README.md

metadata

language:
  - da
  - sv
license: cc-by-4.0
library_name: transformers
tags:
  - merge
  - mergekit
base_model:
  - danish-foundation-models/munin-7b-alpha
  - timpal0l/Mistral-7B-v0.1-flashback-v2

Danish-Swedish Merged Model

This is a merge of the following models, all based on mistralai/Mistral-7B-v0.1:

1. danish-foundation-models/munin-7b-alpha, continued pretraining on Danish data;
2. timpal0l/Mistral-7B-v0.1-flashback-v2, continued pretraining on Swedish data.

Model Details

• Merged by: Dan Saattrup Nielsen
• Model type: Decoder model, based on mistralai/Mistral-7B-v0.1
• Language(s): Danish and Swedish
• License: CC-BY-4.0
• Merge configuration:

  dict(
      models=[
          dict(
              model="danish-foundation-models/munin-7b-alpha",
              parameters=dict(
                  weight=1.0,
                  density=0.6,
              ),
          ),
          dict(
              model="timpal0l/Mistral-7B-v0.1-flashback-v2",
              parameters=dict(
                  weight=1.0,
                  density=0.6,
              ),
          ),
      ],
      merge_method="dare_ties",
      random_seed=4242
      base_model="mistralai/Mistral-7B-v0.1",
      parameters=dict(
          int8_mask=True,
          normalize=True,
      ),
      dtype="bfloat16",
  )