flashinfer-ai/flashinfer-trace
Updated • 5.49k • 15
Zihao Ye
zhye
AI & ML interests
None yet
Recent Activity
Organizations
Vectorize dsa_sparse_attention_h16_ckv512_kpe64_topk2048_ps64 reference
1
#258 opened 3 days ago
by 
ubospica
feat: add gqa_paged_prefill_causal_h24_kv4_d128_ps64 (Mixtral 8x22B TP=2 prefill)
1
#196 opened 14 days ago
by 
averyyh
feat: add gqa_paged_decode_h24_kv4_d128_ps64 workloads (Mixtral 8x22B TP=2)
1
#195 opened 14 days ago
by averyyh
feat: add gqa_paged_prefill_causal_h24_kv4_d128_ps1 workloads, solution, and definition
1
#199 opened 14 days ago
by averyyh
Add gqa_paged_decode_h48_kv8_d128_ps1: solution + workloads + def + tests
#162 opened 15 days ago
by averyyh
workloads: add gqa_paged_prefill_causal_h16_kv1_d128_ps64 (Qwen3-235B-A22B, TP=4)
1
#152 opened 15 days ago
by averyyh
fix: mark missing fp8/scale tensors as random in MoE workload
#228 opened 7 days ago
by averyyh
test claude auto workload collection skill: fuse_add_rms_norm_h5120
1
#28 opened 20 days ago
by averyyh
gdn workload update: all tensors are dumped; add flashinfer gdn baseline
8
#21 opened about 1 month ago
by averyyh
fix: gdn workload path and shape
1
#19 opened about 1 month ago
by averyyh
fix: error when copying definition
#18 opened about 1 month ago
by averyyh
fix: missing mtp tp2 workload
#13 opened about 1 month ago
by averyyh
fix: real k for gdn prefill (tp4)
1
#14 opened about 1 month ago
by averyyh
fix: real k for gdn prefill (tp2)
1
#15 opened about 1 month ago
by averyyh
add gdn tp2 mtp workload
#12 opened about 2 months ago
by averyyh
add gdn tp4 prefill/decode workload
#11 opened about 2 months ago
by averyyh
update gdn tp4 decode: more batch size
#10 opened about 2 months ago
by averyyh
add gdn tp4 mtp workload
#8 opened 2 months ago
by averyyh
updated a
dataset 3 months ago
liked a
dataset 3 months ago