RichardErkhov/Replete-AI_-_Phi-5B-Test-8bits

Quantization made by Richard Erkhov.

Github

Discord

Request more models

Phi-5B-Test - bnb 8bits

• Model creator: https://huggingface.co/Replete-AI/
• Original model: https://huggingface.co/Replete-AI/Phi-5B-Test/

Original model description:

base_model: [] tags: - mergekit - merge license: mit

Untitled Model (1)

This is a merge of pre-trained language models created using mergekit.

Merge Details

Merge Method

This model was merged using the passthrough merge method.

Models Merged

The following models were included in the merge:

• https://huggingface.co/liminerity/Phigments12

• https://huggingface.co/l3utterfly/phi-2-layla-v1-chatml

• and then depth upscaled

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: liminerity/Phigments12
    parameters:
      density: [1, 0.7, 0.1] # density gradient
      weight: 1.0
  - model: l3utterfly/phi-2-layla-v1-chatml
    parameters:
      density: 0.8
      weight: [0, 0.5, 0.7, 1] # weight gradient
 
merge_method: dare_ties
base_model: liminerity/Phigments12
parameters:
  normalize: true
  int8_mask: true
dtype: float16

dtype: float16
merge_method: passthrough
slices:
  - sources:
    - model: phi/
      layer_range: [0,32]
  - sources:
    - model: phi/
      layer_range: [0,32]
 

Join the Replete AI Discord here!