nodes/hao-ai-lab-FastVideo
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
698 Commits 339 Branches 8 Tags
345d2acb19d7248abe41e9e7feeb8a1ffb4d1afd
Commit Graph
698 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
SolitaryThinker 345d2acb19 [fix]: init PackedExpertLinear params with zeros, not empty (review) 
Addresses review thread on fastvideo/models/dits/magi_human.py:263. torch.empty leaves memory uninitialized; if a future caller skips load_state_dict or strict=False loses entries silently, the network would output NaNs. zero-init costs nothing at runtime (overwritten by load) and matches sibling MultiModalityRMSNorm which also uses torch.zeros.
 2026-05-12 14:09:55 -07:00
SolitaryThinker 746d28f16d [fix]: skip MagiHuman DiT parity until pipeline lands (step 4/8) 
The DiT parity tests imported build_packed_inputs from fastvideo.pipelines.basic.magi_human.stages.latent_preparation, which ships in step 4/8 of the PR #1280 decomposition. On this branch (3/8 — DiT only) the import would crash with ImportError when assets were present, instead of cleanly skipping. Wrap the import in try/except so the test is true dead code on this branch, per the decomposition contract: 'every intermediate state is dead code, not broken code.'
 2026-05-12 14:06:38 -07:00
SolitaryThinker 4da00b35f1 [feat]: MagiHuman DiT (transformer) - port + parity tests 
15B-param single-stream transformer for daVinci-MagiHuman: 40 layers,
hidden=5120, head_dim=128, GQA num_query_groups=8. Joint audio-visual
denoising in a unified token sequence (no cross-attention).

Pure addition; the DiT is dead code on the stack tip until the activation
switch in 8/8.

Files:
- fastvideo/models/dits/magi_human.py (architecture port)
- fastvideo/configs/models/dits/magi_human.py (MagiHumanVideoConfig)
- fastvideo/configs/models/dits/__init__.py (+1 export)
- tests/local_tests/helpers/magi_human_upstream.py (loader for the
  upstream daVinci-MagiHuman reference - shared by parity tests)
- tests/local_tests/magi_human/test_magi_human_parity.py (base, bit-exact)
- tests/local_tests/magi_human/test_magi_human_distill_parity.py
  (DMD-2 distill, bit-exact)

Two debug-scratch files were intentionally dropped during decomposition:
_debug_magi_human_block_parity.py and _debug_magi_human_weight_diff.py.

Magi-Stack: 3/8
 2026-05-12 14:01:53 -07:00
William Lin b1ddb1ba33 [feat]: T5-Gemma encoder for MagiHuman pipeline (2/8) (#1296) 2026-05-12 13:59:55 -07:00
Junda SuSolitaryThinkerMatthew NotoXOR-opZhang Peiyuan
 		9ff65c83b6 [feat] Dreamverse 03/14: Add backend skeleton (#1326) 
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com>
Co-authored-by: Matthew Noto <99706358+RandNMR73@users.noreply.github.com>
Co-authored-by: XOR-op <17672363+XOR-op@users.noreply.github.com>
Co-authored-by: Zhang Peiyuan <42993249+jzhang38@users.noreply.github.com>
 2026-05-12 13:58:23 -07:00
Junda SuSolitaryThinkerMatthew NotoXOR-opZhang Peiyuan
 		15473f3c77 [docs] Dreamverse 02/14: Add app documentation (#1325) 
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com>
Co-authored-by: Matthew Noto <99706358+RandNMR73@users.noreply.github.com>
Co-authored-by: XOR-op <17672363+XOR-op@users.noreply.github.com>
Co-authored-by: Zhang Peiyuan <42993249+jzhang38@users.noreply.github.com>
 2026-05-12 13:34:42 -07:00
William Lin 490641cf47 [infra]: MagiHuman housekeeping (gitignore, codespell, skills index) (1/8) (#1295) 2026-05-12 12:50:01 -07:00
Junda SuSolitaryThinkerMatthew NotoXOR-opZhang Peiyuan
 		4ad6880ca6 [docs] Dreamverse 01/14: Add integration provenance (#1324) 
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com>
Co-authored-by: Matthew Noto <99706358+RandNMR73@users.noreply.github.com>
Co-authored-by: XOR-op <17672363+XOR-op@users.noreply.github.com>
Co-authored-by: Zhang Peiyuan <42993249+jzhang38@users.noreply.github.com>
 2026-05-12 11:42:42 -07:00
Raghav K 9d0af28307 [feat] Add Cosmos 2.5 T2W training pipeline (LoRA + full fine-tune) (#1227) 2026-05-11 19:15:00 -07:00
Kaiqin Kongmignonjiaalexzmsmergify[bot]
 		d6dfe95466 [feat] FastVideo World Model Training (#1179) 
Co-authored-by: mignonjia <mhuo@ucsd.edu>
Co-authored-by: alexzms <3036648523@qq.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
 2026-05-11 05:13:12 +00:00
alexzmsandmergify[bot] 636d3b743e [misc] attention hot-path cleanup + denoising loop hoists (#1272) 
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
 2026-05-10 11:51:10 -07:00
William LinandRaghav e3a5c6954f [misc]: import add-model skill stack to .agents/skills/ (#1308) 
Co-authored-by: Raghav <ragg04@gmail.com>
 2026-05-09 13:06:54 -07:00
Aryan KumarAryan Kumaralexzmsmergify[bot]
 		f633e30ebb [feat]: add LongCat bidirectional finetuning support (#1244) 
Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com>
Co-authored-by: alexzms <3036648523@qq.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
 2026-05-09 12:20:43 -07:00
William Lin 5ce4947ac2 [ci] mergify: accept [skill]/[skills] and [infra] PR title tags (#1309) 2026-05-08 17:14:21 -07:00
William Lin 323d74c0d2 [skills]: New skill - decompose-pipeline-pr (#1303) 2026-05-08 16:05:58 -07:00
William Lin d98aeafc86 [feat]: Loader umbrella-repo support + optional component dirs (#1294) 2026-05-08 15:24:26 -07:00
Shao Duan f6396fb8c6 [feat] Add fastvideo.eval video evaluation suite (#1305) 2026-05-07 20:31:15 -07:00
William Lin 6300329cd5 [infra]: Add activation trace hooks for pipeline debugging (#1293) 2026-05-07 16:38:13 -07:00
MookandSolitaryThinker c17d33bf33 [ci] Replace flaky LTX-2 pixel SSIM with latent-slice cosine regression (#1253) 
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com>
 2026-05-05 03:36:58 -07:00
William LinJunda SuMatthew NotoXOR-opZhang Peiyuan
 		2aaeee2ab8 [feat] Improve API: streaming router (multi-replica load balancer + ws proxy) (#1286) 
Co-authored-by: Junda (David) Su <90978028+Davids048@users.noreply.github.com>
Co-authored-by: Matthew Noto <99706358+RandNMR73@users.noreply.github.com>
Co-authored-by: XOR-op <17672363+XOR-op@users.noreply.github.com>
Co-authored-by: Zhang Peiyuan <42993249+jzhang38@users.noreply.github.com>
 2026-05-05 03:00:08 -07:00
William LinJunda SuMatthew NotoXOR-opZhang Peiyuan
 		eb3a394224 [feat] Improve API: streaming auxiliaries (safety, rewrite, logger, mock) (#1284) 
Co-authored-by: Junda (David) Su <90978028+Davids048@users.noreply.github.com>
Co-authored-by: Matthew Noto <99706358+RandNMR73@users.noreply.github.com>
Co-authored-by: XOR-op <17672363+XOR-op@users.noreply.github.com>
Co-authored-by: Zhang Peiyuan <42993249+jzhang38@users.noreply.github.com>
 2026-05-05 00:14:34 -07:00
William LinJunda SuMatthew NotoXOR-opZhang Peiyuan
 		f673423b51 [feat] Improve API: streaming prompt enhancer with LLMProvider abstraction (#1258) 
Co-authored-by: Junda (David) Su <90978028+Davids048@users.noreply.github.com>
Co-authored-by: Matthew Noto <99706358+RandNMR73@users.noreply.github.com>
Co-authored-by: XOR-op <17672363+XOR-op@users.noreply.github.com>
Co-authored-by: Zhang Peiyuan <42993249+jzhang38@users.noreply.github.com>
 2026-05-04 13:44:40 -07:00
William LinJunda SuMatthew NotoXOR-opZhang Peiyuan
 		eb0a41528a [feat] Improve API: streaming server GpuPool + worker subprocess (#1257) 
Co-authored-by: Junda (David) Su <90978028+Davids048@users.noreply.github.com>
Co-authored-by: Matthew Noto <99706358+RandNMR73@users.noreply.github.com>
Co-authored-by: XOR-op <17672363+XOR-op@users.noreply.github.com>
Co-authored-by: Zhang Peiyuan <42993249+jzhang38@users.noreply.github.com>
 2026-05-04 12:56:31 -07:00
William Lin 140bd1a6cf [misc]: standardize install instructions on uv pip install (#1279) 2026-05-02 12:45:50 -07:00
William Lin 11f5a8e582 [misc] pin torch to 2.11.0 (#1277) 2026-05-02 11:48:07 -07:00
Satyam SrivastavaSatyam SrivastavaSatyam SrivastavaSolitaryThinker
 		71b3cb8c34 [ci] Add CI Performance Regression Tracking Changes (#1248) 
Co-authored-by: Satyam Srivastava <satyam53@Mac.lan1>
Co-authored-by: Satyam Srivastava <satyam53@Satyams-MacBook-Air.local>
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com>
 2026-05-02 03:29:53 -07:00
William Lin c85f6a477f [docs] add hierarchical AGENTS.md per-directory guidance (#1278) 2026-05-02 03:28:18 -07:00
Junda Su 40d4930d73 [bugfix] Update fa import (#1271) 2026-05-02 01:25:22 -07:00
William Lin f9be085243 [ci] pre-commit: drop stale excludes + document agent lint flow (#1276) 2026-05-02 01:19:06 -07:00
William Lin 36b53ff350 [bugfix]: classify stable_audio fields in schema parity inventory (#1275) 2026-05-02 00:12:10 -07:00
William Lin 9801037c3d [refactor] tests/local_tests: organize by model family (#1269) 2026-05-01 01:49:54 -07:00
alexzmsandmergify[bot] 74d09b0efd [misc] cleanup: grad-norm asserts, dead offload file, callback names (#1268) 
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
 2026-05-01 01:16:13 -07:00
alexzms 38dc8820ac [ci] add CPU unit tests for train callback system in fastvideo.train (#1267) 2026-05-01 00:53:29 -07:00
William Lin c77a76c6af [feat] Stable Audio Open 1.0: T2A + A2A + RePaint inpainting (native) (#1260) 2026-05-01 00:07:11 -07:00
alexzms d14d5aadea [feat] Cosmos 2.5 training support in fastvideo.train (#1224) 2026-05-01 01:15:02 +00:00
alexzms 4c915b7742 [ci] add CPU unit tests for train checkpoint utilities in fastvideo.train (#1265) 2026-04-29 18:55:39 +00:00
alexzms 9a8bbe18fa [bugfix]: fix SP deadlock in negative prompt encoding during training (#1178) 2026-04-28 01:06:49 +00:00
alexzms ea25441ef0 [ci] add CPU unit tests for fastvideo.train load_run_config (#1264) 2026-04-28 01:06:18 +00:00
Satyam SrivastavaSatyam Srivastavamergify[bot]
 		48957fcde1 [bugfix] Fix modal remote functions crash container on sys exit in CI remote functions (#1261) 
Co-authored-by: Satyam Srivastava <satyam53@Mac.lan1>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
 2026-04-27 21:50:55 +00:00
Mook 7b872cc41e [Perf] Skip bool-mask round-trip in block-sparse VSA attention (#1243) 2026-04-26 15:14:37 -07:00
alexzms 37418946c8 [docs]: clarify real_score_guidance_scale CFG parameterization (#1256) 2026-04-26 16:38:00 +08:00
William Lin 95fd29e0cb [feat] Streaming WebSocket server skeleton (single generator + fMP4) (#1251) 2026-04-26 00:33:49 -07:00
Junda Suandmergify[bot] e17cd2633c [bugfix]: normalize uint8 pil_image in I2V VAE encoding (#1249) 
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
 2026-04-24 09:16:01 +00:00
William Lin e0dc5f2b0c [feat] Add typed LTX-2 continuation state and streaming session store (#1250) 2026-04-24 01:28:07 -07:00
William Lin 70ee5d230c [feat] [6/n] Improve API: LTX-2 public preset + asset wiring + gpu_pool translation (#1239) 2026-04-23 11:36:45 -07:00
William Lin 24ced500f5 [test] add LTX-2 distilled T2V SSIM regression test (#1240) 2026-04-21 12:03:38 -07:00
William Lin 4ddcdf541f [feat] [5.5/n] Improve API: streaming server config surface + serve dispatch (#1238) 2026-04-17 15:36:21 -07:00
William Lin 0e3529869c [feat] [5/n] Improve API: wire ServeConfig.default_request into OpenAI serving (#1237) 2026-04-17 13:26:18 -07:00
William Lin e1e0d91c00 [misc] small cleanup for API handling (#1235) 2026-04-16 16:21:21 -07:00
William Lin 145a3f166b [feat] [4/n] Improve API: refactor sampling param and merge with presets (#1234) 2026-04-16 14:10:02 -07:00