SolitaryThinker
57662a9ea5
[fix] Sync PAN2 padding parity with latest upstream ( #1961 )
2026-10-20 22:28:39 +00:00
SolitaryThinker
f80184af75
[misc] Merge main into PAN2 port ( #1961 )
2026-10-20 22:18:40 +00:00
376e445f0e
[feat]: FastH3 OmniRef inference on consumer devices (RTX 5090 / PRO 6000 / DGX Spark / 4090) ( #1980 )
...
Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com >
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com >
2026-10-19 14:57:36 -07:00
lpc0220 and root
d4ed61cc7c
[refactor] VSA: rename the sm100a block-sparse kernel to ptx_blackwell ( #1956 )
...
Co-authored-by: root <root@ptyche0124.ptyche.clusters.nvidia.com >
2026-10-17 21:10:19 -07:00
William Lin
9f59519959
[misc] Remove DreamX World, HunyuanGameCraft, GLM-Image, LongCat, TurboDiffusion, and Z-Image ( #1926 )
2026-10-14 23:16:32 -07:00
SolitaryThinker
26719dcfe7
[new-model] Add PAN2 text-to-video and image-to-video
...
Port of vllm-project/vllm-omni#8602 . PAN2 is a 29B mixture-of-transformers
video DiT (separate video and text streams joined in one attention per
block) with a Qwen3.5 text encoder and the Wan2.2 VAE. One pipeline serves
text-to-video and first-frame image-to-video.
- Native PAN2 DiT; per-stream Q/K/V fused at load time, Ulysses SP with
replicated text tokens through DistributedAttention.
- Native Qwen3.5 text encoder (Gated DeltaNet + gated attention) with a
PyTorch chunked delta rule.
- Reuses the Wan2.2 VAE and the flow-match Euler scheduler.
- PAN2 stages: chat-template prompt crop at hidden_states[-3], first-frame
conditioning, uniform sigmas with shift 7, fp32 latents. Adds a
DenoisingStage.prepare_scheduler_model_output hook for the last one.
- Registry entry for the Modular Diffusers PAN2ModularPipeline manifest,
pan2_t2v / pan2_i2v presets, example, docs, and local parity tests
against a diffusers/transformers reference on the public tiny checkpoint.
2026-10-15 03:44:35 +00:00
Davide Locatelli
e1e2559359
[perf] FastH3 OmniRef: lossless speed-ups behind default-off switches ( #1930 )
...
These speed-ups for the MiniMax-H3 pipelines keep the output bit-identical. Each
one has a switch that is off by default:
- FASTVIDEO_MINIMAX_H3_EXACT_KERNELS: Triton RoPE, AdaLN modulation, gated
residual and SwiGLU kernels that round to bf16 at the same points as the
eager ops.
- FASTVIDEO_H3_VSA_HEADS_FIRST_TILE: VSA-H3 tiles are written directly in the
heads-first layout of the kernel.
- FASTVIDEO_H3_VAE_TILE_PARALLEL: VAE spatial tiles are split across the
sequence-parallel ranks.
- FASTVIDEO_H3_REF2VA_MEMO_ENTRIES: repeat reference encodes are reused.
The example basic_fasth3_omniref_pdd.py --lossless-accel turns them all on.
On GB200 with 4 GPUs, MiniMax-H3 text to video with audio (124 frames,
1344x768) gives bit-identical frames and audio with the first three switches
on and off.
2026-10-07 15:35:37 -07:00
8530fdc550
[feat] Add external-launcher executor for multi-node inference ( #1746 )
...
Co-authored-by: Will Lin <willlin@nvidia.com >
Co-authored-by: William Lin <8941107+SolitaryThinker@users.noreply.github.com >
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com >
2026-10-07 04:46:26 -07:00
Raghav K and Claude Opus 5.5
315b25d26d
[new-model] Add Cosmos Predict2.5 distilled T2W and DFD V2W inference ( #1767 )
...
Adds the Cosmos Predict 2.5 distilled text-to-world sampler (sCM, per-frame timesteps) and the DFD video-to-world pipeline, with their schedulers, registry entries, presets and examples.
Maintainer review fixes (Swarm orders #718 and #748 ): the DiT receives one shared fps value (its RoPE table is batch-independent, so per-sample vectors crashed batch > 1; differing values now raise), non-finite rollouts raise FloatingPointError instead of being zeroed by nan_to_num, the DFD detector matches "dfd" only in the last path component, the distilled example defaults to 24 fps, and the distilled rollout keeps its state in FP64 through the scheduler step as the official driver does.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com >
2026-10-07 04:42:44 -07:00
af78c759f1
[new-model] Add Wan2.2-Animate-14B character animation/replacement inference ( #1765 )
...
Co-authored-by: SuhaanCommits <suhaan@cedar.build >
Co-authored-by: Claude Fable 5 <noreply@anthropic.com >
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
2026-10-07 04:19:37 -07:00
2452022f15
[new-model] Add LTX-2.5 inference support ( #1704 )
...
Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com >
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: CodeRabbit <noreply@coderabbit.ai >
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com >
2026-10-07 03:34:18 -07:00
4b3f99b224
[new-model] Added LingBot-World-Fast image-to-video support ( #1665 )
...
Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com >
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
2026-10-07 03:32:38 -07:00
37b9dc1836
[new-model] Add native Helios-Distilled T2V pipeline ( #1670 )
...
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com >
2026-10-07 03:18:24 -07:00
4bb5627fa8
[feat] MiniMax H3: block-FP8 text encoder on Hopper, its converter, and a 4xH100 FastH3 profile ( #1829 )
...
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com >
2026-10-07 02:31:52 -07:00
6e4813f8b4
[bugfix] use in-process executor for num_gpus=1 ( #1755 ) ( #1879 )
...
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com >
2026-10-07 02:18:23 -07:00
ef62ac48d1
[new-model] Add Wan2.2-S2V-14B speech-to-video (model port + checkpoint converter) ( #1683 )
...
Co-authored-by: SuhaanCommits <suhaan@cedar.build >
Co-authored-by: Claude Fable 5 <noreply@anthropic.com >
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
2026-10-07 01:36:01 -07:00
274b922d39
[misc]: surface GenerationResult.peak_memory_mb in the MiniMax H3 examples ( #1785 )
...
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com >
2026-10-07 01:29:56 -07:00
02a027c49f
[feat]: FastH3 on DGX Spark and Apple Silicon ( #1920 )
...
Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com >
Co-authored-by: Aryan Kumar <aryank@Aryans-Mac-Studio.local >
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com >
2026-10-06 12:17:11 -07:00
8b51380466
[feat]: FastH3 on single RTX GPUs ( #1919 )
...
Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com >
Co-authored-by: Aryan Kumar <aryank@Aryans-Mac-Studio.local >
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com >
2026-10-06 11:37:01 -07:00
f06d824054
[feat]: add Kandinsky-6.0 TI2VA and VSR pipelines ( #1924 )
...
Co-authored-by: leffff <levnovitskiy@gmail.com >
Co-authored-by: Kirill <kykozlov@gmail.com >
Co-authored-by: shaoxiongduan <shaoxiongduan@gmail.com >
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com >
2026-10-06 10:50:17 -07:00
8444c0897a
[feat] FastH3 Ref2VA PDD inference with reference-video VSA sparsity ( #1907 )
...
Co-authored-by: Davids048 <jundasu@ucsd.edu >
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com >
2026-10-02 11:28:07 -07:00
Aryan Kumar and Aryan Kumar
7f03e03dc6
[docs]: FastH3 V2 cookbook serving and MLX install guide ( #1884 )
...
Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com >
2026-09-29 13:03:18 -07:00
Aryan Kumar
361f919c88
[feat]: add a narrow FastH3 8-step MLX ladder
...
Keep the four-step uniform AdaLN cache, and use explicit DMD rungs only when the snapshot declares its own scheduler shifts.
2026-09-23 11:06:55 -07:00
William Lin
0100218594
[feat] Support the FastH3 8-Step V2 checkpoint: checkpoint-defined shifts, explicit DMD schedule, new example ( #1852 )
2026-09-15 14:13:45 -07:00
a943220c11
[bugfix]: drop dead h3_sequential_load from Spark FastH3 presets ( #1831 )
...
Co-authored-by: Cursor Agent <cursoragent@cursor.com >
Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com >
2026-09-08 12:38:06 -07:00
Junda Su
4e52f47d1e
[feat] add MXFP8 support on H3 ( #1796 )
2026-09-05 17:08:58 -07:00
Aryan Kumar and Aryan Kumar
0bd19a976b
[docs]: add one-Spark FastH3 cookbook runtime with a device-count row ( #1811 )
...
Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com >
2026-09-02 10:05:59 -07:00
aaef49bfc6
[feat]: run FastH3 across two DGX Sparks with Ray sequence parallel ( #1803 )
...
Co-authored-by: Kyle <shh075@ucsd.edu >
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com >
Co-authored-by: Satyam Srivastava <srivastavasatyam53@gmail.com >
Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com >
2026-09-01 01:37:23 -07:00
Aryan Kumar and Aryan Kumar
cf6a00b9be
[feat]: add opt-in CUDA TAEH3 preview decode for FastH3 ( #1795 )
...
Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com >
2026-08-31 23:34:56 -07:00
Aryan Kumar and Aryan Kumar
a28f2bab4b
[feat] Add an optional MLX TAEH3 preview decoder ( #1794 )
...
Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com >
2026-08-31 05:08:06 -07:00
Aryan Kumar and Aryan Kumar
f82d8be4bf
[perf] Sequential MiniMax H3 start with GPU-direct DiT load ( #1793 )
...
Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com >
2026-08-31 04:20:21 -07:00
Suhaan Khurana
29ff16ec96
[feat] Add MiniMax H3 MLX spatial fast mode ( #1789 )
2026-08-30 17:47:02 -07:00
a4d9a75e2c
[perf] Add MiniMax H3 MLX VSA and SIMD attention ( #1776 )
...
Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com >
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
2026-08-30 13:44:54 -07:00
ac48bb3cd1
[feat] Add MiniMax H3 MLX T2VA inference ( #1770 )
...
Co-authored-by: Aryan Kumar <aryank@Aryans-Mac-Studio.local >
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
Co-authored-by: CodeRabbit <noreply@coderabbit.ai >
2026-08-28 12:54:33 -07:00
William Lin and shaoxiongduan
a534ba20a0
[feat] Add MiniMax H3 LoRA inference and preview launchers ( #1771 )
...
Co-authored-by: shaoxiongduan <shaoxiongduan@gmail.com >
2026-08-27 14:43:39 -07:00
William Lin
9212f4f218
[ci] make GPU validation change-aware ( #1747 )
2026-08-25 21:26:25 -07:00
Aryan Kumar and Aryan Kumar
6388db815b
[bugfix] FastMetal-QAD MLX support: refuse CUDA QAD trees, use packed mlx_dit config, stream loads ( #1736 ) ( #1758 )
...
Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com >
2026-08-25 14:51:40 -07:00
William Lin
a837fe841a
[docs] Update FastH3 README ( #1749 )
2026-08-23 05:05:58 -07:00
William Lin
f9e3680f11
[perf] Align FastH3 optimized inference profile ( #1748 )
2026-08-23 02:01:03 -07:00
William Lin
13213395b4
[perf] Parallelize MiniMax-H3 VAE over sequence ranks ( #1744 )
2026-08-22 18:05:31 -07:00
William Lin
d3cff517cd
[perf] Add opt-in regional fullgraph compile for DiT inference ( #1741 )
2026-08-22 12:14:39 -07:00
William Lin
6d6a10be7a
[feat] FastVideo-Minimax-FastH3-Preview few-step example + 64-token-tile VSA-H3 inference path ( #1731 )
2026-08-21 12:40:09 -05:00
Shao Duan
c4ad4227c0
[misc] MiniMax-H3: move the AdaLN converter into scripts/checkpoint_conversion ( #1712 )
2026-08-21 02:15:00 -05:00
Aryan Kumar and Aryan Kumar
8537dcd6de
[feat]: Apple Silicon MLX runtime — INT8 Wan2.1 and Wan2.2 inference ( #1638 )
...
Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com >
2026-08-18 15:08:00 -07:00
Kai
0653f8f3af
[new-model] Add V2A: native MMAudio inference pipeline ( #1622 )
2026-08-09 17:29:27 -07:00
Shao Duan
541ef014ee
[perf] MiniMax-H3: rank-reduced AdaLN pruned model option (-39% params, -23 GiB VRAM) ( #1699 )
2026-08-09 12:31:57 -07:00
William Lin
9028953625
[misc] yapf pass under CI's interpreter (3.12) + pin hook language_version ( #1702 )
2026-08-08 22:24:17 -07:00
William Lin
15568f27db
[perf]: H3 torch.compile + CUDA graphs (1.2-1.3x) with denoising step marking ( #1689 )
2026-08-06 16:23:33 -07:00
Kaiqin Kong
e8b0e4c61e
[feat] Add MiniMax H3 ( #1674 )
2026-08-04 13:54:39 -07:00
William Lin
af2934dd6b
[feat]: FA4-FP4 ATTN_QAT_INFER on sm_100/sm_103 + NVFP4 weight purge ( #1647 )
2026-07-27 15:46:52 -07:00
SolitaryThinker