SolitaryThinker
4d8713f6e4
[bugfix]: bound WanCausalModel num_frames_per_block override to <= 3
...
The constructor override only checked >= 1, bypassing the <= 3 limit the
config path enforces in CausalWanTransformer3DModel. Apply the same upper
bound with a clear error message.
2026-07-05 15:01:25 -07:00
SolitaryThinker
fda43fc610
[bugfix]: refuse teacher forcing with a local attention window
...
_prepare_teacher_forcing_mask silently ignored local_attn_size while the
block-wise causal mask honors it, so a configured attention window was
dropped on the teacher-forcing path. Raise NotImplementedError instead of
training with a mask that contradicts the config.
2026-07-05 15:01:09 -07:00
SolitaryThinker
5a8329ddf6
[bugfix]: set causal-CD timesteps before the teacher CFG forwards
...
training_batch.timesteps was assigned t_pf only after the two teacher
CFG passes, so their set_forward_context(current_timestep=...) carried
the stale random timesteps from prepare_batch — wrong VSA sparsity
gating when attn_kind == "vsa". Assign before any forward.
2026-07-05 15:00:47 -07:00
SolitaryThinker
fcb5b465c5
[bugfix]: checkpoint causal-CD EMA consistency target on save/resume
...
The base TrainingMethod.checkpoint_state() only persists trainable roles,
so CausalConsistencyDistillationMethod's EMA target (frozen but mutated by
_update_ema every step) was never saved; a resume reloaded it from
init_from, snapping the consistency target back to the base checkpoint.
Persist roles.ema.transformer via the same full-state DCP wrapper
DiffusionNFT already uses for its frozen 'old' role, moving _FullModelState
into fastvideo/train/utils/checkpoint.py so both methods share it.
2026-07-05 15:00:27 -07:00
H1yori233
4f13fb0fee
fix
2026-07-05 14:56:08 -07:00
H1yori233
d96ec99b4b
fix
2026-07-05 14:56:08 -07:00
H1yori233
2555f25cce
fix scheduler
2026-07-05 14:56:08 -07:00
H1yori233
fbe56fce8d
update scheduler and ema
2026-07-05 14:56:08 -07:00
H1yori233
5064bcdc47
cleanup
2026-07-05 14:56:08 -07:00
H1yori233
2aa824b615
cleanup
2026-07-05 14:56:08 -07:00
H1yori233
f934efc58d
add cf
2026-07-05 14:56:08 -07:00
William Lin
76b0550c15
[ci]: run pre-commit on fork PRs without manual approval ( #1555 )
2026-07-05 14:18:16 -07:00
William Lin
384c1e9493
[misc]: update reseed-performance-baseline skill for the hf_store move ( #1545 follow-up) ( #1553 )
2026-07-05 14:16:55 -07:00
William Lin
b1dbcc93f6
[misc]: reformat fastvideo/performance to the repo yapf config ( #1554 )
2026-07-05 14:16:20 -07:00
William Lin
b93833772e
[ci]: guard against test directories no CI lane collects ( #1552 )
2026-07-05 14:07:38 -07:00
Mac Lee
30b523edd6
[ci] Normalize performance stage component metrics ( #1475 ) ( #1550 )
2026-07-05 14:05:26 -07:00
Mac Lee and SolitaryThinker
6aab7f3832
[ci] cover Hunyuan 1.5 chat-list text preprocessing ( #1518 )
...
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
2026-07-05 12:05:25 -07:00
Mac Lee
9cd53fe5f8
[ci] Add metric-specific performance thresholds ( #1545 )
2026-07-05 12:04:33 -07:00
Mac Lee
6a32cf3a5e
[ci]: expose LoRA extraction slash command ( #1542 )
2026-07-05 06:45:31 -07:00
William Lin
98be9b3da2
[bugfix]: address the three remaining #1447 review findings ( #1549 )
2026-07-05 06:43:27 -07:00
Mac Lee
c53e85b767
[ci] Add v2 performance benchmark config identity fields ( #1544 )
2026-07-05 06:18:15 -07:00
William Lin
40a8bd2d3b
[bugfix]: skip ThunderKittens kernels on aarch64 and document the kernel build matrix ( #1548 )
2026-07-05 06:13:11 -07:00
Mac Lee and SolitaryThinker
31aa115611
[bugfix]: preserve FSDP hooks for RMSNorm qk norms ( #1513 )
...
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
2026-07-05 04:15:49 +00:00
zainnh and SolitaryThinker
98ac10a528
[infra] Auto-rebuild CUDA images when docker/Dockerfile changes on main ( #1526 )
...
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
2026-07-04 16:25:46 -07:00
William Lin
a5a6d171e5
[attn] Make FA4 explicit opt-in via FASTVIDEO_FA4 and delete the runtime fallback machinery ( #1540 )
2026-07-04 14:51:00 -07:00
William Lin
51ed1ea423
[build] Bump fastvideo-kernel pin to 0.3.2 ( #1541 )
2026-07-03 14:48:15 -07:00
Mac Lee
00ec3e7388
[bugfix]: compute VSA topk from padded blocks ( #1517 )
2026-07-03 14:16:51 -07:00
William Lin
0d626ef2d1
[kernel] Bump fastvideo-kernel pin to 0.3.1 and version the FA4 tile_mn port as 0.3.2 ( #1539 )
2026-07-03 14:11:08 -07:00
William Lin
b36d0ef085
[infra] Add DGX Spark and multi-architecture CUDA support ( #1447 )
2026-07-03 13:41:43 -07:00
William Lin
10c8c5df4d
[misc] reorg: relocate ui/ and performance_dashboard/ under apps/ ( #1537 )
2026-07-02 20:29:17 -07:00
William Lin
31e26abec4
[chore] release fastvideo-kernel 0.3.1 ( #1520 )
2026-06-30 12:25:27 -07:00
sumyyyyy and SolitaryThinker
9e83ba630c
[kernel] Extract VSA utility functions into fastvideo_kernel ( #1408 )
...
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
2026-06-30 12:19:25 -07:00
William Lin
fc02a9ce8e
[ci] aarch64 kernel wheel: build for Blackwell (sm_100a/sm_120a), not Hopper ( #1516 )
2026-06-29 16:10:42 -07:00
William Lin
3d36160fc4
[bugfix] -fsigned-char so ThunderKittens compiles on aarch64 (Grace Hopper) ( #1515 )
2026-06-29 14:05:38 -07:00
William Lin
a11ec43de2
[ci] build + publish aarch64 (Grace Hopper) kernel wheels alongside x86_64 ( #1514 )
2026-06-29 12:28:34 -07:00
William Lin
2656d6530c
[bugfix] compile FP4 (attn_qat_infer) kernels for sm_120a only via per-arch split ( #1508 )
2026-06-29 11:07:27 -07:00
Kevin Lin
e3f54e7169
[bugfix] Fix causal attention mask for Blackwell FP4 MMA column layout ( #1506 )
2026-06-28 20:36:40 -07:00
William Lin
4ba5681307
[bugfix] guard ThunderKittens Hopper kernels for non-sm_90a device passes ( #1507 )
2026-06-28 20:07:43 -07:00
alexzms
8c23c86994
[docs]: add raw-video preprocess script for the QAD MixKit recipe ( #1487 )
2026-06-28 18:01:13 -07:00
William Lin
78d606a3a7
[feat]: make VSA tile cache configurable for training ( #1444 )
2026-06-28 02:18:31 -07:00
William Lin
c4e108de78
[docs]: modernize documentation build ( #1503 )
2026-06-28 02:02:09 -07:00
Kaiqin Kong
16bf2eaf77
[feat] Relativistic RoPE re-indexing for long causal rollouts ( #1454 )
2026-06-28 01:43:00 -07:00
William Lin
1d15d974fa
[misc]: cleanup outdated or unneeded agent infrastructure ( #1504 )
2026-06-27 19:37:49 -07:00
alexzms
5e57868b76
[ci] train-framework model coverage: Cosmos/MatrixGame2 finetune grad-norm + Cosmos/LongCat/MatrixGame2 loading smokes ( #1497 )
2026-06-27 17:16:00 -07:00
William Lin
6be280c914
[bugfix]: fix docs build ( #1502 )
2026-06-27 13:00:09 -07:00
KyleNeverGivesUp and Claude Opus 4.8
3ccdec9798
[bugfix] Make enable_torch_compile_vae actually compile the Wan VAE ( #1498 )
...
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com >
2026-06-27 12:30:59 -07:00
Satyam Srivastava
8658f774f3
[docs] Restructure contributing CI/CD and testing docs ( #1501 )
2026-06-27 11:46:18 -07:00
Raghav K
4f3ad3f6df
[perf] Default Wan VAE decode to bf16 (lossless, faster) ( #1472 )
2026-06-26 12:42:29 -07:00
William Lin
b454aa56c3
[misc] fix pre-commit ( #1500 )
2026-06-26 12:39:28 -07:00
William Lin
f9b8e30ff3
Update README.md
2026-06-26 09:23:30 -07:00
SolitaryThinker