metadata
inference: false
language:
- en
pipeline_tag: text-generation
tags:
- llama
- llama-2
airoboros-limarpv3-l2-70b-gpt4-1.4.1-exl2
Exllama v2 quant of Panchovix/airoboros-l2-70b-gpt4-1.4.1-limarpv3-qlora
Branches:
- main: measurement.json calculated at 2048 token calibration rows on PIPPA
- 5.0bpw-h6: 5 decoder bits per weight, 6 head bits
- ideal for 2x 24gb GPUs at 8192 context, or 1x 48gb GPU at 8192 context with CFG cache