Note FP32 rather than FP16
Browse files
README.md
CHANGED
|
@@ -13,6 +13,8 @@ The model recipe was as follows:
|
|
| 13 |
- Linear merge of **Orca-2-13b** @0.8 and **cat-v1.0-13b** @0.2 = OrcaCat-13b (no plans to release)
|
| 14 |
- Gradient SLERP merge of **Noromaid-v0.1.1** @0.5 and **OrcaCat-13b** @0.5 = OrcaMaid-13b
|
| 15 |
|
|
|
|
|
|
|
| 16 |
The overall goal of this merge is to create a model that sounds uniquely human and natural, without sacrificing intelligence.
|
| 17 |
|
| 18 |
The prompt format is Alpaca. You can use the standard format as shown, but for best results, you should customize the system prompt to your specific needs.
|
|
|
|
| 13 |
- Linear merge of **Orca-2-13b** @0.8 and **cat-v1.0-13b** @0.2 = OrcaCat-13b (no plans to release)
|
| 14 |
- Gradient SLERP merge of **Noromaid-v0.1.1** @0.5 and **OrcaCat-13b** @0.5 = OrcaMaid-13b
|
| 15 |
|
| 16 |
+
Both merges were done in FP32 rather than FP16, due to Orca being realeased as FP32. I didn't want to risk losing any precision.
|
| 17 |
+
|
| 18 |
The overall goal of this merge is to create a model that sounds uniquely human and natural, without sacrificing intelligence.
|
| 19 |
|
| 20 |
The prompt format is Alpaca. You can use the standard format as shown, but for best results, you should customize the system prompt to your specific needs.
|