Astrohermes-3B / README.md
Aryanne's picture
Update README.md
e1e6580
metadata
language:
  - en
library_name: transformers
tags:
  - gpt
  - llm
  - stablelm
inference: true
license: cc-by-sa-4.0

This model is a mix of PAIXAI/Astrid-3B + jondurbin/airoboros-3b-3p0 + cxllin/StableHermes-3b, as shown in the yaml(see Astrohermes.yml or below). Aryanne/Astridboros-3B = PAIXAI/Astrid-3B + jondurbin/airoboros-3b-3p0

slices:
  - sources:
    - model: Aryanne/Astridboros-3B
      layer_range: [0, 15]
  - sources:
    - model: cxllin/StableHermes-3b
      layer_range: [15, 16]
  - sources:
    - model: Aryanne/Astridboros-3B
      layer_range: [16, 17]
  - sources:
    - model: cxllin/StableHermes-3b
      layer_range: [17, 18]
  - sources:
    - model: Aryanne/Astridboros-3B
      layer_range: [18, 19]
  - sources:
    - model: cxllin/StableHermes-3b
      layer_range: [19, 20]
  - sources:
    - model: Aryanne/Astridboros-3B
      layer_range: [20, 21]
  - sources:
    - model: cxllin/StableHermes-3b
      layer_range: [21, 22]
  - sources:
    - model: Aryanne/Astridboros-3B
      layer_range: [22, 23]
  - sources:
    - model: cxllin/StableHermes-3b
      layer_range: [23, 24]
  - sources:
    - model: Aryanne/Astridboros-3B
      layer_range: [24, 32]
merge_method: passthrough
dtype: float16

I recommend the use of alpaca prompt format.

GGUF Quants: afrideva/Astrohermes-3B-GGUF