Skip to content

Commit 3907d55

Browse files
fix eval_llama_qnn custom annotation
1 parent 350ea3c commit 3907d55

File tree

1 file changed

+3
-1
lines changed

1 file changed

+3
-1
lines changed

examples/qualcomm/oss_scripts/llama/eval_llama_qnn.py

Lines changed: 3 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -316,7 +316,9 @@ def eval_llm(args):
316316
if args.ptq is not None:
317317
quant_dtype = getattr(QuantDtype, f"use_{args.ptq}")
318318
decoder_model_config = SUPPORTED_LLM_MODELS[args.decoder_model]
319-
custom_annotations = decoder_model_config.custom_annotation
319+
custom_annotations = (
320+
decoder_model_config.quant_recipe().recipe.custom_quant_annotations
321+
)
320322

321323
quantizer = make_custom_quantizer(
322324
quant_dtype, args.range_setting, custom_annotations, args.quant_linear_only

0 commit comments

Comments
 (0)