A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 9, 10:00 UTC
[UX] Remove noisy CT UnquantizedLinearMethod warn (#33273)
Signed-off-by: mgoin <mgoin64@gmail.com>
Michael Goin committed
Jan 28, 2026 at 19:09 UTC
141cd43967d20b2206353975ad0ebcce4d7d7c69
1 file changed
+1
-5
vllm/model_executor/layers/quantization/compressed_tensors/compressed_tensors.py
+1
-5
@@ -751,11 +751,7 @@ class CompressedTensorsConfig(QuantizationConfig):
751
model_compression_config=model_compression_config,
752
)
753
elif weight_quant is None:
754
- logger.warning_once(
755
- "Acceleration for non-quantized schemes is "
756
- "not supported by Compressed Tensors. "
757
- "Falling back to UnquantizedLinearMethod"
758
- )
754
+ # Falling back to UnquantizedLinearMethod
755
return None
756
757
else: