The torch adapter boundary marshalled the latent host<->device on EVERY denoise
step: _t uploaded the latent (and re-uploaded the text embeds) and _n downloaded
the velocity with a forced CUDA sync — 2*N PCIe copies + N syncs per generation,
buying nothing, since the latent could stay resident on the GPU the whole loop.
Root cause was a numpy loop surface. But the loop MATH is already array-agnostic
(CFG combine + flow-match Euler are pure arithmetic; the solver kernel already
passes torch through). So introduce a per-platform array namespace (v2/platform/
array_ns): numpy on CPU (torch-free — the parity mini is unchanged), torch-on-
device on cuda. The latent is seeded with numpy and uploaded ONCE; it then stays
resident through forward -> CFG combine -> solver -> next step; a single host
marshal happens at the request/output boundary (engine._to_artifact).
Opt-in per recipe via ModelCard.device_io (set on the Wan cards). When set on a
GPU box, build_component flips the components' TorchComponent.device_io so _out
keeps tensors on-device (in fp32, matching the old _to_numpy cast so the combine
dtype is unchanged). Un-migrated families and the CPU toy keep numpy in/out.
Also: PrecisionPolicy.cast is array-preserving; _t accepts resident tensors.
Verified BIT-IDENTICAL on real Wan2.1-1.3B / H100: the on-device latent equals
the pre-change numpy-path latent exactly (max_abs_diff 0.0, np.array_equal True).
CPU mini holds 237 passed / 5 pre-existing; pre-commit clean.
Other WanDenoiseLoop families can flip device_io next (per-family GPU re-verify);
non-Wan loops migrate to the xp namespace later.
platform/backends/ had become a flat dump of 15 per-model torch_<model>.py
adapters next to the genuinely-shared infra. Each adapter is referenced from
its card by a plain 'module:Class' string loaded via importlib, so there was
no real coupling forcing it into platform/ — the Cosmos/Flux/etc adapter
belongs WITH its recipe (card/loop/program).
Move each torch_<model>.py -> v2/recipes/<model>/adapter.py and flip the card
strings to v2.recipes.<model>.adapter:<Class>. backends/ now holds only the
shared substrate (torch_backend base, torch_cuda registration, torch_kernels,
toy/cpu/accel). Each recipe is now a self-contained package.
Cross-refs updated: gen3c/adapter imports CosmosT5Encoder from cosmos2/adapter;
sd35/program + stable_audio/card import from their own package. Recipes still
import torch-free (adapters pulled only via the string on a GPU box) — CPU mini
holds 237 passed / 5 pre-existing (bucket_c torch-absent). pre-commit clean.
Sweep all of v2/ through pre-commit (was previously only run on changed
files). Fixes surfaced across untouched modules:
- ruff: isinstance-tuple -> X | Y (UP038), try/except/pass ->
contextlib.suppress (SIM105), negated-return (SIM103), %-format ->
f-string (UP031).
- mypy: add annotations for no-untyped-call + var-annotated across recipes,
training methods, torch/toy backends, and serving.
- yapf reflow of the SF-Wan KV-cache call sites (semantics unchanged).
yapf/ruff/codespell/mypy all pass; v2 tests 237 passed / 5 pre-existing
(bucket_c torch-absent on CPU venv).
The causal adapter never passed a kv_cache, so CausalWanTransformer3DModel.forward routed to
_forward_train (no cross-chunk KV) on every chunk instead of _forward_inference (the CausVid
Alg-2 KV-cache path). Each chunk denoised blind to the previous ones; the loop's cross-chunk
"context" was a toy mean(prior_latents) the adapter ignored. Result: hard discontinuities at
every chunk boundary (frame-to-frame absdiff spikes 43-55 every ~12 frames).
Fix (cuda path only; toy/CPU path and the 237-test suite untouched):
- WanDiT.alloc_causal_caches(): allocate the persistent per-block KV + cross-attn caches sized
from the model config (mirrors CausalDenoisingStage._initialize_kv_cache).
- WanDiT.__call__: thread kv_cache/crossattn_cache/current_start/cache_start/start_frame/
frame_seqlen so the model runs _forward_inference.
- wan_causal/loop.py: own the caches in LoopState (per-request -> interleave-safe); pass
current_start = chunk_idx*chunk_size*frame_seqlen per chunk; do the clean-KV write
(timestep ~0) after each chunk so the next attends to it.
Verified on H100: frame-to-frame absdiff mean 12.8->4.4, max 55.4->8.3; chunk-boundary spikes
eliminated; coherent across all 7 chunks. CPU causal toy tests unchanged (24 passed).
Sweep all 296 v2 modules: simplify verbose docstrings/comments and remove 376 dangling
"(design_vN §X)" citations to the now-deleted design docs (v2/README.md is the source of
truth). Comment/docstring-only — AST-verified code-identical; the CPU suite holds at 237
passed / 5 pre-existing. Also applies yapf + ruff --fix auto-fixes (import ordering,
forward-ref annotation de-quoting under `from __future__ import annotations`; behavior-
neutral, suite-confirmed) and adds the legitimate domain terms mot/clen/te to the codespell
ignore-list. Remaining ruff (24) + mypy (68 no-untyped-call) findings are pre-existing v2
debt, untouched here.
Unify the four design docs (design.md, designv2.md, design_v3.md, designv4.md) into a single
authoritative v2/README.md: the (recipe, runtime) thesis, driven loops, planes, one-WorkUnit
scheduler, the parity ladder + interleave gate, training-on-shared-loops, the weight-sharing
topology catalog, the current GPU status (20+ models + the BAGEL/Qwen-Omni/Cosmos3 trio
verified), and a prioritized roadmap. Recast design_summary.md as a one-page design philosophy
pointing to it. Add .agents/exploration/mstar-v2-roadmap.md (the adversarially-verified M*
Walk-Graph gap analysis driving the roadmap). Delete the four superseded design docs.
MatrixGame2 (causal DMD loop): bump st.step_idx on every executed work unit (each
DMD step and each clean-context pass). The loop drives its own control flow off
block_idx/dmd_idx/phase, but the runtime's no-progress watchdog keys on
st.step_idx, so a multi-block causal rollout was seen as stalled. Mirrors what
every other recipe loop does.
MatrixGame3 (5B WanModel): patch-align the latent H/W (patch_size (1,2,2)) before
denoise. The model folds (H/2, W/2) tokens, so an odd latent dim made the
unpatchified velocity come back one row/col short of the noise latent. Crop to
(latent // patch) * patch, faithful to MatrixGame3DenoisingStage.
v2 mini: 240 passed.
Two distilled few-step video models rendered incorrectly on GPU; root-caused via
dense frame sampling (contact sheets) and fixed in the recipe cards.
SF-Wan2.1 (self-forcing causal, wan_causal card): was oversaturated/overcooked.
The distilled student is CFG-FREE (guidance 1.0, single forward/step) and denoises
with the 4-step DMD schedule [1000,750,500,250] (warped by FlowShiftPolicy(5.0)),
at a native causal block of 3 latent frames. Defaults were ClassicCFG@6.0 + 2 steps
+ block 2 -> overcooked AND under-denoised. Fixed: num_chunks=7, chunk_size=3,
steps_per_chunk=4; SamplingDefaults num_steps=4, guidance_scale=1.0 (7x3=21 latent
-> 81 frames). Renders a clean raccoon-in-sunflowers across all 81 frames.
LTX2-Distilled 2-stage SR (ltx2 card + LTX2VAE): was temporally blocky. Root cause
was an OOM-forced 57-frame reduction (only 8 latent temporal frames); the model is
designed for 121 (16 latent frames). Enable VAE tiling in LTX2VAE so the 121-frame
full-res decode fits the 80 GiB GPU; keep base cfg_scale=3.0 (drives brightness;
cfg=1 washed out) with stg_scale=0.0 (v2's drop-text perturbation is not real
skip-layer STG). Renders the on-prompt backyard shot, bright + temporally coherent.
torch_backend.py: enable LTX2VAE tiling; clear pre-existing mypy no-untyped-call /
yapf debt on the file (surfaced once per-file linting bypassed the duplicate-module
flakiness) by annotating the helper/constructor/maker signatures.
Tests: update the 3 affected CPU defaults/chunk-count tests. v2 mini: 240 passed.
Second GPU pass (distilled + huge dense, 2-wide). 5 verified end-to-end with real
weights; CPU toy path kept green (240 passed, 2 skipped).
VERIFIED:
* matrixgame3 — mp4 (9,256,256,3), zero fixes. 6.47B, standard Wan attn (NOT
sparse-attn-blocked, like its mg2 sibling); degenerate single-clip.
* fastwan — FastWan2.2-TI2V-5B-FullAttn DMD 3-step, mp4 (17,256,256,3), zero
fixes. The FULL-ATTENTION variant has no VSA params -> the generic
Wan loader maps it cleanly (reuses WanDiT via load_id, no adapter).
* longcat — LongCat-Video-T2V 13.58B, mp4 (17,256,256,3), zero fixes, CPU
offload (~40GB peak).
* sfwan22 — Self-Forcing Wan2.2-A14B causal+MoE (2x14B), mp4 (29,288,288,3),
CPU expert offload (~80GB peak).
* lingbotworld — Wan2.2-class 2x14B camera world model, mp4 (9,256,256,3), offload
(~98GB peak transient).
BLOCKED: turbowan-i2v-a14b — SLA sparse-attn params (attn1.attn_impl.proj_l on all
40 layers of both experts) cannot load into the dense Wan build + needs the
fastvideo-kernel SLA Triton kernels (no nvcc here). Confirmed via the safetensors
header (no 60GB download). Same SLA family as turbowan-1.3b.
Fixes (own-port only): sfwan22/loop.py, lingbotworld/{card,program}.py +
torch_lingbotworld.py. pre-commit clean per-package.
VideoGenerator._result hardcoded out.artifacts['video'].frames, so an audio-only
(Stable Audio) or image-only (SD3.5 / FLUX.2 T2I) generation crashed with
KeyError 'video' even though the engine had correctly produced the AudioArtifact /
image TensorArtifact (surfaced during stable_audio GPU bring-up). _result now
guards on the artifact present: video -> mp4 (unchanged), else image -> png
([C,H,W]/[B,..] normalized to [H,W,C]), plus the existing audio -> sibling .wav;
image_path recorded in result.extra. The video path is byte-for-byte unchanged.
v2 mini 240 passed, 2 skipped; pre-commit clean.
Ran each dense public port through the real VideoGenerator on H100 (2-wide across
both GPUs). 8 produce real finite output end-to-end; per-model adapter/loop fixes
landed in each port's OWN files (no shared/fastvideo edits). CPU toy path kept
green (240 passed, 2 skipped) — GPU-only conditioning gated to the cuda backend.
VERIFIED (real GPU output):
* stable_audio — stereo audio (2, 441000) @44.1kHz. Fixes: dedicated
'conditioner' component kind (SA owns its T5, empty
text_encoder_configs) + ConditionerLoader from conditioner/
+ VDenoiser c_noise = atan(sigma)/(pi/2).
* matrixgame2 — mp4 (9,256,256,3). Loads CLEAN (not sparse-attn-blocked).
Fixes: 20ch cond_concat (4ch mask + 16ch img), mandatory i2v
(synth blank first frame), pre-sized kv_cache/crossattn_cache
(SDPA inference path, avoids flex_attention compile), bf16
autocast, per-request reset_caches.
* gen3c — video (9,256,256,3), zero fixes (worked first try).
* wan_fun_control — video (9,256,256,3).
* lucy_edit — video (17,256,256,3).
* hunyuangamecraft— video (9,256,256,3).
* hunyuan_video — video (3,9,256,256), dual LLaMA+CLIP + Hunyuan VAE.
* hunyuan_video15 — video (9,256,256,3), dual Qwen+ByT5 (gated to cuda; CPU passes
single embed).
BLOCKED (not v2 recipe bugs — load/run reached, then a fastvideo/env wall):
* cosmos25 — DiT + VAE ran finite on GPU; the Qwen2.5-VL Reason1 encoder hits a
transformers 5.12.1 incompat in fastvideo shared code
(Qwen2_5_VLConfig.pad_token_id).
* kandinsky5 — fastvideo registry.py registers it with a bare PipelineConfig (no
Kandinsky5 config) -> load fails fastvideo-side. (latent z=16 +
visual_cond adapter corrected; toy decoupled to its own channels.)
* hyworld — fastvideo's hyworld DiT hardcodes flash_attn (no SDPA fallback);
flash_attn kernel not built here.
CPU-safety fixes (mine): kandinsky5 toy ToyDiT/ToyVAE use the toy LATENT_CHANNELS
(not the real z=16); hunyuan_video15 dual-encoder packing gated to cuda.
Sparse-attn distilled (turbowan/SLA, fastwan/VSA) + gated (cosmos2/flux2/sd35) +
huge (>80GB) handled separately. pre-commit clean per-package.
Adds explicit ModelEntry aliases for the LTX-2 (FastVideo/LTX2-Diffusers,
LTX2-base, Lightricks/LTX-2 -> single-stage base) and LTX-2.3 (LTX2.3-Diffusers,
LTX2.3-Distilled-Diffusers, LTX2.3-base, Lightricks/LTX-2.3, lightricks/ltx-2.3
-> the distilled joint-A/V card) naming variants of already-ported LTX
checkpoints (the arch fallback also resolves LTX2Transformer3DModel from a root).
v2 now resolves 63/64 fastvideo registry ids by exact id; the only remaining id,
FastVideo/Wan2.1-VSA-T2V-14B-720P-Diffusers, is ENV-BLOCKED (VSA Sparse-Linear
Attention kernels require nvcc, not built in this bring-up; it arch-resolves to
the base Wan card but needs the VSA kernel build to run faithfully).
v2 mini 240 passed, 2 skipped; pre-commit clean.
Closes the bucket-B sampler/conditioning gap — each reuses the Wan/Causal ARCH
(no new torch adapter) with a new in-package loop/sampler/conditioning, declared
in _BUCKET_C as explicit-HF-id-only (transformer_cls="" so the generic Wan/Causal
arch fallback is NOT hijacked — only the exact id distinguishes the capability
variant from a base Wan of the same class).
* turbowan — TurboWan rCM (Reparameterized Consistency Model) few-step: a
faithful in-package RCMScheduler port (TrigFlow->RectifiedFlow
schedule + stochastic consistency SDE step), 1.3B/14B T2V +
TurboWan2.2-I2V-A14B (MoE i2v, boundary 0.9 in raw-sigma space)
* lucy_edit — Lucy-Edit v2v editor: a video_vae_encode node (the input video
-> 48ch cond latent) threaded via the shared i2v_cond hook ->
96ch Lucy DiT input (faithful to denoising.py is_lucy_edit)
* wan_fun_control — Wan2.1-Fun-Control: control-video conditioning (reuses the
i2v [mask|cond] concat pattern)
* sfwan22 — Self-Forcing Wan2.2-A14B: causal chunk_rollout + Wan2.2 MoE
boundary routing, i2v (boundary 0.9) + t2v (boundary 0.875)
* fastwan — FastWan DMD 3-step: TI2V-5B-FullAttn loadable; the VSA-trained
variants + non-strict to_gate_compress load are BRINGUP
All 12 residual ids resolve+build; base Wan/Causal resolution unchanged (arch
fallback not hijacked). The _BUCKET_C regression test auto-extended -> v2 mini
240 passed, 2 skipped; pre-commit clean (per-package + registry).
Data-driven from registry._BUCKET_C (+ cosmos2): each net-new ported arch
resolves through the registry (exact id + arch fallback) AND runs end-to-end on
the CPU toy backend via the public Engine path (resolve -> build card+program ->
load_card -> Engine.run), emitting exactly one modality-correct artifact
(video / image / audio) + latents. Auto-covers future _BUCKET_C rows.
21 tests pass; full v2 mini 232 passed, 2 skipped.
First net-new architecture ported via the self-contained recipe-package pattern
(card-declared adapter + new loop, no shared-dispatch edit):
* CosmosDenoiseLoop (v2/recipes/cosmos2/loop.py): EDM preconditioning folded into
a flow-match Euler integrator. Faithful port of CosmosDenoisingStage — Karras
sigma schedule (rho=7, sigma_max=80 -> sigma_min=0.002, terminal clamp), latent
init randn*sigma_max, per-step c_in/c_skip/c_out (sigma_data=1) -> x0, CFG in x0
space, x0 -> velocity (x-x0)/sigma, FLOW_MATCH_STEP. video2world frame-replace
conditioning threaded but inert for the t2v preset.
* build_karras_sigmas helper added to v2/loop/sampler.py.
* CosmosDiT + CosmosT5Encoder adapters (v2/platform/backends/torch_cosmos.py),
declared on the card via ComponentSpec.adapter (the ac29750b enabler) — DiT
returns raw EDM output + builds the mandatory zero condition/padding masks + fps;
T5 uses the raw last_hidden_state (no Wan zero-pad). Reuses the WanVAE adapter.
* card/program/registry (HF id nvidia/Cosmos-Predict2-2B-Video2World + arch
fallback on CosmosTransformer3DModel) + COSMOS_NEG prompt + SamplingDefaults
(35 steps, gs 7, 704x1280, 93f, 16fps).
CPU-verified: Karras schedule, card/program build, registry resolve (id + arch),
EDM loop runs end-to-end on the CPU toy backend. GPU load/run is BRINGUP.
v2 mini green (211 passed, 2 skipped); pre-commit clean.
A new architecture can declare its own torch adapter on the card
(ComponentSpec.adapter="module:Class") instead of editing the shared _make_dit/
_make_vae/_make_text_encoder dispatch. _explicit_adapter() constructs it as
cls(module, *extra, device=, dtype=) and short-circuits the built-in Wan/LTX2
class-name dispatch when set. This makes each bucket-C port a self-contained
recipe package (card + adapter module + loop + program) with no shared-file edit
-> conflict-free parallel porting. Unset -> unchanged built-in dispatch.
CPU mini green (211 passed, 2 skipped).
Reuses everything: 2 WanTransformer3DModel experts + BoundaryTimestepRouting (from the A14B MoE pattern),
the CLIP image encoder + first-frame [mask|cond] conditioning + the i2v program (from the Fun-InP i2v
port), and the shared WanDenoiseLoop (i2v hooks + the boundary expert). No new adapter. CPU-verified: the
toy MoE i2v runs end-to-end (2 experts + boundary + conditioning -> finite video); resolves with i2v caps.
Structural (GPU-pending: 2x14B, like the A14B T2V). Wan family now largely covered (T2V 1.3B/14B/TI2V-5B/
A14B, causal SF, i2v 1.3B/14B/A14B). CPU mini 211/2.
Real image-to-video, unlocking the i2v cluster. v2/recipes/wan21/i2v.py: CLIP image-encode -> the DiT's
encoder_hidden_states_image; first-frame VAE conditioning + a 4-channel mask -> the 20ch [mask|cond] that
the Wan adapter concatenates with the 16ch noise -> the 36ch i2v DiT input (mirrors fastvideo's
ImageEncodingStage + ImageVAEEncodingStage; v2's WanVAE.encode already applies the matching (z-mean)/std).
Reuses the shared WanDenoiseLoop (its None-default i2v hooks) + the Wan torch adapter unchanged. Adds
ToyImageEncoder + the image_encoder checkpoint subfolder stamp. Registered Wan2.1-Fun-1.3B-InP.
GPU-verified: loads via the generic Wan loader (1.56B, no param-mapping issue), runs the full i2v
conditioning, produces real video (3,9,256,384, std 0.44, finite, motion 0.041). CPU mini 211/2 (T2V
unregressed). BRINGUP: visual confirmation that the output follows the conditioning image is
human-in-the-loop.
Threads i2v conditioning through the SHARED WanDenoiseLoop with zero T2V risk: init() reads optional
slots i2v_cond (the [mask|cond] latent) + i2v_img_embeds (CLIP) into scratch; _velocity passes them to
the dit (context=, cond=); WanDiT concats cond (16->36ch) and uses the embeds as
encoder_hidden_states_image; capture is disabled only when i2v conditioning is present. For T2V both are
None -> the dit call, CFG, and cudagraph capture are byte-identical (CPU mini 211 pass, 2 skip — no
regression). ToyDiT accepts+ignores cond (image-conditioning is a GPU-path concern). Completes the i2v
backend seam; the program's mask+cond construction + the Wan i2v card/registry + GPU verify follow.
Adds CLIPImageEncoder (encode_image -> the DiT's encoder_hidden_states_image) + the generic builder's
image_encoder maker (ImageEncoderLoader + ImageProcessorLoader), registered as the cuda 'image_encoder'
kind. Mirrors fastvideo's ImageEncodingStage. CPU-verified (component-kinds + lazy invariant, 211/2);
GPU path marked BRINGUP/written-not-run (processor subfolder + dtype to confirm on a real i2v checkpoint),
matching how the rest of the torch backend was originally landed. Reusable by the Wan i2v cluster + many
bucket-C models (Hunyuan/Cosmos i2v). Next i2v increments: the mask+cond latent construction + the
concat-into-DiT-input loop, then the card + registry + GPU-verify with Wan2.1-Fun-1.3B-InP.
GPU verification (Wan2.1-Fun... no: FastWan2.1-T2V-1.3B) failed at load: 'Parameter blocks.0.to_gate_compress.bias
not found in custom model state dict' — the FastVideo/* DMD-distilled checkpoints carry gated-attention
params (to_gate_compress) that the generic WanTransformer3DModel loader can't map. v2/registry.py's
select_by_architecture ALREADY rejects WanDMDPipeline for exactly this reason; my explicit ModelEntry
wrongly bypassed it. Reverted FastWan (1.3B + 14B-480P) and the unverified VSA-14B alias (same FastVideo/*
risk). Kept the official Wan2.1-T2V-14B (standard weights, same loader path as the GPU-verified 1.3B).
Lesson recorded in V2_PORTING_STATUS.md: FastWan/Turbo/VSA need a param-mapping fix (like LTX-2.3 did),
not just a schedule — bucket-C-effort. GPU-verify every port before claiming support. CPU mini 211/2.
Same WanTransformer3DModel arch (VSA is an attention-backend choice, not a weight/arch difference); v2
runs dense TORCH_SDPA, so it resolves to build_wan_t2v_14b_card. Registry resolution verified; the GPU
forward path is the 1.3B-proven Wan adapter (specific 14B/VSA weights not separately GPU-run).
First bucket-B port toward 'support every fastvideo model': Wan2.1-T2V-14B reuses the Wan recipe +
torch adapter unchanged (same WanTransformer3DModel/AutoencoderKLWan/UMT5) — only a registry entry +
build_wan_t2v_14b_card (720p, flow_shift 5.0) + SamplingDefaults differ. Without the entry the arch
fallback would give it the 1.3B 480p defaults; the explicit entry gives 50 steps / 720x1280.
Also recorded the full backlog in V2_PORTING_STATUS.md: 63 fastvideo models = 8 ported / 21 bucket-B
(reuse Wan/Causal/LTX2 arch — registry+recipe+defaults, no new adapter) / 34 bucket-C (13 new
architectures needing a TorchComponent adapter). Updated the stale 'how to add a model' steps to the
post-redesign structure (v2/recipes/, torch_backend.py, SamplingDefaults). CPU mini 211 pass, 2 skip.
Addresses the adapter-setup pains: collapses the torch_cuda(trampolines)/torch_adapters/torch_ltx2 split
+ 11 near-identical adapter classes + 6 build_torch_* builders into:
- v2/platform/backends/torch_backend.py: a TorchComponent base centralizing .to/.eval, the numpy<->torch
marshalling (ONE place), the set_forward_context wrap, and the weight surface; thin per-model subclasses
(WanDiT/LTX2DiT/WanVAE/LTX2VAE/T5Encoder/Gemma/LTX2Upsampler/LTX2AudioVAE/LTX2Vocoder) carrying only
forward semantics; and ONE build_component(spec) dispatching by spec.kind via _MAKERS.
- torch_cuda.py: registers that single generic builder for all 6 cuda kinds (no per-kind trampolines).
- v2 owns its namespace via re-export STUBS (facade, marked '# STUB'): v2/forward_context, v2/fastvideo_args,
v2/distributed, v2/loader (the load_component seam), v2/api, v2/models/{dits,audio,upsamplers}/*. All v2
code imports v2.*; 'from fastvideo' now lives ONLY in those 8 stub files -> a future per-module vendored
cutover swaps a stub body, no caller changes. No divergence (stubs run fastvideo's live code).
- Deleted torch_adapters.py + torch_ltx2.py.
Verified: CPU mini 210 pass/2 skip; lazy invariant (platform load imports no torch); GPU bit-parity LTX-2.3
T2VS (audio std 0.04304, identical to pre-redesign) + Wan2.1 (video std 31.94, motion 5.997).
Frees v2/models/ to become the vendored-architecture namespace that mirrors fastvideo/models
(part of making v2 self-contained / able to replace fastvideo). The v2 recipe layer (per-family
card.py/loop.py/program.py + common.py + the build_*_engine re-exports) is the recipe, not the
architectures, so it moves to v2/recipes/. The pure-numpy toy/parity implementations (ToyDiT etc.)
move from v2/models/backend.py to v2/platform/backends/toy.py (alongside cpu.py/accel.py/torch_*).
Mechanical: all imports are absolute, so v2.models.<x> -> v2.recipes.<x> and v2.models.backend ->
v2.platform.backends.toy across v2/ + examples/ (89 files, 178 refs). No behavior change.
CPU mini green (202 passed, 2 skipped); all v2 files compile.
Mechanical conversion of every relative import under v2/ to an absolute v2.* path
(from .x / ..x / ...x -> from v2.<pkg>.x) so imports are unambiguous, grep-able, and
stable when code is copied/moved between entrypoints (VideoGenerator, CLI, server).
Surgical prefix-only rewrite: only the 'from <dots><module>' prefix changed — import
names, parentheses, multi-line formatting, comments, and ordering are byte-for-byte
preserved (no collapsing, no reorder, no unrelated reformatting).
- 471 imports across 125 files; v2/tests/ was already absolute (untouched).
- Validated: all 125 files compile, every 'from v2.* import' target resolves to a real
module/package, zero relative imports remain (full sweep), CPU mini suite green
(202 passed, 2 skipped).
Makes the 'Full A/V' LTX-2.3 deliverable reachable from the user-facing entrypoint, not just the
engine. A model advertising TEXT_TO_VIDEO_SOUND (LTX-2.3) now auto-issues a T2VS request, so generate()
/ generate_video() return BOTH modalities in one joint pass:
- VideoGenerator stores the resident instance + a supports_av flag (from card.capabilities); generate()
gains want_audio (None=auto-by-capability, True/False to force) and routes T2V vs T2VS+{video,audio}.
- _result saves the stereo waveform as a sibling .wav at the vocoder's REAL rate (24000) — read off the
built audio_vae adapter (TorchLTX2AudioVAE.sample_rate = Vocoder.output_sample_rate), since the
AudioArtifact default rate is a placeholder. Populates GenerationResult.audio/.audio_sample_rate and
extra['audio_path']. scipy IEEE-float WAV; [channels,samples] auto-transposed.
- Rewrote v2_basic_ltx2_3_distilled.py: registry routes to build_ltx2_3_card (its own joint T2VS A/V
card, not the LTX-2 base/2-stage card); the example prints both the mp4 and the wav.
- GPU-verified via the convenience API: ev.mp4 + ev.wav (24000 Hz, stereo 61920x2, nonzero, std 0.043).
CPU mini green (202 passed, 2 skipped); engine/program/toy paths untouched (test_ltx2_av pins 44100).
The full joint text->video+audio LTX-2.3 path now generates on the real 18.99B model:
- GPU audio components: build_torch_audio_vae (AudioDecoderLoader -> LTX2AudioDecoder, chains the
vocoder) + build_torch_vocoder (VocoderLoader -> LTX2Vocoder); registered the 'audio_vae'/'vocoder'
cuda component kinds; stamped their checkpoint subfolders (_WAN21_SUBFOLDERS).
- Fix: TorchGemma.encode_av must pass output_hidden_states=True — the 2.3 connector's SEPARATE audio
projection lives in hidden_states[0] only then (gemma.py:703); without it the audio text fell back to
the video embedding (4096 vs 2048 -> audio cross-attn shape mismatch).
- GPU-verified T2VS: video (3,33,256,384, std 0.68) + audio (stereo 2x61920 @24kHz, nonzero, std 0.059).
- test_torch_backend: cuda component kinds now include audio_vae + vocoder. CPU suite green (202+2).
Makes LTX-2.3 a first-class, faithful card (was wrongly merged into the single-stage base):
- LTX23DenoiseLoop (loop.py): single-pass joint A/V denoise — one DiT forward per step cross-attends
video<->audio via the adapter's (v_vel,a_vel) return; full-res video latent + a [8,T,16] audio latent;
distilled few-step schedule (BASE_SIGMAS). Video-only when no audio requested.
- build_ltx2_3_card (model_id 'ltx2.3-distilled' — the name now correctly names the REAL 2.3): 5
components incl. audio_vae (AudioDecoder) + vocoder (required_for t2vs, optional_for t2v); caps
T2V + T2VS. build_ltx2_3_program: dual-connector text-encode -> joint denoise -> video + audio decode.
- registry.py routes FastVideo/LTX-2.3-Distilled-Diffusers -> this card (split from the base entry).
- Toy support: ToyTextEncoder.encode_av (separate video/audio text), ToyDiT joint A/V (audio now a
keyword-only arg so positional callers like the talker are unaffected), channel-agnostic
ToyAudioVAE (np.resize identity for the existing 2-stage T2VS).
- CPU-verified: toy T2VS -> video+audio (8 steps), T2V -> video-only; CPU suite green (202+2).
GPU audio-VAE/vocoder loaders + the real-T2VS GPU verify are the next step.
Foundation for the LTX-2.3 T2VS port (card/loop/program wiring + GPU verify to follow):
- TorchLTX2DiT.__call__ gains an optional joint audio path: pass audio_latent[8,T,16] + audio_text and it
feeds audio_hidden_states/audio_encoder_hidden_states/audio_timestep/audio_sigma in ONE forward
(LTX-2.3 cross-attends video<->audio) and returns (video_velocity, audio_velocity). Video-only call is
byte-for-byte unchanged (audio_latent=None).
- TorchGemma.encode_av returns the SEPARATE (video_text, audio_text) projections from the 2.3 connector
(video=last_hidden_state, audio=hidden_states[0]); 2.0 returns them equal.
- TorchLTX2AudioVAE (AudioDecoder -> Vocoder -> waveform@24kHz) + TorchLTX2Vocoder wrapper.
Additive + backward-compatible; CPU suite unaffected (adapters are GPU-lazy).
Per review: dispatch should be a directly-mapped HF-string -> card registry (like fastvideo), shared
by every entrypoint (VideoGenerator + a future CLI / server), not buried in the generator.
- New v2/registry.py mirrors fastvideo's fastvideo/registry.py hybrid resolution: (1) exact HF repo id
in an explicit ModelEntry registry [PRIMARY — correct per-model card/capabilities, and the only way to
split same-architecture capability variants like Wan2.1 T2V vs the i2v 'InP' 1.3B], (2) short repo-name
match, (3) architecture inference [FALLBACK — local paths / unregistered repos]. resolve(model_path[,
root]) is the single shared entry point.
- video_generator.py: moved _read_arch_signature/_select_builders into the registry; from_config now
calls resolve() — registered ids resolve with no config read, else arch inference on a cheap *.json
snapshot. Reconciles the earlier 'no brittle table' refactor with the 'map hf string -> card' ask: one
clean registry + a fallback, not three coupled structures.
- Verified: registry resolves exact-id / short-name / arch-fallback / unregistered correctly; CPU suite
green (202+2); wan21 GPU smoke generates via the new resolve path.
Addresses the 'how is ltx2 separate from ltx2.3' confusion + a wrong capability:
- LTX-2 cards renamed by ARCHITECTURE (the version labels did not map to it): build_ltx2_card model_id
'ltx2.3-distilled' -> 'ltx2-2stage-distilled' (two-stage base->upsample->refine; serves the
upsampler-having FastVideo/LTX2-Distilled-Diffusers); build_ltx2_base_card 'ltx2.base' ->
'ltx2-single-stage' (one loop; serves Davids048 base + the single-stage FastVideo/LTX-2.3-Distilled,
which has NO spatial_upsampler). Dispatch already splits on has_spatial_upsampler. Updated the
model-id refs in the mini's tests/examples.
- Wan2.1 base is T2V-only: dropped the wrong Capability.IMAGE_TO_VIDEO + narrowed components'
required_for to {t2v} (i2v is the separate InP variant; v2 has no i2v path yet). build_wan21_card is
shared by wan21 + wan2.2-ti2v; the A14B card was already T2V-only.
- CPU suite green (202 passed, 2 skipped).
Within the bounded Wan+LTX-2 scope:
- Wan2.2-A14B MoE now GENERATES on a single 80GB GPU via CPU offload (TorchWanDiT offload_group):
the two 14B experts live on CPU and only the active one is swapped onto the GPU at the boundary-
timestep transition (a single swap, not per-step). GPU-verified: 60GB peak (vs 79GB OOM), produced
wan22_a14b_lion.mp4 (17x480x832, std 54.7, motion 8.36). Single-expert Wan stays resident.
- LTX-2 base (single-stage) port: build_ltx2_base_card + build_ltx2_base_program reuse the LTX-2
adapters at FULL latent res with a request-driven many-step flow-match (LTX2DenoiseLoop full_res/
request_steps/base_flow_sigmas; distilled base/refine path preserved via False defaults). The SAME
single-stage card serves LTX-2.3-Distilled (also single-stage: no spatial_upsampler) — dispatched by
the new has_spatial_upsampler discriminator in _select_builders. v2_basic_ltx2.py added; VideoGenerator
gains shutdown() for API parity.
- Fixes: from_config 'os' scoping (shadowed module import); LTX-2 upsampler per_channel_statistics
source (the AE's .decoder, not the top-level module).
Verification status: A14B offload, the upsampler, and the arch-dispatch refactor were GPU-verified
earlier this session. LTX-2 base/2.3 are CPU-verified (cards/programs build, dispatch routes, schedule
correct); their GPU smoke tests were pending when the box was rescheduled aarch64->x86 mid-session,
which broke the aarch64 venv (numpy/torch unrunnable on x86) — GPU re-verify blocked on the env.
Two reviewer asks + the next Wan port, all within the bounded Wan+LTX-2 scope:
1. Architecture-driven dispatch (replaces the HF-id table + substring fallback): from_config reads the
checkpoint's pipeline/transformer/VAE class names (+ z_dim, transformer_2) and picks the v2 card via
_select_builders — mirroring fastvideo's get_pipeline_config_cls_from_name. Resolves local paths /
renamed repos / new distilled variants of a known arch with no table edits, and cleanly REJECTS
FastWan (detected by WanDMDPipeline) with a precise message instead of a confusing load crash.
2. Real LTX-2 spatial upsampler (was a nearest-neighbor np.repeat stand-in): new 'upsampler' component
kind -> TorchLTX2Upsampler wraps the real LTX2LatentUpsampler and applies the repo's upsample_video
(un_normalize via the VAE decoder's per_channel_statistics -> learned 2x upsample -> normalize). CPU
keeps ToyUpsampler (np.repeat) via the factory terminal, so the program calls
component('spatial_upsampler').upsample(...) on both backends with no device branch. GPU-verified:
9x512x768, std 70.5, motion 6.51.
3. Wan2.2-T2V-A14B MoE card (build_wan22_a14b_card): two WanTransformer3DModel experts +
BoundaryTimestepRouting @0.875. GPU-verified that both experts denoise; OOMs in VAE decode on one
80GB GPU (~70GB resident) — upstream offloads the DiT for MoE; documented as offload-blocked.
CPU suite green (202 passed, 2 skipped). FastWan root-caused (non-strict load of VSA gate_compress +
VSA not built); roadmap (V2_PORTING_STATUS.md) updated with the bounded scope + per-model status.
- VideoGenerator gains the convenience surface most basic examples use: from_pretrained(model,
num_gpus/*_cpu_offload/...) and generate_video(prompt, sampling_param=, **kwargs) -> result, on top
of the typed from_config/generate. Accepts SamplingParam.
- v2 examples matching the upstream convenience-API examples for the verified models: v2_basic.py
(Wan2.1) and v2_basic_self_forcing_causal.py (SF-causal).
- V2_PORTING_STATUS.md: honest per-family roadmap. Working: wan21, wan_causal, ltx2-distilled. Each
further model needs per-model work (FastWan: WanDMD to_gate_compress param mapping; TurboWan: RCM
consistency sampler; Wan2.2: MoE card; LTX2 base/i2v; new families: new cards/adapters; gated Flux2 /
local GEN3C / audio StableAudio / interactive MatrixGame blocked in this env).
Official FastVideo/LTX2-Distilled-Diffusers. New torch_ltx2.py adapters (build_torch_* dispatch on
class):
- TorchLTX2DiT: patchify-internal; per-token timestep ones(B,tok,1)*sigma (sigma direct) + per-sample
video_sigma; DiT predicts x0 so the adapter returns velocity=(x_t-x0)/sigma for the v2 flow-match step.
- TorchLTX2VAE: CausalVideoAutoencoder decode (internal per-channel un_normalize).
- TorchGemma: LTX2GemmaTextEncoderModel (Gemma + feature-extractor + connectors) -> last_hidden_state.
- ltx2 loop: real 128-ch latent geometry on cuda (32x spatial / 8x temporal; half-res base, 2x upsample).
e2e two-stage (8+3 steps) -> coherent, high-quality video (surfers at sunset), (3,9,512,768). NOTE:
still uses the v2 program's np.repeat upsampler between stages (the refine regenerates from noise so
output is faithful-quality); real LTX2LatentUpsampler swap-in is a follow-up.
Official SF checkpoint wlsaidhi/SFWan2.1-T2V-1.3B-Diffusers (reuses TorchWanVAE + TorchT5Encoder).
- TorchWanDiT detects the causal transformer: ignores the chunk_rollout loop's latent `context` (the
real model conditions across chunks via an internal kv_cache, not a forward arg) -> dispatches to
full-attention _forward_train, and passes a per-latent-frame timestep [B, num_frames] (the causal
block asserts a per-frame temb), uniform per chunk.
- chunk_rollout real geometry on cuda (16ch; chunk_size latent frames; 8x spatial).
e2e: chunk_rollout over the SF student -> coherent video (cat in a garden), (3,21,480,832). Fidelity
gap (artifacts): the v2 loop's per-chunk few-step sampling != the official kv-cache streaming + SF
schedule (a follow-up).
The cuda-availability probe imports torch by design; the no-torch-import invariant is only
verifiable when torch is absent. skipif torch installed -> green on GPU box (202 passed, 2 skipped),
still enforced in torchless CI.
- latent_shape(req, model): real Wan geometry (16ch; (T-1)//4+1, H/8, W/8) on the cuda backend;
toy stand-in stays on accel/cpu.
- stamp_wan21_checkpoints(card, model_root): map a root (local dir or HF id) onto the 3 components'
ComponentSpec.checkpoint; build_wan21_card(checkpoint_root=...) optional.
Take the written-not-run torch adapters to runs-and-generates on 1x H100 (aarch64):
- A: FastVideoArgs.from_kwargs(model_path=root) builds the real pipeline_config; single-GPU dist
init; load each component from its subfolder; tokenizer from the sibling <root>/tokenizer.
- B/C: DiT forward wrapped in set_forward_context(attn_metadata=None) (SDPA dense path);
timestep=sigma*1000 + bare-velocity output confirmed.
- D: VAE decode denormalizes z*std+mean; removed the double-mean (it re-added
shift_factor==latents_mean) that washed out the video.
- E: UMT5 from config; text embeds zero-padded to text_len (Wan t5_postprocess_text) - the fix
that took output from a dark blur to a coherent prompt-matching scene.
- Components run at native precision (DiT bf16, VAE/text fp32); checkpoint check before dist init.
Orientation + process doc for an agent on a GPU branch: the 6 commits already
landed, the files to touch, the gating tasks (Risk A FastVideoArgs/checkpoint),
the verification bar (CPU suite stays 204; GPU generation matches a reference via
the SSIM harness), commit/push rules (no Claude co-author; don't rewrite history;
wandb token referenced not embedded), and the gotchas. Points to
GPU_BRINGUP.md for the detailed checklist + risk table.
Adversarial cross-check of the written-not-run torch adapters against the real
fastvideo source confirmed the interface contracts (DiT returns bare velocity
tensor; timestep=sigma*1000; encode().mode() + bare decode; .last_hidden_state;
no fused solver kernel) but caught a wrong construction layer. Fixed in code:
- Construction: WanTransformer3DModel / AutoencoderKLWan have NO from_pretrained.
Replace it with the real FastVideo loaders (TransformerLoader / VAELoader /
TextEncoderLoader + TokenizerLoader, each load(model_path, fastvideo_args)).
The loader resolves the class from the checkpoint config — so UMT5-vs-T5 is
chosen correctly instead of hardcoded (was BLOCKER #1/#4/#5).
- Text encoder: wrap the forward in set_forward_context(...) — the (U)MT5
attention reads global state via get_forward_context(); a bare call mis-encodes
(was BLOCKER #2). Drop the wrong padding="max_length".
- VAE: apply latent normalization the DiT expects — (z-mean)*inv_std on encode,
inverse on decode, with latents_std stored as its reciprocal; shift_factor
before decode (was BLOCKER #3). Skipping it yields washed-out video, not error.
The remaining unknowns are genuinely box-dependent (FastVideoArgs fields,
shift_factor placement, exact tokenizer kwargs, FSDP) — GPU_BRINGUP.md reconciled
to mark what's now fixed-in-code vs what still needs the box. 204 CPU tests pass.
Implement the GPU backend the substrate was built for: Platform.detect() ->
cuda resolves real torch adapters + torch solver ops instead of the numpy
rungs, with the existing loops/policies/scheduler/training unchanged.
WRITTEN-NOT-RUN: this environment has no GPU/torch, so the torch code is
grounded in the verbatim real fastvideo APIs (DiT forward signature confirmed
from source) but cannot be executed/verified here. It is gated available=False
(CPU mini stays green; importing the backends never imports torch), with every
on-box confirm point marked `# BRINGUP` and an ordered checklist in
platform/backends/GPU_BRINGUP.md.
- torch_adapters.py: TorchWanDiT / TorchWanVAE / TorchT5Encoder wrap the real
module named by each card's load_id and bridge it to the mini's duck-typed
surface (numpy<->torch at the boundary; loop math stays numpy fp32). DiT
weight-surface (copy_from/blend_from/clone) for serving sync; mse_grad_step
raises (GPU training is a separate workstream).
- torch_kernels.py: flow_match_step / flow_sde_step as plain torch elementwise.
Grounded conclusion from the kernel audit: fastvideo-kernel ships NO fused
solver kernel (only attention/norm/quant primitives), so the cuda solver is
torch, registered at arch generic with an honest source string.
- torch_cuda.py: rewritten as lazy trampolines (torch imported only inside
builder/kernel bodies). Adds the missing vae + text_encoder cuda components
(they'd otherwise silently fall back to the toy) and corrects the dishonest
"fastvideo-kernel:flow_*" labels.
- ComponentSpec.checkpoint: the weights source for the torch adapter (risk A;
the one field the cards didn't carry). Empty on the CPU toys.
- 7 CPU-verifiable wiring tests: honest registration/sources, torch-free import,
cuda-resolves-real-cells-not-toy, build-fails-loudly-without-torch.
204 tests pass (CPU). The torch path needs a GPU box to verify (GPU_BRINGUP.md).
Close the loudest deferred gap from the cudagraph audit: the capturable step
now binds its I/O to address-stable static buffers (modeling real CUDA static
I/O buffers), instead of allocating fresh arrays per call.
- StaticWorkspace: address-stable buffers allocated once per capture key; bind()
copies the current step's inputs in place via np.copyto, which RAISES on a
shape/dtype mismatch — turning the weak peak_activation_bytes proxy into a real
key-soundness backstop (a step whose shape doesn't fit can't replay an
incompatible graph). Output written into a static buffer too.
- WorkPlan.graph_fn / graph_inputs: a capturable step exposes its deterministic
op-structure as graph_fn(model, workspace) reading EVERY per-step input (latent,
sigmas, conditioning, scale) from the workspace — never from closure over
per-step data — plus the dict of current values. Loops without both stay on the
eager path. wan21 factors a shared _velocity() so graph_fn and the eager run
stay bit-identical.
- Capturer dispatch captures/replays via graph_fn against the keyed workspace;
the workspace is shared per key on the instance. Correct under the engine's
synchronous step execution (bind+graph_fn atomic per dispatch, output returned
as a copy) — proven by the batch-of-N interleave gate running two same-key
requests through the shared workspace bit-identically. A concurrent/multi-stream
executor would need a per-stream pool (documented).
- 2 new tests (no-static-form eager-break, static-buffer shape-mismatch raises);
workspace collision test split into bytes-proxy vs shape-backstop.
197 tests pass.
Wire the capture/replay lifecycle into the driven-loop step boundary — the
other half of Path A (hand-fused kernels behind the registry + piecewise
cudagraphs, no compiler). Models and tests the correctness-critical control
logic; replay re-runs the current step thunk (CPU models the lifecycle, not the
GPU speedup).
- GraphCapturer on the instance (cross-request cache), wired into
RuntimeLoopContext.execute and gated by LoopSpec.graph_capture ==
"breakable_cudagraph". Capture key = (device, arch, loop, shape_sig,
resident-weight-versions, graph_key). Eager-break for non-capturable (SDE) and
interceptor-overridden steps. Never executes a stored thunk, so interleaved
requests can't smear state.
- WorkPlan.capturable / graph_key. wan21 sets capturable=not sde and folds
compute dtype (shape_sig.dtype) + CFG branch set + expert + scheduler-precision
into the key — closing a cross-precision key-collision corruption path a
non-fp32 build would otherwise hit on a real GPU (audit finding).
- Version-in-key auto-invalidation + real eviction: set_weights_version evicts
the synced component's graphs (duck-typed, so card/ imports no runtime),
preventing a GPU graph leak across FlowGRPO's per-iteration syncs.
- register_kernel gains the workspace_bytes capture-safety contract (declared in
the matrix; cuda cells declare real scratch, numpy reference is 0).
- 11 tests: capture-once/replay-many, eager-break (SDE + override), capture ≡
pure-eager bit-identical, recapture on shape + weight-version change with
eviction, eager-loop gating, accel-backend capture, and capturer unit tests
(key discrimination, eager-break, workspace-collision safety net, eviction).
Honestly deferred (bite a real GPU, not the CPU tests): the static-buffer
refactor of the step body, admission budgeting of capture cost (GRAPH_CAPTURE),
and per-card opt-in beyond wan21 — all documented in cudagraph.py + README.
195 tests pass.
SolitaryThinker