Commit
•
33209f9
1
Parent(s):
f37d8cb
Update README.md
Browse files
README.md
CHANGED
|
@@ -179,6 +179,8 @@ Probably coming soon, and I'll update as they do. I would upload them myself but
|
|
| 179 |
### Merge Method
|
| 180 |
|
| 181 |
A whole lot of SLERPing. I've found that SLERP and Model Stock are the only two reliable methods for merging Llama 3. The other methods seem to frazzle the weights leading to broken blends.
|
|
|
|
|
|
|
| 182 |
|
| 183 |
### Models Merged
|
| 184 |
|
|
|
|
| 179 |
### Merge Method
|
| 180 |
|
| 181 |
A whole lot of SLERPing. I've found that SLERP and Model Stock are the only two reliable methods for merging Llama 3. The other methods seem to frazzle the weights leading to broken blends.
|
| 182 |
+
This model was not the result of a thought-out process. I just kept mixing this and that until I got something that felt like the right balance of various factors.
|
| 183 |
+
The merge recipe for this model is ugly and I'm almost ashamed to show it, but it is what it is.
|
| 184 |
|
| 185 |
### Models Merged
|
| 186 |
|