From 8a917e930081427fe3fcbfe63d58e9f10eb078cb Mon Sep 17 00:00:00 2001 From: Rohan Joshi Date: Fri, 20 Jun 2025 08:41:11 -0700 Subject: [PATCH] Fix typo for 16a4w_block quantization (#11821) Summary: Pull Request resolved: https://github.com/pytorch/executorch/pull/11821 Differential Revision: D76949897 --- examples/qualcomm/oss_scripts/llama/llama.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/examples/qualcomm/oss_scripts/llama/llama.py b/examples/qualcomm/oss_scripts/llama/llama.py index 33482090b28..99f346eccbc 100755 --- a/examples/qualcomm/oss_scripts/llama/llama.py +++ b/examples/qualcomm/oss_scripts/llama/llama.py @@ -398,7 +398,7 @@ def quantize(self, quant_dtype, args, tokenizer, custom_annotations=()): self.llama_graph_module, self.inputs, strict=True ).module() - if QuantDtype == QuantDtype.use_16a4w_block: + if quant_dtype == QuantDtype.use_16a4w_block: conv_nodes = [ n for n in fx_graph_module.graph.nodes if "conv" in n.name ]