Skip to content

Actions: mudler/vllm.cpp

Actions

containers

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
containers.yml
866 workflow runs
866 workflow runs

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
fix(ROCM): AttnQkNormRopeGate output-dtype dispatch -- the 0.8B GDN forward divergence root-caused and fixed (RED anchor turns GREEN) containers #873: Pull request #570 synchronize by VikashLoomba
Queued VikashLoomba:row/ROCM-GDN-08B-FIX
VikashLoomba:row/ROCM-GDN-08B-FIX
Queued
gate(ROCM): M4 -- pinned vLLM-ROCm oracle on gfx1100 + ROCm device-golden lane in the Qwen3-0.6B near-tie gate containers #872: Pull request #549 synchronize by VikashLoomba
Queued VikashLoomba:row/ROCM-M4-GATE
VikashLoomba:row/ROCM-M4-GATE
Queued
feat(rocm): grouped + non-grouped keep-quant expert GEMM — the MoE-path terminus (issue #41) containers #871: Pull request #523 synchronize by VikashLoomba
Queued VikashLoomba:row/ROCM-GG-ON-MOE
VikashLoomba:row/ROCM-GG-ON-MOE
Queued
feat(rocm): MoE combine/gate ops — SharedExpertGate, MoeCombine, MoeCombineGate (issue #41) containers #870: Pull request #509 synchronize by VikashLoomba
Queued VikashLoomba:row/ROCM-MOE-CHAIN
VikashLoomba:row/ROCM-MOE-CHAIN
Queued
test(rocm): ReshapeAndCache->PagedAttention composition at real dims (issue #41) containers #869: Pull request #497 synchronize by VikashLoomba
Queued VikashLoomba:row/ROCM-ATTN-COMPOSE-TEST
VikashLoomba:row/ROCM-ATTN-COMPOSE-TEST
Queued
feat(MODEL-FP8-BLOCK-WEIGHT): a block-wise FP8 checkpoint now LOADS, and says by name that it cannot run yet (#1189) containers #868: Pull request #1228 opened by localai-bot
Queued row/MODEL-FP8-BLOCK-WEIGHT
row/MODEL-FP8-BLOCK-WEIGHT
Queued
record(ENG-CUDAGRAPH-DEDUP): the device A/B ran, and the fold it was filed for NEVER HAPPENS (#1162, #1184, #1226) containers #867: Pull request #1227 opened by localai-bot
Queued row/ENG-CUDAGRAPH-DEDUP-RESULT
row/ENG-CUDAGRAPH-DEDUP-RESULT
Queued
fix(GATE-SYMBOL-ANCHORS): cite the symbol, and gate that it is still there (#1143, #1139) containers #866: Pull request #1203 synchronize by localai-bot
Queued fix/citation-anchor-freshness
fix/citation-anchor-freshness
Queued
feat(LTX25-KEYFRAME-INTERP): the interpolation pipeline, on the conditioning builder that has no frame-0 branch (#1096) containers #865: Pull request #1209 synchronize by localai-bot
Queued row/LTX25-KEYFRAME-INTERP
row/LTX25-KEYFRAME-INTERP
Queued
feat(VT-MATMUL-FP8-BLOCK-REF): add vt::MatmulFp8BlockScaled, the block-scaled fp8 GEMM whose scales cannot live in an epilogue containers #864: Pull request #1224 opened by localai-bot
52m 41s row/VT-MATMUL-FP8-BLOCK-REF
row/VT-MATMUL-FP8-BLOCK-REF
52m 41s
feat(ENG-CUDAGRAPH-DEDUP): fold captures that share a topology onto one graph executable containers #863: Pull request #1178 synchronize by localai-bot
6m 29s row/ENG-CUDAGRAPH-DEDUP
row/ENG-CUDAGRAPH-DEDUP
6m 29s
feat(ENG-CUDAGRAPH-BREAK): a forward can be captured as SEGMENTS split at break points, and CUDA permits the re-begin W0 could not test (#1192, #1163) containers #862: Pull request #1207 synchronize by localai-bot
11m 18s row/ENG-CUDAGRAPH-BREAK-W1
row/ENG-CUDAGRAPH-BREAK-W1
11m 18s
perf(GDN-MOE-BF16-OUT): the GDN recurrence output and the z gate are bf16 on MoE too, as both references keep them containers #861: Pull request #1223 opened by localai-bot
25m 36s row/GDN-MOE-BF16-OUT-IMPL
row/GDN-MOE-BF16-OUT-IMPL
25m 36s
feat(ENG-CUDAGRAPH-BREAK): a forward can be captured as SEGMENTS split at break points, and CUDA permits the re-begin W0 could not test (#1192, #1163) containers #860: Pull request #1207 synchronize by localai-bot
16m 7s row/ENG-CUDAGRAPH-BREAK-W1
row/ENG-CUDAGRAPH-BREAK-W1
16m 7s
fix(#1157): NemotronH decoded with STALE input ids, and the recurrent carry was never the cause containers #859: Pull request #1221 synchronize by localai-bot
Queued row/MODEL-NEMOTRON-H-ABI-A2P-1157-fix
row/MODEL-NEMOTRON-H-ABI-A2P-1157-fix
Queued
fix(#1157): NemotronH decoded with STALE input ids, and the recurrent carry was never the cause containers #858: Pull request #1221 opened by localai-bot
11m 20s row/MODEL-NEMOTRON-H-ABI-A2P-1157-fix
row/MODEL-NEMOTRON-H-ABI-A2P-1157-fix
11m 20s
feat(ENG-CUDAGRAPH-BREAK): a forward can be captured as SEGMENTS split at break points, and CUDA permits the re-begin W0 could not test (#1192, #1163) containers #857: Pull request #1207 synchronize by localai-bot
59m 15s row/ENG-CUDAGRAPH-BREAK-W1
row/ENG-CUDAGRAPH-BREAK-W1
59m 15s
fix(GATE-SYMBOL-ANCHORS): cite the symbol, and gate that it is still there (#1143, #1139) containers #856: Pull request #1203 synchronize by localai-bot
2h 35m 10s fix/citation-anchor-freshness
fix/citation-anchor-freshness
2h 35m 10s
feat(ENG-RESIDENCY-CONFIG): the device budget is a config key, and --offload-config reaches vllm-cli and the pooling server (#1127, #1135) containers #855: Pull request #1216 opened by localai-bot
Queued row/ENG-RESIDENCY-CONFIG-W2
row/ENG-RESIDENCY-CONFIG-W2
Queued
record(ENV-LEASE-RUNTIME-STAGING): the leased worker compiles, downloads and runs Python, and AGENTS.md said it could do none of the three containers #854: Pull request #1215 opened by localai-bot
52m 2s row/RECORDS-LEASE-RUNTIME-TRUTH
row/RECORDS-LEASE-RUNTIME-TRUTH
52m 2s
feat(VT-QUANT-FP8-GROUP): add vt::QuantFp8Group, the dynamic per-token per-group fp8 activation quant, CPU and CUDA (#1189) containers #853: Pull request #1212 synchronize by localai-bot
56m 40s row/VT-QUANT-FP8-GROUP
row/VT-QUANT-FP8-GROUP
56m 40s
record(LTX25-DECODE-SPEED): three scalar single-threaded loops in the LTX CPU path, and the number ltx2_video.cpp said was owed containers #852: Pull request #1214 opened by localai-bot
Queued record/ltx25-scalar-cpu-path
record/ltx25-scalar-cpu-path
Queued
feat(VT-QUANT-FP8-GROUP): add vt::QuantFp8Group, the dynamic per-token per-group fp8 activation quant, CPU and CUDA (#1189) containers #851: Pull request #1212 opened by localai-bot
22m 41s row/VT-QUANT-FP8-GROUP
row/VT-QUANT-FP8-GROUP
22m 41s
fix(ENG-EXPERT-STREAM): the CUDA blocker is a missing vt::MatmulBTAlphaBeta, not a slower third path, and the arm now refuses by name (#1126, #1205, #1197) containers #850: Pull request #1200 synchronize by localai-bot
Queued row/ENG-EXPERT-STREAM-1126
row/ENG-EXPERT-STREAM-1126
Queued
docs: pin and reproduce Qwen3.8-2.4T-A95B UD-Q1_0 on one GB10 (#1194) containers #849: Pull request #1211 opened by localai-bot
Queued docs/qwen38-spark-recipe
docs/qwen38-spark-recipe
Queued