mirror of
https://github.com/KoboldAI/KoboldAI-Client.git
synced 2025-06-05 21:59:24 +02:00
4-bit as Default
This commit is contained in:
@@ -52,5 +52,3 @@ dependencies:
|
|||||||
- einops
|
- einops
|
||||||
- peft==0.3.0
|
- peft==0.3.0
|
||||||
- scipy
|
- scipy
|
||||||
- --find-links=https://0cc4m.github.io/exllama/exllama-whl-links.html
|
|
||||||
- exllama==0.0.6
|
|
||||||
|
@@ -61,7 +61,7 @@ class model_backend(HFTorchInferenceModel):
|
|||||||
"default": temp['quantization'] if 'quantization' in temp else 'none',
|
"default": temp['quantization'] if 'quantization' in temp else 'none',
|
||||||
"tooltip": "Whether or not to use BnB's 4-bit or 8-bit mode",
|
"tooltip": "Whether or not to use BnB's 4-bit or 8-bit mode",
|
||||||
"menu_path": "Layers",
|
"menu_path": "Layers",
|
||||||
"children": [{'text': 'None', 'value':'none'},{'text': '4-bit', 'value': '4bit'}, {'text': '8-bit', 'value': '8bit'}],
|
"children": [{'text': '4-bit', 'value': '4bit'}, {'text': '8-bit', 'value': '8bit'}, {'text': '16-bit', 'value':'16-bit'}],
|
||||||
"extra_classes": "",
|
"extra_classes": "",
|
||||||
"refresh_model_inputs": False
|
"refresh_model_inputs": False
|
||||||
})
|
})
|
||||||
|
Reference in New Issue
Block a user