Commit Graph
27 Commits
Author SHA1 Message Date
5a6a1bb883 [mxfp8-kv] Skip writes to the reserved CUDA-graph padding slot (#35351)
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: Ke Bao <ispobaoke@gmail.com>
Co-authored-by: Sam Shleifer <sam@thinkingmachines.ai>
2026-09-22 01:21:25 +08:00
8ae4a39b50 Gate mamba extra-buffer predicates on uses_mamba_radix_cache (#37474)
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
Co-authored-by: Ke Bao <ispobaoke@gmail.com>
Co-authored-by: Sam Shleifer <sam@thinkingmachines.ai>
2026-09-17 14:39:12 +08:00
76e06febab [CI] Add e2e test for dp-attention local control broadcast (#39437)
Co-authored-by: Sam Shleifer <sam@thinkingmachines.ai>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
2026-09-16 18:11:40 +08:00
6afb5e1771 Gate the idle-loop tree-cache sanity check behind a default-off env (#36205)
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
Co-authored-by: hzh0425 <hzh0425@apache.org>
2026-08-30 00:10:28 +08:00
Sam ShleiferandClaude Fable 5 a2e88279c2 metrics: don't clock-rebase unset time sentinels in ReqTimeStats deserialization (#34335)
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
2026-08-12 14:27:47 +08:00
Sam ShleiferandClaude Fable 5 afb4f37ca5 [Inkling] silu_and_mul: replace helion kernels with plain Triton (#33903)
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
2026-08-08 15:00:23 +08:00
Sam ShleiferandClaude Fable 5 d238e36b24 [Fix] Restore data_parallel_rank alias on native /generate (dp-aware gateway routing is silently dropped) (#33565)
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
2026-08-08 14:59:16 +08:00
Sam ShleiferandClaude Fable 5 62a28197c0 Autotune flashinfer extend buckets at warmup (#32556)
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
2026-08-07 14:06:05 -07:00
Sam Shleifer 8e7d361def [perf] Compute input logprobs without materializing the full-vocab log-softmax (#31958) 2026-08-07 14:01:08 -07:00
Sam Shleifer 58974ca16c [perf] Assemble flat prompt top logprobs scheduler-side as numpy arrays (#32223) 2026-07-31 18:00:51 -07:00
Sam ShleiferandClaude Fable 5 62dfaaa0e0 [Nemotron] Fix decode track-save reading the stale tail of the CUDA-graph track buffer (#32555)
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
2026-07-29 19:03:44 -07:00
Sam Shleifer d0e69d3881 [feat] Optional base64 encoding for the flat prompt top logprob arrays (#31960) 2026-07-29 12:15:56 -07:00
Sam Shleifer 5cc273a780 [feat] Opt-in flat response format for prompt top logprobs (#32078) 2026-07-26 23:44:34 -07:00
Sam Shleifer b8bb1b4e5a [lora] Fix WAR race: never write MoE runner output into hidden_states in place (#31870) 2026-07-24 00:58:55 -07:00
Sam Shleifer 1f9d778d1b Skip dist_init/nccl port prechecks when the dist init method is overridden (#31410) 2026-07-23 12:40:41 -07:00
Sam Shleifer b3a0185cab model_runner: extract post-memory-pool wiring into _init_post_memory (#31601) 2026-07-19 00:57:54 +08:00
Sam Shleifer d86ae51fcf metrics: allow extra labels on HTTP request/response Prometheus metrics (#31530) 2026-07-19 00:57:26 +08:00
Sam Shleifer 0675d3033f Fix mamba track-boundary seqlen under overlap scheduler (#31369) 2026-07-17 10:22:07 +08:00
Sam ShleiferandAlex Nails dec0836302 Fix processor config loading for object-storage model paths (#31211)
Co-authored-by: Alex Nails <alex.nails@radixark.ai>
2026-07-15 00:21:31 -07:00
Sam Shleifer 133254086b Plug mamba_extra_buffer ping-pong slot leaks (#26941) 2026-06-04 21:46:36 +08:00
Sam Shleifer ce2506e1c6 Deprecate record_nolora_graph dual MoE CUDA graph capture (#24314) 2026-05-15 19:20:58 -07:00
Sam Shleifer e6f252e9b8 Cache FlashInfer autotune configs (#24156) 2026-05-05 02:00:40 +08:00
Sam Shleifer 63f225ca2e [session] fix mamba pool leak in StreamingSession.release_session + plumb idle leak check (#23496) 2026-05-02 11:38:08 +08:00
d41e8c459d Support RunAI loading for quantized checkpoints (#23850)
Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Co-authored-by: Sam Shleifer <sam@thinkingmachines.ai>
2026-05-02 11:11:40 +08:00
Sam Shleifer b9e33d6a5b Dual MoE CUDA graph capture for lora/nolora batches (#22809) 2026-04-22 14:11:11 -07:00
Sam Shleifer 1100b9865c Fix MxInt4 MoE returning wrong output variable (#21348) 2026-03-26 10:57:09 +08:00
Sam Shleifer 0d49b13fdd Fix circular import in quantization modules (#17372) 2026-01-21 15:47:09 +08:00