Commit
·
a43dc52
1
Parent(s):
86dcf67
Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: cc-by-nc-4.0
|
| 3 |
+
---
|
| 4 |
+
|
| 5 |
+
A 11b Mistral model, based on the NeverSleep recipe. Not meant for direct deployment, but who knows?
|
| 6 |
+
|
| 7 |
+
### Recipe
|
| 8 |
+
|
| 9 |
+
slices
|
| 10 |
+
|
| 11 |
+
- sources:
|
| 12 |
+
-
|
| 13 |
+
- model:
|
| 14 |
+
- model: maywell/PiVoT-0.1-Starling-LM-RP
|
| 15 |
+
-
|
| 16 |
+
layer_range: [0, 24]
|
| 17 |
+
|
| 18 |
+
- sources:
|
| 19 |
+
-
|
| 20 |
+
- model: maywell/PiVoT-0.1-Starling-LM-RP
|
| 21 |
+
-
|
| 22 |
+
layer_range: [8, 32]
|
| 23 |
+
|
| 24 |
+
merge_method: passthrough
|
| 25 |
+
|
| 26 |
+
dtype: bfloat16
|