Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
datasets:
|
| 4 |
+
- teknium/openhermes
|
| 5 |
+
- argilla/ultrafeedback-binarized-preferences
|
| 6 |
+
- Intel/orca_dpo_pairs
|
| 7 |
+
language:
|
| 8 |
+
- en
|
| 9 |
+
library_name: transformers
|
| 10 |
+
---
|
| 11 |
+
# DPOpenHermes 11B
|
| 12 |
+
|
| 13 |
+
This is a mergekit merge of DPOpenHermes-7B from seperate versions of it.
|
| 14 |
+
|
| 15 |
+
```
|
| 16 |
+
slices:
|
| 17 |
+
- sources:
|
| 18 |
+
- model: openaccess-ai-collective/DPOpenHermes-7B
|
| 19 |
+
revision: dpo-v0
|
| 20 |
+
layer_range: [0, 24]
|
| 21 |
+
- sources:
|
| 22 |
+
- model: openaccess-ai-collective/DPOpenHermes-7B
|
| 23 |
+
layer_range: [8, 32]
|
| 24 |
+
merge_method: passthrough
|
| 25 |
+
dtype: bfloat16
|
| 26 |
+
```
|