Commit
•
a07ecdf
1
Parent(s):
d930560
"Update README.md"
Browse files
README.md
CHANGED
|
@@ -25,8 +25,18 @@ I'm constantly enhancing these model descriptions to provide you with the most r
|
|
| 25 |
- Model creator: [stabilityai](https://huggingface.co/stabilityai)
|
| 26 |
- Original model: [stablelm-3b-4e1t](https://huggingface.co/stabilityai/stablelm-3b-4e1t)
|
| 27 |
|
|
|
|
| 28 |
|
| 29 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 30 |
|
| 31 |
|
| 32 |
# About GGUF format
|
|
|
|
| 25 |
- Model creator: [stabilityai](https://huggingface.co/stabilityai)
|
| 26 |
- Original model: [stablelm-3b-4e1t](https://huggingface.co/stabilityai/stablelm-3b-4e1t)
|
| 27 |
|
| 28 |
+
StableLM is a familiy of Models by Stability AI
|
| 29 |
|
| 30 |
|
| 31 |
+
---
|
| 32 |
+
# Brief
|
| 33 |
+
## Note:
|
| 34 |
+
Current (as of. 2023-11-15) implementations of Llama.cpp only support GPU offloading up to 34 Layers.
|
| 35 |
+
The model will crash immediately if -ngl is larger than 34.
|
| 36 |
+
The model works fine however without any gpu acceleration.
|
| 37 |
+
|
| 38 |
+
---
|
| 39 |
+
|
| 40 |
|
| 41 |
|
| 42 |
# About GGUF format
|