Commit Graph
12 Commits
Author SHA1 Message Date
KyleandClaude Opus 5 f247a307a8 [perf]: make deferred loading opt-in per pipeline and clean up on failure
Review of the first version found that the flag was unsafe as a global
default. Releasing a component and loading it again is only correct when
nothing outside the loader has changed it, and two habits in this tree break
that without raising:

  * mutating a component after load. LongCatPipeline.initialize_pipeline turns
    on block-sparse attention and writes parameters into every transformer
    block. That runs once, so a re-materialized component silently comes back
    with the feature off and the second generation quietly degrades.
  * reading a component's attributes while stages are built. The shared
    DenoisingStage.__init__ derives the attention backend from
    transformer.hidden_size, which materializes the DiT during post_init and
    defeats the deferral it was meant to gain.

_lazy_module_names is now empty in the base class, so an unchecked pipeline
gets no deferral and says so. MiniMax-H3 opts in to the four components this
PR measured, and nothing else changes behaviour.

Also from review:

Releasing on the failure path. A stage's hook frees only what that stage is
the last user of, and it ran only after a successful forward. Both the stage
and the whole run now release on the way out, so the retry a memory
constrained caller attempts does not start from a worse position than the
request that just failed. A failing release cannot replace the exception
being propagated.

Walking into nested stages. Cosmos25AutoDenoisingStage keeps the transformer
inside child stages, so a one-level scan called it unreferenced and never
freed it. The scan now recurses through stages and containers with cycle
protection.

Keeping the proxy when torch.compile is skipped. The FSDP check ran after the
proxy had already been replaced by the real module, so an FSDP-wrapped
component lost both the compile and its release hook.

Not attaching the activation trace to a deferred component, since the hook
manager pins every module it wraps.

test_parser.py asserted on a whole serialized config dict, so adding a field
to OffloadConfig broke it. Grepping the field name could not find that; only
running the suite could.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-28 22:36:51 -07:00
William Lin a159b63c67 [bugfix] Harden OpenAI serving after post-merge review (#1782) 2026-08-28 22:29:24 -07:00
Kaiqin Kong e8b0e4c61e [feat] Add MiniMax H3 (#1674) 2026-08-04 13:54:39 -07:00
Junda Su 755a4e4470 [new-model] Add LingBot-Video Dense and MoE/refiner T2V inference (#1595) 2026-07-18 20:18:16 -07:00
Junda Su d4bdd3621a [new-model] Port LingBot-World-v2 (#1579) 2026-07-16 19:51:53 -07:00
6966f9e0bc [fix] Z-Image (#1236) draft port: rebase + strict-load contract + bf16 encoder parity + PORT_STATUS (#1339)
Co-authored-by: Mrinaal Dogra <mdogra@ucsd.edu>
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com>
2026-07-16 18:18:25 -07:00
b063f8ca41 [feat] Fix FLUX.1-dev port: native RoPE, parity tests, SSIM reference (#1321)
Co-authored-by: Ishan Vaish <ivaish@ucsd.edu>
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-13 01:49:42 -07:00
6392bd40a9 [feat] Dreamverse 10/14: Add serving API contracts (#1333)
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>
2026-05-14 10:07:01 -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 4ddcdf541f [feat] [5.5/n] Improve API: streaming server config surface + serve dispatch (#1238) 2026-04-17 15:36:21 -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
William Lin 76c62d7a00 [feat] [1/n] API improvements: add intial files for new fastvideo public API (#1218) 2026-04-05 18:13:19 -07:00