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 16, 23:15 UTC
[CI][BugFix] Flip is_quant_method_supported condition (#5577)
Michael Goin committed
Jun 16, 2024 at 10:07 UTC
4a6769053ab2616f7f490e6ec5b8241e76ef0c2a
1 file changed
+1
-1
tests/quantization/utils.py
+1
-1
@@ -10,5 +10,5 @@ def is_quant_method_supported(quant_method: str) -> bool:
10
11
capability = torch.cuda.get_device_capability()
12
capability = capability[0] * 10 + capability[1]
13
- return (capability <
13
+ return (capability >=
14
QUANTIZATION_METHODS[quant_method].get_min_capability())