Add --no_use_cuda_fp16 param for AutoGPTQ

2023-06-23 12:22:56 -03:00 · 2023-06-23 12:22:56 -03:00 · 3ae9af01aa
commit 3ae9af01aa
parent 5646690769
5 changed files with 6 additions and 2 deletions
--- a/modules/loaders.py
+++ b/modules/loaders.py
@ -9,6 +9,7 @@ loaders_and_params = {
        'triton',
        'no_inject_fused_attention',
        'no_inject_fused_mlp',
+        'no_use_cuda_fp16',
        'wbits',
        'groupsize',
        'desc_act',