 Jaret Burkett
|
7690ea62c8
|
Add variable precision adam convrot optimizer with stochastic rounding and accumulation.
|
2026-09-06 10:51:35 -06:00 |
|
|
Jaret Burkett
|
bbb1fda457
|
Fix gradient accumulation scaling so grads scale to the same size as if it was one big batch.
|
2026-09-06 10:48:10 -06:00 |
|
|
Jaret Burkett (Ostris)
|
fea99cf5d0
|
Merge pull request #813 from Booyaka101/fix/dataloader-pin-memory
perf: enable pin_memory on CUDA DataLoaders for faster CPU->GPU transfer
|
2026-09-05 08:09:26 -06:00 |
|
|
Jaret Burkett (Ostris)
|
2a2ca26ee8
|
Merge branch 'main' into fix/dataloader-pin-memory
|
2026-09-05 08:08:39 -06:00 |
|
|
Jaret Burkett
|
b36bb3998a
|
Version bump
|
2026-09-03 08:44:44 -06:00 |
|
|
Jaret Burkett
|
90515c5396
|
Fix issue where we gave the wrong dtype hidden states
|
2026-09-03 08:43:15 -06:00 |
|
|
Jaret Burkett
|
ead3738773
|
Fixed connection cap for SSE in the ui
|
2026-09-02 17:08:12 -06:00 |
|
|
Jaret Burkett
|
ecb2bfbe5e
|
Fixed issue where some unmanaged modules could be lefs on cpu
|
2026-09-02 09:22:39 -06:00 |
|
|
Jaret Burkett
|
9d6a9a0803
|
Fixed embedding scale for offloading for ltx 2.3. Was a new bug added today.
|
2026-08-31 15:53:58 -06:00 |
|
|
Jaret Burkett
|
6940ebf533
|
Skip files where caching fails in a dataset instead of crashing.
|
2026-08-30 16:49:12 -06:00 |
|
|
Jaret Burkett
|
e98109f213
|
Version bump
|
2026-08-30 11:02:22 -06:00 |
|
|
Jaret Burkett
|
74ed5fddb0
|
Fix issue with ltx caching
|
2026-08-30 11:01:52 -06:00 |
|
|
Jaret Burkett
|
764b5064fb
|
Migrate to a new DTO for latents to carry more information that a normal tensor such as audio.
|
2026-08-30 10:30:21 -06:00 |
|
|
Jaret Burkett
|
2a69c1e7de
|
Add initial support for Minimax H3 VSA sparse attention
|
2026-08-30 08:24:43 -06:00 |
|
|
Jaret Burkett
|
be995185f5
|
Fixed issue with offloading with new model loader on nvfp4 weights.
|
2026-08-29 12:01:40 -06:00 |
|
|
Jaret Burkett
|
7380476b9c
|
Version bump
|
2026-08-29 11:24:16 -06:00 |
|
|
Jaret Burkett
|
683fe8afc0
|
Stability improvements to model offloading. Added D-OPSD bleed loss as well.
|
2026-08-29 11:23:47 -06:00 |
|
|
Jaret Burkett (Ostris)
|
64a20f51a6
|
Merge pull request #1025 from ostris/models/v2
Reworked the entire model loading to make everything testable, hot swappable, and uniform.
|
2026-08-29 09:06:19 -06:00 |
|
|
Jaret Burkett
|
bb55d38958
|
Version bump
|
2026-08-29 08:58:15 -06:00 |
|
|
Jaret Burkett
|
7195abc32b
|
Performance fixes and bug fixes
|
2026-08-28 22:21:09 -06:00 |
|
|
Jaret Burkett
|
5ddc5f8ca7
|
Optimize quantization path
|
2026-08-28 11:47:29 -06:00 |
|
|
Jaret Burkett
|
ce3df32101
|
Fix speed on wan conv patch
|
2026-08-28 10:54:26 -06:00 |
|
|
Jaret Burkett
|
92df289931
|
Fix quantization issues
|
2026-08-28 10:48:34 -06:00 |
|
|
Jaret Burkett
|
85a6880643
|
Move the offload and quantization to the base model
|
2026-08-28 08:37:56 -06:00 |
|
|
Jaret Burkett
|
702254688d
|
Added legacy paths
|
2026-08-28 07:55:19 -06:00 |
|
|
Jaret Burkett
|
c3bc8b0b4e
|
Full test of ui arch compatability
|
2026-08-28 06:56:05 -06:00 |
|
|
Jaret Burkett
|
520d96aac3
|
Phase 2
|
2026-08-27 18:18:31 -06:00 |
|
|
Jaret Burkett
|
45886f01b2
|
Phase 2
|
2026-08-27 16:16:35 -06:00 |
|
|
Jaret Burkett
|
9113420b61
|
Phase 2
|
2026-08-27 15:37:13 -06:00 |
|
|
Jaret Burkett
|
9ed2e0b8e7
|
Test model loading
|
2026-08-27 14:31:53 -06:00 |
|
|
Jaret Burkett
|
8db198ec0a
|
Phase 1
|
2026-08-27 11:53:08 -06:00 |
|
|
Jaret Burkett
|
e8d9cf6d35
|
Models v2 - phase 0
|
2026-08-27 10:51:43 -06:00 |
|
|
Jaret Burkett
|
5497a001cb
|
Fix issue where errors on qwen image omni were not being fully surfaced.
|
2026-08-26 20:54:24 -06:00 |
|
|
Jaret Burkett
|
da79ebce99
|
Add D-OPSD as a distillation handeling option for MiniMax H3 ref2va
|
2026-08-26 11:06:33 -06:00 |
|
|
Jaret Burkett
|
8a912564ce
|
Fix video DOP methods
|
2026-08-26 10:05:14 -06:00 |
|
|
Jaret Burkett
|
8436c407f6
|
Add ability to delete the loss logs from a selected range on the loss graph
|
2026-08-23 09:57:47 -06:00 |
|
|
Jaret Burkett
|
27a03a91f2
|
Rework progress bar and ui speed string so it gets the same value at a more accurate and snappier wall clock time
|
2026-08-21 13:31:02 -06:00 |
|
|
Jaret Burkett
|
b96476a841
|
Allow overriding the batch size on a per dataset level
|
2026-08-21 12:56:04 -06:00 |
|
|
Jaret Burkett
|
89102f76dc
|
Allow user to pick if reference images are sent in the image or video reference stream for Hidream H3 ref2va
|
2026-08-19 13:32:06 -06:00 |
|
|
Jaret Burkett
|
afd1d92722
|
Fix issue with double sampling qwen vl video frames for hidream h3 for reference videos
|
2026-08-17 15:39:07 -06:00 |
|
|
Jaret Burkett
|
42dfe9c661
|
Improve video frame reader
|
2026-08-16 10:21:11 -06:00 |
|
|
Jaret Burkett
|
b982a03ae4
|
Add training adapter for MiniMax H3 ref2va and default to contrastive loss with adapter
|
2026-08-16 09:11:07 -06:00 |
|
|
Jaret Burkett
|
2042481914
|
On MiniMax H3 set the default to use both contrastive guidance and the training adapter as it is yielding better results than one or the other.
|
2026-08-16 08:43:31 -06:00 |
|
|
Jaret Burkett
|
61310c6397
|
Inprove cpu data acuisition on backend
|
2026-08-16 08:42:36 -06:00 |
|
|
Jaret Burkett
|
2cbc2bb097
|
On minimax h3, trim cached text embeddings to max tokens if they are longer than specified
|
2026-08-16 08:42:01 -06:00 |
|
|
Jaret Burkett
|
0f788923ae
|
Animate action bar so starting a job has a 'working' indicator.
|
2026-08-15 10:33:40 -06:00 |
|
|
Jaret Burkett
|
e6cffbc002
|
Adjust encoded paths so filenames resolve as a basename properly when downloading with things like wget.
|
2026-08-15 08:56:55 -06:00 |
|
|
Jaret Burkett
|
151ad0e959
|
Adjust MiniMah h3 sizing for videos to downscale to match
|
2026-08-15 07:50:42 -06:00 |
|
|
Jaret Burkett
|
70b1089359
|
Version bump
|
2026-08-15 07:14:13 -06:00 |
|
|
Jaret Burkett
|
127d6f626d
|
Rework img/video reference in Minimax H3 to more closely match the comfy ui implementation.
|
2026-08-15 07:13:54 -06:00 |
|
|
Jaret Burkett
|
f1faa7725b
|
version bump
|
2026-08-15 06:20:32 -06:00 |
|
|
Jaret Burkett
|
97bf49edad
|
Add support for video references in MiniMax H3 ref2va
|
2026-08-15 06:18:09 -06:00 |
|
|
Jaret Burkett
|
4900e5e866
|
Version bump
|
2026-08-14 19:54:51 -06:00 |
|
|
Jaret Burkett
|
5f53ecde54
|
Make the viewer thumbnails pul the thumbs
|
2026-08-14 18:14:12 -06:00 |
|
|
Jaret Burkett
|
247cb45c3e
|
Improve gpu cpu monitor efficiency.
|
2026-08-14 17:14:00 -06:00 |
|
|
Jaret Burkett
|
695b0baccf
|
Switch to a live always active gpu and cpu device monitor that can be connected to with an SSE connection for real time device information stream. Much more efficient and faster than previous method of polling.
|
2026-08-14 12:40:01 -06:00 |
|
|
Jaret Burkett
|
5261d3fcca
|
Add checkpointing to the Wan2.1 encoder
|
2026-08-14 07:31:16 -06:00 |
|
|
Jaret Burkett
|
0e4b6e8695
|
Version bump
|
2026-08-13 12:36:34 -06:00 |
|
|
Jaret Burkett
|
6ea281973d
|
Add support for MiniMax H3 Ref2Vid training
|
2026-08-13 12:36:03 -06:00 |
|
|
Jaret Burkett
|
ab18528fdb
|
Let omni captioner handle just images as well. Set it as the new default captioner.
|
2026-08-13 10:30:17 -06:00 |
|
|
Jaret Burkett
|
6b7fb60a22
|
Make contrastive guidance loss do constant instead of sigma loss schedule by default
|
2026-08-13 09:52:02 -06:00 |
|
|
Jaret Burkett
|
4e91fb2d0a
|
Add thinking and abliterated versions of qwen omni.
|
2026-08-13 08:56:31 -06:00 |
|
|
Jaret Burkett
|
6c88e3d138
|
Add layer offloading to omni captioner
|
2026-08-13 07:21:54 -06:00 |
|
|
Jaret Burkett
|
a69f3e8710
|
Version bump
|
2026-08-12 21:38:00 -06:00 |
|
|
Jaret Burkett
|
742a4c8cef
|
Added a Download Full Dataset and Download Captions option to the datasets page so they captions or dataset can be easily downloaded.
|
2026-08-12 21:37:35 -06:00 |
|
|
Jaret Burkett
|
81adcc2176
|
Add caption prompt template picker to the ui. A few styling improvements thrown in for a good time.
|
2026-08-12 21:23:28 -06:00 |
|
|
Jaret Burkett
|
ca42a72f4c
|
Fixed issue with flash attention on omni captioner
|
2026-08-12 21:14:23 -06:00 |
|
|
Jaret Burkett
|
7f9a142dfd
|
Generate thumbnails for dataset items so things load faster and the ui is more stable. Especially good for videos.
|
2026-08-12 20:33:42 -06:00 |
|
|
Jaret Burkett
|
175cc1e151
|
Add Qwen 3 Omni for captioning videos with sound.
|
2026-08-12 19:57:01 -06:00 |
|
|
Jaret Burkett
|
4b00b61257
|
When doing an audio loss, show the img and audio loss in the loss log
|
2026-08-12 18:58:54 -06:00 |
|
|
Jaret Burkett
|
e16e04f123
|
On the ui, keep the last caption job bar visible at the bottom of a dataset so it can be viewed and edited
|
2026-08-12 17:32:48 -06:00 |
|
|
Jaret Burkett
|
18645d93b7
|
Fix trailing slashes in path for settings
|
2026-08-12 17:04:22 -06:00 |
|
|
Jaret Burkett
|
a1ddeeef13
|
Fixed issue with offloading text encoder on ltx 2.5
|
2026-08-12 10:46:58 -06:00 |
|
|
Jaret Burkett
|
0fd3e61c4c
|
Expand loras to match if they are not the same size in lora merger
|
2026-08-12 10:04:44 -06:00 |
|
|
Jaret Burkett
|
0bacc88e47
|
Version Bump
|
2026-08-12 07:55:21 -06:00 |
|
|
Jaret Burkett
|
7eb65b837a
|
Switch MiniMax H3 to default to contrastive guidance loss. Add a custom model select toggle to automatically fill out settings for the model.
|
2026-08-12 07:54:57 -06:00 |
|
|
Jaret Burkett
|
cbf910ac02
|
Add support for LTX 2.5
|
2026-08-12 05:50:30 -06:00 |
|
|
Jaret Burkett
|
924c426675
|
Resolve the new ltx 2.5 path subfolder for ltx 2.5. Sill checking compatability.
|
2026-08-11 13:30:17 -06:00 |
|
|
Jaret Burkett
|
62017a915a
|
Version bump
|
2026-08-11 08:30:09 -06:00 |
|
|
Jaret Burkett
|
21dc65972d
|
Only add a blank control if the model has to have it, for unconditionals. Previously it always encoded a blank control on unconditional if the model could take one. Affects DOP and blank prompt preservations.
|
2026-08-11 08:25:23 -06:00 |
|
|
Jaret Burkett
|
f421542df4
|
When dropping out, caching, and doing DOP, make sure we select a cahced trigger word when dropped out so DOP matches the drop out embeddings.
|
2026-08-11 06:59:20 -06:00 |
|
|
Jaret Burkett
|
8d4beedd04
|
Move guidance loss target selection when given a range before preservation so it is avaliable during preservation.
|
2026-08-11 06:57:33 -06:00 |
|
|
Jaret Burkett
|
ab5fef8970
|
Fix caption dropout so it now works identically when caching text embeddings.
|
2026-08-10 18:53:33 -06:00 |
|
|
Jaret Burkett
|
356ce7e84e
|
Version bump
|
2026-08-10 09:21:27 -06:00 |
|
|
Jaret Burkett
|
257da9b586
|
Rework DOP so it works with caching text embeddings
|
2026-08-09 22:13:49 -06:00 |
|
|
Jaret Burkett
|
5ff8a0435a
|
Make v1 the default training adapter
|
2026-08-09 14:56:39 -06:00 |
|
|
Jaret Burkett
|
61da9c95d3
|
Version bump
|
2026-08-09 12:39:30 -06:00 |
|
|
Jaret Burkett
|
72623ed3d6
|
When doing auto frame count. Ensure the time is not squeezed or expanded to fit tempooral spacing. trime the few extra frames. Also fixed frame counts of buckets.
|
2026-08-09 12:31:02 -06:00 |
|
|
Jaret Burkett
|
682b27c6ee
|
Reworked freeing memory manager for removing text encoder completly when not needed.
|
2026-08-09 07:08:24 -06:00 |
|
|
Jaret Burkett
|
3a28c4b1b7
|
Keep alive cron server to prevent failed connections
|
2026-08-08 20:27:01 -06:00 |
|
|
Jaret Burkett
|
8c1a4082fd
|
Allow images to work with auto frame count, and include images in video datasets if they exist.
|
2026-08-08 20:26:28 -06:00 |
|
|
Jaret Burkett
|
6d8afa5684
|
Version bump
|
2026-08-08 18:58:11 -06:00 |
|
|
Jaret Burkett (Ostris)
|
c596d4ab27
|
Merge pull request #1002 from whatsthisaithing/codex/fix-convrot-offload-stream-lifetime
Fix offload buffer stream lifetime
|
2026-08-08 18:55:18 -06:00 |
|
|
Jaret Burkett
|
f4e9130547
|
Fix race condition that can corrupt grads under certain conditions.
|
2026-08-07 14:52:47 -06:00 |
|
|
Jaret Burkett
|
817f3dcbcb
|
Fix finite check on inverted masked prior
|
2026-08-07 09:49:12 -06:00 |
|
|
Jaret Burkett
|
685ce37a8d
|
Adjust sample timestep sigmas to be model evals for h3 for 1 extra step
|
2026-08-06 19:54:04 -06:00 |
|
|
Jaret Burkett
|
9171d5ec1d
|
Update training adapter path
|
2026-08-06 11:21:28 -06:00 |
|
|
Jaret Burkett
|
71625d1207
|
Add the alpha version of the MiniMax H3 trianing adapter and set it as the new default training method.
|
2026-08-06 11:11:31 -06:00 |
|
|
Jaret Burkett
|
b904b99705
|
varsion bump
|
2026-08-06 09:25:35 -06:00 |
|
|
Jaret Burkett
|
b811636ae4
|
Check for finite vs isnan on loss before backpropigating. to catch infinity overflows
|
2026-08-06 09:18:56 -06:00 |
|