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 14, 05:30 UTC
[rocm] enable torchao quantization for rocm (#24400)
Signed-off-by: Lifan Shen <lifans@meta.com>
Lifans committed
Sep 10, 2025 at 06:16 UTC
d6069887c6868f5f9683b4eb3e85a123d5124e53
1 file changed
+1
-1
vllm/platforms/rocm.py
+1
-1
@@ -171,7 +171,7 @@ class RocmPlatform(Platform):
171
172
supported_quantization: list[str] = [
173
"awq", "gptq", "fp8", "compressed-tensors", "fbgemm_fp8", "gguf",
174
- "quark", "ptpc_fp8", "mxfp4", "petit_nvfp4"
174
+ "quark", "ptpc_fp8", "mxfp4", "petit_nvfp4", "torchao"
175
]
176
177
@classmethod