metadata
license: llama2
pipeline_tag: text-generation
tags:
- llama-2
Description
Converted to f16 using llama_cpp convert.py script, then quantized to q6_K using quantize from the same llama_cpp repository.
Resulting file was split into 2 parts.
Note: HF does not support uploading files larger than 50GB.
File require joining
To join the files, do the following:
cat codellama-70b-python-q6_K.gguf-split-* > codellama-70b-python-q6_K.gguf && rm codellama-70b-python-q6_K.gguf-split-*