DeepSeek/DeepEP
1
0
Fork 0
mirror of https://github.com/deepseek-ai/DeepEP synced 2025-06-26 18:28:11 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Remove useless control metadata for low-latency combine

Browse Source
This commit is contained in:
Chenggang Zhao
2025-04-07 09:55:39 +08:00
parent 2a0b3d7a5d
commit 42494864ba
3 changed files with 8 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
tests/test_low_latency.py
View File

@@ -84,7 +84,7 @@ def test_main(num_tokens: int, hidden: int, num_experts: int, num_topk: int,
if do_check:
diff = calc_diff(x * topk_weights.masked_fill(topk_idx == -1, 0).sum(dim=1).view(-1, 1), combined_x)
assert torch.isnan(combined_x).sum().item() == 0
assert diff < 1e-5, f'Error: diff={diff}'
assert diff < 1e-5, f'Error: {diff=}, {zero_copy=}'
hash_value ^= hash_tensor(combined_x)
def create_test_cast_with_outliers(num_outliers):