metadata

license: llama3
language:
  - en
library_name: CTranslate2
pipeline_tag: text-generation
tags:
  - facebook
  - meta
  - llama
  - llama-3
  - kaltcit
  - cat
  - ct2
  - quantized model
  - int8
base_model: turboderp/llama3-turbcat-instruct-8b

CTranslate2 int8 version of turbcat 8b

This is a int8_float16 quantization of turbcat 8b
See more on CTranslate2: Docs | Github

This model and it's dataset was created by Kaltcit, an admin of the Exllama Discord server.

This model was converted to ct2 format using the following commnd:

ct2-transformers-converter --model kat_turbcat --output_dir turbcat-ct2 --quantization int8_float16 --low_cpu_mem_usage

no converstion needed using the model from this repository as it is already in ct2 format.