SolitaryThinker and Claude Opus 4.8
3ab66290dc
[bugfix] QAD 5090: emit sm_120a in build.sh so attn_qat_infer kernels build
...
build.sh auto-detected Blackwell (sm_120) and exported TORCH_CUDA_ARCH_LIST=12.0 without the arch-conditional 'a' suffix. CMake's AUTO gate for the attn_qat_infer (modified SageAttention3 FP4) kernels only matches 12.0a/120a/sm_120a, so fp4attn_cuda/fp4quant_cuda were silently skipped and the ATTN_QAT_INFER backend fell back to Flash Attention at runtime. Exporting the env var also bypassed CMake's local-GPU fallback that would otherwise have enabled them.
Mirror the existing 9.0 -> 9.0a Hopper handling for 12.0 -> 12.0a.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com >
2026-06-23 07:29:53 +00:00
7fa0fed781
[feat] QAD 5090: env-gate attention torch.compile via FASTVIDEO_DISABLE_ATTENTION_COMPILE
...
DistributedAttention.forward (and the VSA subclass) are hard-decorated with
@torch.compiler.disable, which keeps attention out of the surrounding
torch.compile graph unconditionally. That blocks the inference compile path
even after the FP4 linear and SageAttention3 graph-break fixes land, since
the attention forward itself can never be traced.
Make the disable conditional on FASTVIDEO_DISABLE_ATTENTION_COMPILE:
- unset / "1" / "true" (default): keep torch.compiler.disable — current behavior
- "0" / "false" / "no" / "off": drop it so attention can fold into the graph
The env var is read at import time (decorators are applied at class
definition), which is the right granularity for the multiproc spawn path:
each worker re-imports and inherits the parent's env.
Co-authored-by: Loay Rashid <42599591+loaydatrain@users.noreply.github.com >
Co-authored-by: Kaiqin Kong <k1kong@ucsd.edu >
Co-authored-by: William Lin <SolitaryThinker@users.noreply.github.com >
2026-06-23 06:44:05 +00:00
loaydatrain
9096310b5c
precommit stuff
2026-06-23 06:44:05 +00:00
loaydatrain
fce6ed516d
Adding TAEHV script, sage_attn3 more torch compile friendly, weight popping+torch compile+single level quant changes to nvfp4_qat_config
2026-06-23 06:44:05 +00:00
Kevin Lin
82ed9fe58d
[feat] QAD 5090: FP8 linear layer inference ( #1465 )
2026-06-22 18:25:06 -07:00
Satyam Srivastava
3d8cc4f0a0
[bugfix] Fix performance component timing extraction ( #1473 )
2026-06-22 13:05:35 -07:00
Satyam Srivastava
0557f7a7d9
[ci] Add performance dashboard metadata and visualizations ( #1470 )
2026-06-19 14:27:01 -07:00
dc66cd97ef
[feat] QAD 5090: FP8 QAT linear training (14/12) ( #1464 )
...
Co-authored-by: William Lin <SolitaryThinker@users.noreply.github.com >
Co-authored-by: Loay Rashid <42599591+loaydatrain@users.noreply.github.com >
Co-authored-by: Kaiqin Kong <k1kong@ucsd.edu >
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
2026-06-19 01:03:51 -07:00
6da206e196
[feat] QAD 5090: FP4 QAT linear STE for training (13/12) ( #1463 )
...
Co-authored-by: William Lin <SolitaryThinker@users.noreply.github.com >
Co-authored-by: Loay Rashid <42599591+loaydatrain@users.noreply.github.com >
Co-authored-by: Kaiqin Kong <k1kong@ucsd.edu >
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-06-18 15:55:47 -07:00
sumyyyyy
87f98c9b8b
[kernel] Add varlen support for block-sparse attention ( #1319 )
2026-06-18 01:06:00 +00:00
e60601df7f
[feat] QAD 5090: QAT training recipe — finetune + DMD distillation (12/12) ( #1462 )
...
Co-authored-by: William Lin <SolitaryThinker@users.noreply.github.com >
Co-authored-by: Loay Rashid <42599591+loaydatrain@users.noreply.github.com >
Co-authored-by: Kaiqin Kong <k1kong@ucsd.edu >
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-06-17 15:18:52 -07:00
eed9c4bfbf
[kernel] QAD 5090: Add Attn-QAT training Triton kernels (11/12) ( #1460 )
...
Co-authored-by: William Lin <SolitaryThinker@users.noreply.github.com >
Co-authored-by: Loay Rashid <42599591+loaydatrain@users.noreply.github.com >
Co-authored-by: Kaiqin Kong <k1kong@ucsd.edu >
2026-06-17 13:35:12 -07:00
1dee77f4a4
[feat] QAD 5090: Wire the Attn-QAT training attention backend (10/12) ( #1459 )
...
Co-authored-by: William Lin <SolitaryThinker@users.noreply.github.com >
Co-authored-by: Loay Rashid <42599591+loaydatrain@users.noreply.github.com >
Co-authored-by: Kaiqin Kong <k1kong@ucsd.edu >
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
Co-authored-by: alexzms <26690162+alexzms@users.noreply.github.com >
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-06-17 13:32:56 -07:00
Satyam Srivastava
b80148819c
[ci] Add performance dashboard visualisation scripts ( #1469 )
2026-06-16 23:29:29 -07:00
c3b971488e
[docs] QAD 5090: Add NVFP4 + Attn-QAT inference example and how-to (9/12) ( #1458 )
...
Co-authored-by: William Lin <SolitaryThinker@users.noreply.github.com >
Co-authored-by: Loay Rashid <42599591+loaydatrain@users.noreply.github.com >
Co-authored-by: Kaiqin Kong <k1kong@ucsd.edu >
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
Co-authored-by: alexzms <26690162+alexzms@users.noreply.github.com >
2026-06-16 14:36:50 -07:00
88e753f281
[feat] QAD 5090: Wire the Attn-QAT inference attention backend (8/12) ( #1457 )
...
Co-authored-by: William Lin <SolitaryThinker@users.noreply.github.com >
Co-authored-by: Loay Rashid <42599591+loaydatrain@users.noreply.github.com >
Co-authored-by: Kaiqin Kong <k1kong@ucsd.edu >
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
Co-authored-by: alexzms <26690162+alexzms@users.noreply.github.com >
2026-06-16 13:12:33 -07:00
77832059cc
[kernel] QAD 5090: Add modified SageAttention3 FP4 inference kernels (7/12) ( #1455 )
...
Co-authored-by: William Lin <SolitaryThinker@users.noreply.github.com >
Co-authored-by: Loay Rashid <42599591+loaydatrain@users.noreply.github.com >
Co-authored-by: Kaiqin Kong <k1kong@ucsd.edu >
Co-authored-by: Edenzzzz <wtan45@wisc.edu >
2026-06-16 12:02:30 -07:00
alexzms and mergify[bot]
633d393568
[ci] layer-0 grad-norm regression for per-method training tests (5a-ii) ( #1396 )
...
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-06-12 04:45:07 +00:00
Junda Su and Peiyuan Zhang
5854aec2ce
[feat] Add Wan RL DiffusionNFT training ( #1450 )
...
Co-authored-by: Peiyuan Zhang <a1286225768@gmail.com >
2026-06-11 21:18:59 -07:00
Mook
30e45c2411
[bugfix] Classify new config/sampling fields in schema parity inventory ( #1446 )
2026-06-10 13:38:36 -07:00
alexzms
2a4fe697a6
[docs] LTX-2.3 distilled i2v: typed-API example (from_config + generate) ( #1448 )
2026-06-10 10:58:14 -07:00
alexzms
921db7479d
[perf] LTX-2.3 distilled i2v: drop max-autotune from compile kwargs ( #1445 )
2026-06-10 10:06:41 -07:00
Aryan Kumar and Aryan Kumar
7f539424cb
[feat]: add Lucy Edit inference scaffold ( #1363 )
...
Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com >
2026-06-09 15:21:43 -07:00
19a838f54f
[bugfix]: release VSA tile cache during training ( #1434 )
...
Co-authored-by: Aryan Kumar <aryan5v@users.noreply.github.com >
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-06-09 15:01:22 -07:00
d922ab2cbc
[model] Flux2 Klein Port ( #1349 )
...
Co-authored-by: Gnav3852 <63612880+Gnav3852@users.noreply.github.com >
Co-authored-by: Mac Lee <macthecadillac@gmail.com >
2026-06-09 14:55:55 -07:00
Kaiqin Kong and mergify[bot]
9ea77d37f3
[bugfix] EMA shadow on resume and EMA under MoE path ( #1441 )
...
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-06-09 00:55:43 +00:00
2e35b0c6bd
[refactor]: linear/mlp FP4 path additions for Wan-2.1 (Attn-QAT 6/12) ( #1390 )
...
Co-authored-by: Peiyuan Zhang <a1286225768@gmail.com >
Co-authored-by: Matthew Noto <notomatthew31@gmail.com >
2026-06-08 14:51:57 -07:00
William Lin
1c627a3f98
[bugfix]: build fastvideo-kernel on GPU-less Docker runners ( #1437 )
2026-06-06 17:52:45 -07:00
Kaiqin Kong
a931efe33a
[bugfix] EMA in distillation pipeline ( #1440 )
2026-06-06 17:38:03 -07:00
William Lin
041e5e9029
[bugfix]: unblock PyPI publish (flash-attn-cute direct dep) ( #1436 )
2026-06-05 10:07:12 -07:00
Kaiqin Kong and mergify[bot]
efcc245c2e
[feat] VLM as judge for WM ( #1429 )
...
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-06-05 04:25:54 +00:00
alexzms and mergify[bot]
922e7e0813
[docs] LTX-2.3 distilled i2v example with compile + timing breakdown ( #1430 )
...
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-06-05 01:26:50 +00:00
William Lin
c62a8514b0
[chore]: release v0.2.0 ( #1432 )
2026-06-04 14:22:49 -07:00
William Lin
3eb8081801
[chore]: unpin runtime deps in pyproject.toml ( #1431 )
2026-06-04 13:46:49 -07:00
Raghav K
3505d09564
[bugfix] tests: include ltx2_3_base in expected LTX2 preset set ( #1427 ) ( #1428 )
2026-06-04 12:17:18 -07:00
Shao Duan and SolitaryThinker
570607945c
[feat] dreamverse: sequence parallelism for serving ( #1424 )
...
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
2026-06-01 23:13:49 -07:00
Kaiqin Kong and SolitaryThinker
d3a821cdcf
[feat] LoRA controls and integration for Dreamverse ( #1420 )
...
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
2026-06-01 18:54:11 -07:00
Kaiqin Kong
3f24578139
[bugfix] LTX2: honor video_position_offset_sec in the DiT ( #1422 )
2026-06-01 16:59:15 -07:00
Kevin Lin
89fcf08378
[bugfix] Fix STFT dtype mismatch ( #1419 )
2026-05-31 20:56:07 -07:00
c2930b2aa1
[ci] Add additional Dreamverse UI tests ( #1417 )
...
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-05-31 19:53:29 -07:00
KUAN-HAO HUANG and SolitaryThinker
019239690b
[perf] Add Adaptive Guidance (CFG gating) for stale-uncond reuse ( #1372 )
...
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
2026-05-30 13:58:55 -07:00
alexzms
d6119c1f82
[bugfix]: dreamverse modal bypasses ENTRYPOINT — set ffmpeg env + key check ( #1413 )
2026-05-29 16:50:01 -07:00
alexzms
84214c80bb
[feat] LTX-2.3 audio: BWE vocoder path ( #1398 )
2026-05-29 16:43:26 -07:00
alexzms and mergify[bot]
c2d7143c72
[feat] LTX-2.3 transformer support (config-gated extension of LTX-2) ( #1397 )
...
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
2026-05-29 16:42:37 -07:00
Kaiqin Kong and SolitaryThinker
afdb6fbfa5
[feat] Add MatrixGame3.0 ( #1201 )
...
Co-authored-by: SolitaryThinker <wlsaidhi@gmail.com >
2026-05-27 21:07:01 +00:00
alexzms
ba4c02d883
[docs]: highlight Dreamverse deployment paths + add Server B200 (SSH) guide ( #1409 )
2026-05-27 09:30:58 -07:00
Junda Su
2c137931f3
[bugfix] Fix Dreamverse Modal compile warmup latency ( #1394 )
2026-05-26 17:58:38 -07:00
Shao Duan
36682797a0
[feat] eval: input ergonomics + Evaluator features + bug fixes ( #1392 )
2026-05-26 13:59:45 -07:00
William Lin
0ef1357a77
[docs]: surface activation-trace utility in add-model skills ( #1399 )
2026-05-26 13:45:12 -07:00
William Lin
a75d19786a
[docs]: Wire activation trace into mkdocs nav + perf/troubleshooting ( #1304 )
2026-05-26 13:14:36 -07:00
