_from_file: failed to load model
Traceback (most recent call last):
File "koboldcpp.py", line 3540, in
File "koboldcpp.py", line 3231, in main
File "koboldcpp.py", line 411, in load_model
OSError: exception: access violation reading 0x0000000000000074
[17624] Failed to execute script 'koboldcpp' due to unhandled exception!
I tried "L3-SnowStorm-v1.15-4x8B-A.i1-Q4_K_S.gguf" and "L3-SnowStorm-v1.15-4x8B-A.i1-IQ4_XS.gguf"
With the latest Koboldcpp Client 1.66.1
Am I doing something wrong?
Your koboldcpp is outdated, you need a newer version that supports smaug.