rixty/KoboldAI-Client
1
1
Fork 0
mirror of https://github.com/KoboldAI/KoboldAI-Client.git synced 2025-06-05 21:59:24 +02:00
Code Issues Projects Releases Wiki Activity

Update to GPTQ module 0.0.2, add support for upstream cuda quantizations, automatic detection

Browse Source
This commit is contained in:
0cc4m
2023-05-09 22:19:18 +02:00
parent 6121598142
commit a2d01bb9e4
6 changed files with 46 additions and 29 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
modeling/inference_models/hf.py
View File

@@ -181,7 +181,8 @@ class HFInferenceModel(InferenceModel):
if "gptq_bits" in dir(self.model_config):
utils.koboldai_vars.gptq_model = True
utils.koboldai_vars.gptq_bits = self.model_config.gptq_bits
utils.koboldai_vars.gptq_groupsize = self.model_config.gptq_groupsize
utils.koboldai_vars.gptq_groupsize = self.model_config.gptq_groupsize if getattr(self.model_config, "gptq_groupsize", False) else -1
utils.koboldai_vars.gptq_version = self.model_config.gptq_version if getattr(self.model_config, "gptq_version", False) else 1
utils.koboldai_vars.gptq_file = None
else:
utils.koboldai_vars.gptq_model = False