Skip to content

Actions: artalis-io/bitnet.c

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
190 workflow runs
190 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

GPU performance: Q4 matvec rewrite, compute COPY, flash attention, ze… CI #190: Commit a9bfc5a pushed by farkasmark
11m 1s main
main
11m 1s
Add benchmark suite, fix bench_kernels for current API CI #189: Commit f5653be pushed by farkasmark
10m 56s main
main
10m 56s
Enable TQ on draft model, update README + CLAUDE.md docs CI #188: Commit f303052 pushed by farkasmark
12m 30s main
main
12m 30s
Enable TQ on draft model, update CLAUDE.md for prompt cache + TQ docs CI #187: Commit 222e973 pushed by farkasmark
12m 19s main
main
12m 19s
Fix TQ cache reset, add full-turn prompt caching, verify WASM TQ path CI #186: Commit 1fbef8e pushed by farkasmark
12m 21s main
main
12m 21s
TQ-aware prompt cache + audit fixes: bn_alloc, size_t casts, warning … CI #185: Commit 6790097 pushed by farkasmark
12m 18s main
main
12m 18s
Q4_0 AVX2: inline dot_i8_float, eliminate redundant integer accumulator CI #184: Commit b44bb05 pushed by farkasmark
11m 26s main
main
11m 26s
reindex CI #183: Commit 1fb8d56 pushed by farkasmark
11m 27s main
main
11m 27s
SSM delta: Kahan compensated summation for S@k and S^T@q CI #182: Commit ae701ee pushed by farkasmark
11m 27s main
main
11m 27s
reindex CI #181: Commit 10887dd pushed by farkasmark
11m 32s main
main
11m 32s
Q4_0 AVX2: inline dot_i8_float, eliminate redundant integer accumulator CI #180: Commit b6f774b pushed by farkasmark
10m 1s main
main
10m 1s
Q4_0 AVX2: inline dot_i8_float, eliminate redundant integer accumulator CI #179: Commit 9331d72 pushed by farkasmark
11m 25s main
main
11m 25s
SSM delta: Kahan compensated summation for S@k and S^T@q CI #178: Commit ae701ee pushed by farkasmark
11m 23s main
main
11m 23s
Q4_K GPU matvec: integer accumulation for NEON-matched precision CI #177: Commit e74832b pushed by farkasmark
11m 20s main
main
11m 20s
Fix all audit findings: C1-C2 critical, H1-H4 high, M1-M7 medium CI #176: Commit 77ef475 pushed by farkasmark
11m 22s main
main
11m 22s
Fix SSM GPU shaders: add per-layer state and conv_state offsets CI #175: Commit 36896cb pushed by farkasmark
11m 24s main
main
11m 24s
Q4_0 AVX2: float-domain accumulation eliminates per-block integer hsum CI #174: Commit dfebd1e pushed by farkasmark
11m 29s main
main
11m 29s
Revert Q-gated debug changes, keep CPU fallback for correctness CI #173: Commit d2ae7f8 pushed by farkasmark
11m 38s main
main
11m 38s
Add Q4_0 AVX2 tiled matmul kernel for batched prefill CI #172: Commit 84345ba pushed by farkasmark
11m 26s main
main
11m 26s
GPU Q-gated: keep rejection, clean up debug prints CI #171: Commit 3cc2d97 pushed by farkasmark
11m 25s main
main
11m 25s
Batch SSM projections in prefill for 37% faster time-to-first-token CI #170: Commit 55dd17e pushed by farkasmark
11m 35s main
main
11m 35s
Fix GPU Q-gated: skip stacked QKV for gated layers, reject until fixed CI #169: Commit e696ea9 pushed by farkasmark
9m 33s main
main
9m 33s
Fix GPU MoE coherence: reject Q-gated attention on GPU path CI #168: Commit 83d1dff pushed by farkasmark
10m 49s main
main
10m 49s
GPU MoE: batch expert submissions + mappedAtCreation + slab allocator CI #167: Commit 798e0d8 pushed by farkasmark
11m 1s main
main
11m 1s
Audit fixes: Q6K overflow guard, hoist prefill Q8K scratch CI #166: Commit ca8d029 pushed by farkasmark
10m 59s main
main
10m 59s