grimjim_-_Mistral-Starling-merge-trial1-7B-8bits

Quantization made by Richard Erkhov.

Github

Discord

Request more models

Mistral-Starling-merge-trial1-7B - bnb 8bits

• Model creator: https://huggingface.co/grimjim/
• Original model: https://huggingface.co/grimjim/Mistral-Starling-merge-trial1-7B/

Original model description:

base_model:

• Nexusflow/Starling-LM-7B-beta
• grimjim/Mistral-7B-Instruct-demi-merge-v0.2-7B library_name: transformers tags:
• mergekit
• merge license: apache-2.0

---

Mistral-Starling-merge-trial1-7B

This is a merge of pre-trained language models created using mergekit. The goal was to combine strong reasoning with 32K context length.

Merge Details

Merge Method

This model was merged using the SLERP merge method.

Models Merged

The following models were included in the merge:

• Nexusflow/Starling-LM-7B-beta
• grimjim/Mistral-7B-Instruct-demi-merge-v0.2-7B

Configuration

The following YAML configuration was used to produce this model:

slices:
  - sources:
      - model: grimjim/Mistral-7B-Instruct-demi-merge-v0.2-7B
        layer_range: [0, 32]
      - model: Nexusflow/Starling-LM-7B-beta
        layer_range: [0, 32]
# or, the equivalent models: syntax:
# models:
merge_method: slerp
base_model: grimjim/Mistral-7B-Instruct-demi-merge-v0.2-7B
parameters:
  t:
    - filter: self_attn
      value: [0, 0.5, 0.3, 0.7, 1]
    - filter: mlp
      value: [1, 0.5, 0.7, 0.3, 0]
    - value: 0.5 # fallback for rest of tensors
dtype: bfloat16