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

Update Flashinfer from v0.4.1 to v0.5.2 (#27952)

Browse Source 
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
This commit is contained in:
Harry Mellor
2025-11-07 16:24:42 -08:00
committed by GitHub
parent 67a2da890e
commit 811df41ee9
4 changed files with 11 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
tests/kernels/attention/test_flashinfer_trtllm_attention.py
View File

@@ -238,9 +238,11 @@ def test_flashinfer_trtllm_decode_with_baseline(
if q_quant_dtype == FP8_DTYPE and o_quant_dtype == FP4_DTYPE:
rtol, atol = 7e-2, 9e-2
elif q_quant_dtype == FP8_DTYPE and o_quant_dtype == FP8_DTYPE:
rtol, atol = 2e-2, 4e-2
rtol, atol = 3e-2, 4e-2
elif q_quant_dtype == FP8_DTYPE and o_quant_dtype == dtype:
rtol, atol = 1e-2, 2e-2
rtol, atol = 2e-2, 2e-2
elif kv_quant_dtype == FP8_DTYPE:
rtol, atol = 4e-2, 6e-2
else:
rtol, atol = 1e-2, 1e-2