Skip to content

GitLab

Commit 96d4afdc authored by zhuwenwen's avatar zhuwenwen
Browse files

update moe configs name

parent 2b90ecd3
Show whitespace changes
Inline Side-by-side
Showing with 3 additions and 3 deletions
vllm/utils/__init__.py
View file @ 96d4afdc
...@@ -84,9 +84,9 @@ DEFAULT_MAX_NUM_BATCHED_TOKENS = 2048 ...@@ -84,9 +84,9 @@ DEFAULT_MAX_NUM_BATCHED_TOKENS = 2048
POOLING_MODEL_MAX_NUM_BATCHED_TOKENS = 32768 POOLING_MODEL_MAX_NUM_BATCHED_TOKENS = 32768
MULTIMODAL_MODEL_MAX_NUM_BATCHED_TOKENS = 5120 MULTIMODAL_MODEL_MAX_NUM_BATCHED_TOKENS = 5120
gpuname = torch.cuda.get_device_properties(torch.cuda.current_device()).name GPU_ARCH = torch.cuda.get_device_properties("cuda").gcnArchName
is_kme = gpuname.startswith('K100_AI') or gpuname.startswith('K500SM_AI') is_kme = any(arch in GPU_ARCH for arch in ["gfx928"])
SUPPORT_TC = gpuname.startswith('K100_AI') or gpuname.startswith('K500SM_AI') or gpuname.startswith('BW') SUPPORT_TC = any(arch in GPU_ARCH for arch in ["gfx928", "gfx936"])
def _generate_random_int8( def _generate_random_int8(
tensor: torch.Tensor, tensor: torch.Tensor,
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment