Gemma-Merge-Test-7b / README.md
rombodawg's picture
Update README.md
54eea28 verified
|
raw
history blame
623 Bytes
---
license: apache-2.0
---
This is a test merge of some Gemma-7b finetunes using task_arithmatic. So far its not working, but this is uploaded for testing and diagnosing purposes.
Merge config:
```yaml
models:
- model: gemma-7b-it-fp16
parameters:
weight: 1
- model: CorticalStack_gemma-7b-ultrachat-sft
parameters:
weight: 1
- model: cloudyu_google-gemma-7b-it-dpo-v1
parameters:
weight: 1
- model: abideen_gemma-7b-openhermes
parameters:
weight: 1
merge_method: task_arithmetic
base_model: gemma-7b-base
parameters:
normalize: true
int8_mask: true
dtype: float16
```