CoderCowMoo commited on
Commit
cdcdc4e
Β·
verified Β·
1 Parent(s): f084fbc

use_exllama=false?

Browse files
Files changed (1) hide show
  1. app.py +2 -1
app.py CHANGED
@@ -16,7 +16,8 @@ tokenizer = AutoTokenizer.from_pretrained(model_id)
16
  quantize_config = BaseQuantizeConfig(
17
  bits=4,
18
  group_size=128,
19
- desc_act=False
 
20
  )
21
  model = AutoGPTQForCausalLM.from_quantized(
22
  model_id,
 
16
  quantize_config = BaseQuantizeConfig(
17
  bits=4,
18
  group_size=128,
19
+ desc_act=False,
20
+ use_exllama=False
21
  )
22
  model = AutoGPTQForCausalLM.from_quantized(
23
  model_id,