Qianfeng Zhang
|
3f9f2fa736
|
Remove max_target 3200 cases from cross_attention testing and benchmarking
|
2026-04-17 09:17:38 +00:00 |
|
Qianfeng Zhang
|
5c84f54fd9
|
Add scripts for testing/benchmarking cross_attention cases
|
2026-04-16 15:45:57 +00:00 |
|
Qianfeng Zhang
|
302537c5a8
|
Update to support grouped mode hstu attention
|
2026-03-09 16:15:58 +00:00 |
|
Qianfeng Zhang
|
f2a555dac7
|
Align the masking logic in HstuCrossAttentionBlockMask with pytorch mask_v2 scripts
|
2026-02-09 15:55:13 +00:00 |
|
Qianfeng Zhang
|
8af5e26717
|
Add softmax selection to two of the testing scripts
|
2026-02-05 15:27:15 +00:00 |
|
Qianfeng Zhang
|
c360e0cbc4
|
Add scripts for benchmark sparsity 0.9 cases with mattn256 & full256
|
2026-01-30 10:02:31 +00:00 |
|
Qianfeng Zhang
|
749e83f2fd
|
Update to use BottomRight-Diagonal masking when seqlen_kv is bigger than seqlen_q
|
2026-01-26 13:45:42 +00:00 |
|
Qianfeng Zhang
|
f79a29ac80
|
Rename and add scripts for testing hdim96
|
2025-12-12 16:16:43 +00:00 |
|
Qianfeng Zhang
|
f01e0ef37d
|
Enable the using of WarpTile-32x32x16 and add scripts to verify
|
2025-11-30 04:58:28 +00:00 |
|
Qianfeng Zhang
|
8408ec0a02
|
Add scripts for testing the using of separate sequence lengths for k/v
|
2025-11-02 03:16:22 +00:00 |
|
Qianfeng Zhang
|
7c4012266a
|
Update to benchmark scripts to consider for using softmax
|
2025-10-23 10:09:37 +00:00 |
|
Qianfeng Zhang
|
d1505786f8
|
Add support of softmax in hstu attention
|
2025-10-20 14:26:55 +00:00 |
|
Qianfeng Zhang
|
fb09061b0c
|
Add norm_dist parameter for hstu example to select either normal or uniform distribution to initialize data
|
2025-08-12 03:06:35 +00:00 |
|
Qianfeng Zhang
|
1404336200
|
Update HstuBlockMaskWithLocal::GetTileRangeAlongX, add comments and test cases for causal == false
|
2025-08-10 06:10:14 +00:00 |
|
Qianfeng Zhang
|
971d0d98d4
|
Update to support min_full_attn_seqlen be bigger than max_uih_len
|
2025-08-08 09:26:55 +00:00 |
|
Qianfeng Zhang
|
f27d8cefb7
|
Add attn_scale MakeKargs() parameter support and update in example, reference codes
|
2025-08-03 03:37:28 +00:00 |
|
Qianfeng Zhang
|
3483af0516
|
Fix added case in test_hstu_attention.sh
|
2025-07-25 15:12:05 +00:00 |
|
Qianfeng Zhang
|
29d3dc9662
|
Update in GetTileRangeAlongX to consider for non-causal+local_size>0 situation and add test case to test_hstu_attention.sh
|
2025-07-25 14:56:13 +00:00 |
|
Qianfeng Zhang
|
43a97681b8
|
Add three scripts for verification of jagged causal cases
|
2025-07-25 11:20:46 +00:00 |
|
Qianfeng Zhang
|
c87a217475
|
Update to test_ck_hstu_mask.sh and test_pytorch_hstu_mask.py to align their testings
|
2025-06-22 16:26:53 +00:00 |
|
Qianfeng Zhang
|
09ac14604c
|
Align the -seqlens=xxx in the mattn0_full0 and mattn256_full256 scripts with the required benchmarks
|
2025-06-18 16:02:04 +00:00 |
|
Qianfeng Zhang
|
9e6a24010a
|
Move all test and bench scripts to folder scripts
|
2025-06-06 08:22:38 +00:00 |
|