grenade
0 Followers · 0 Following
  • Joined on 2026-02-17
Repositories
6
 Projects Packages Public Activity Starred Repositories
grenade pushed to feat/98-batched-decode at helexa/helexa 2026-07-03 12:24:04 +00:00
d3ef9513f8 fix(neuron): as_deref for Arc'd prefix_cache in cuda-only worker path (#98)
grenade pushed to feat/98-batched-decode at helexa/helexa 2026-07-03 12:03:48 +00:00
ec7a5b750a feat(neuron): lockstep batched decode engine (#98)
grenade pushed to feat/98-batched-decode at helexa/helexa 2026-07-03 11:46:53 +00:00
7f2fcc3527 feat(neuron): batch-row extraction for rebatch on join/leave (#98)
grenade pushed to feat/98-batched-decode at helexa/helexa 2026-07-03 11:42:11 +00:00
9ac1a1586c feat(neuron): device-worker jobs for batched decode (#98)
grenade pushed to feat/98-batched-decode at helexa/helexa 2026-07-03 11:34:36 +00:00
05b6c70a3a feat(neuron): batched lockstep decode groundwork for qwen3_5 (#98)
grenade created branch feat/98-batched-decode in helexa/helexa 2026-07-03 11:34:36 +00:00
grenade deleted branch fix/95-flash-prefill-only from helexa/helexa 2026-07-03 09:29:55 +00:00
grenade pushed to main at helexa/helexa 2026-07-03 09:29:53 +00:00
9c037c5a0b Merge pull request 'perf(neuron): route only prefill through FlashAttention (#95)' (#120) from fix/95-flash-prefill-only into main
cb6ddbbd98 perf(neuron): route only prefill through FlashAttention (#95)
Compare 2 commits »
grenade closed issue helexa/helexa#95 2026-07-03 09:29:52 +00:00
F4a — perf(neuron): enable flash-attention on full-attn layers
grenade merged pull request helexa/helexa#120 2026-07-03 09:29:52 +00:00
perf(neuron): route only prefill through FlashAttention (#95)
grenade created pull request helexa/helexa#120 2026-07-03 09:29:41 +00:00
perf(neuron): route only prefill through FlashAttention (#95)
grenade pushed to fix/95-flash-prefill-only at helexa/helexa 2026-07-03 09:15:08 +00:00
cb6ddbbd98 perf(neuron): route only prefill through FlashAttention (#95)
grenade created branch fix/95-flash-prefill-only in helexa/helexa 2026-07-03 09:15:07 +00:00
grenade deleted branch feat/95-flash-attn from helexa/helexa 2026-07-03 08:47:28 +00:00
grenade pushed to main at helexa/helexa 2026-07-03 08:47:26 +00:00
e824ea3274 Merge pull request 'perf(neuron): FlashAttention on the qwen3_5 full-attention layers (#95)' (#119) from feat/95-flash-attn into main
9139b07d71 fix(neuron): candle-flash-attn must be a direct optional dependency (#95)
4762537225 perf(neuron): FlashAttention on the qwen3_5 full-attention layers (#95)
Compare 3 commits »
grenade merged pull request helexa/helexa#119 2026-07-03 08:47:25 +00:00
perf(neuron): FlashAttention on the qwen3_5 full-attention layers (#95)
grenade created pull request helexa/helexa#119 2026-07-03 08:47:14 +00:00
perf(neuron): FlashAttention on the qwen3_5 full-attention layers (#95)
grenade pushed to feat/95-flash-attn at helexa/helexa 2026-07-03 08:31:32 +00:00
9139b07d71 fix(neuron): candle-flash-attn must be a direct optional dependency (#95)
grenade pushed to feat/95-flash-attn at helexa/helexa 2026-07-03 08:14:51 +00:00
4762537225 perf(neuron): FlashAttention on the qwen3_5 full-attention layers (#95)
grenade created branch feat/95-flash-attn in helexa/helexa 2026-07-03 08:14:51 +00:00