biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Bugfix] Fix some narrowing conversion warnings (#20141)

Browse Source 
Signed-off-by: Tyler Michael Smith <tyler@neuralmagic.com>
This commit is contained in:
Tyler Michael Smith
2025-06-27 12:01:28 -04:00
committed by GitHub
parent c6c983053d
commit e8c3bd2cd1
6 changed files with 8 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
csrc/quantization/fp4/nvfp4_experts_quant.cu
View File

@@ -561,7 +561,7 @@ void scaled_fp4_experts_quant_sm100a(
TORCH_CHECK(output_scale.size(1) * 4 == padded_k);
auto in_dtype = input.dtype();
at::cuda::CUDAGuard device_guard{(char)input.get_device()};
const at::cuda::OptionalCUDAGuard device_guard(device_of(input));
const cudaStream_t stream =
at::cuda::getCurrentCUDAStream(input.get_device());
if (in_dtype == at::ScalarType::Half) {
@@ -579,4 +579,4 @@ void scaled_fp4_experts_quant_sm100a(
} else {
TORCH_CHECK(false, "Expected input data type to be half or bfloat16");
}
}
}