![](https://cdn-avatars.huggingface.co/v1/production/uploads/6033e34a9aa44495c80dd043/FlLD2t0N7xRbIKcCKAD2n.png)
Merge Crew
AI & ML interests
Merging models
Organization Card
The merge crew is the mergiest crew. 😎
Merge Crew Planning Document
https://docs.google.com/document/d/1fP2FIrCifWcLGdTBmqeogdCdZJOwxqPfEyO-HA76_qc/edit?usp=sharing
Merging tutorial
https://huggingface.co/blog/mlabonne/merge-models
Colab for merging
Lazy merge kit notebook for merging models. https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing
Model Merging Discord
Discord channel for discussions of model mergin. https://discord.com/channels/905500133343518791/1202582325146161183
Merge methods
When working you use a YAML file like the following
models:
- model: timpal0l/BeagleCatMunin
# No parameters necessary for base model
- model: bineric/NorskGPT-Mistral-7b
parameters:
density: 0.53
weight: 0.6
merge_method: dare_ties
base_model: timpal0l/BeagleCatMunin
parameters:
int8_mask: true
dtype: bfloat16
random_seed: 42
The dare_ties methods seems to perform better than other merging methods. Learn more about merge ties here. https://arxiv.org/pdf/2306.01708.pdf
Collections
1
models
12
![](https://cdn-avatars.huggingface.co/v1/production/uploads/6033e34a9aa44495c80dd043/FlLD2t0N7xRbIKcCKAD2n.png)
merge-crew/da-sv-slerp
Text Generation
•
Updated
•
19
•
2
![](https://cdn-avatars.huggingface.co/v1/production/uploads/6033e34a9aa44495c80dd043/FlLD2t0N7xRbIKcCKAD2n.png)
merge-crew/da-sv-ties
Text Generation
•
Updated
•
16
![](https://cdn-avatars.huggingface.co/v1/production/uploads/6033e34a9aa44495c80dd043/FlLD2t0N7xRbIKcCKAD2n.png)
merge-crew/da-sv-dare-ties-density-0.6
Text Generation
•
Updated
•
17
![](https://cdn-avatars.huggingface.co/v1/production/uploads/6033e34a9aa44495c80dd043/FlLD2t0N7xRbIKcCKAD2n.png)
merge-crew/da-sv-dare-ties-density-0.3
Text Generation
•
Updated
•
17
![](https://cdn-avatars.huggingface.co/v1/production/uploads/6033e34a9aa44495c80dd043/FlLD2t0N7xRbIKcCKAD2n.png)
merge-crew/da-sv-dare-ties-density-0.9
Text Generation
•
Updated
•
17
![](https://cdn-avatars.huggingface.co/v1/production/uploads/6033e34a9aa44495c80dd043/FlLD2t0N7xRbIKcCKAD2n.png)
merge-crew/da-sv-task-arithmetic
Text Generation
•
Updated
•
18
![](https://cdn-avatars.huggingface.co/v1/production/uploads/6033e34a9aa44495c80dd043/FlLD2t0N7xRbIKcCKAD2n.png)
merge-crew/MOE-SWE-DAN-NO-CODE
Text Generation
•
Updated
•
5
•
1
![](https://cdn-avatars.huggingface.co/v1/production/uploads/6033e34a9aa44495c80dd043/FlLD2t0N7xRbIKcCKAD2n.png)
merge-crew/munin-neuralbeagle-7b-density-very-low
Text Generation
•
Updated
•
6
![](https://cdn-avatars.huggingface.co/v1/production/uploads/6033e34a9aa44495c80dd043/FlLD2t0N7xRbIKcCKAD2n.png)
merge-crew/munin-neuralbeagle-7b-density-low
Text Generation
•
Updated
•
7
![](https://cdn-avatars.huggingface.co/v1/production/uploads/6033e34a9aa44495c80dd043/FlLD2t0N7xRbIKcCKAD2n.png)
merge-crew/munin-neuralbeagle-7b-density-high
Text Generation
•
Updated
•
5
datasets
None public yet