09e3f3dd5f
move residual arg.
2025-01-30 00:03:12 +00:00
be1cb10894
make residual optional.
2025-01-30 00:02:31 +00:00
8fe0ed225a
Conditioned Adapter make residual optional.
2025-01-29 23:49:22 +00:00
41eec16c67
Renamed a few variables to reduce cog load.
2025-01-29 23:26:39 +00:00
e4a14af8d7
Add back in the pose for both channels.
2025-01-29 22:40:18 +00:00
bca44f6d7f
Remove some comments.
2025-01-29 21:11:55 +00:00
536dfc8f6a
Figure out why classifier free guidence breaks the loop. I suspect it wasn't baked into pipeline should have been 0
2025-01-29 06:10:29 +00:00
a16285141e
Noted place that requires intervention.
2025-01-29 05:30:17 +00:00
707e001f7b
The encoder hidden states and encode attention mask sizes do not match up in the validation pipeline.
2025-01-29 05:24:06 +00:00
4ab3b25c13
Make a fix to the video tensors.
2025-01-28 23:03:04 +00:00
8ec35d58fb
trying training again but broke something so going back.
2025-01-28 22:36:09 +00:00
dfa03f1ed1
Fix before trying again.
2025-01-27 08:30:45 +00:00
80f653f336
Passed data in for validation.
2025-01-25 03:07:12 +00:00
e99d70ff46
Fixed some issues with imports and the original lora.py code.
2025-01-25 02:22:05 +00:00
06e26ec092
test and trace the changes.
2025-01-24 22:58:44 +00:00
c6490c0d51
Reducing cogload on the framework.
2025-01-24 22:46:32 +00:00
305a4a994d
Need to pass in image ref video.
2025-01-24 05:21:53 +00:00
440738e535
Add conditioned pipeline ...
2025-01-24 02:47:58 +00:00
64898ef20d
Test the conditioned pipeline and ensure it loads.
2025-01-24 02:47:23 +00:00
b7f26bd2e6
removed some code made things clearer to replace the pipeline inference which is needed for validation.
2025-01-23 23:49:24 +00:00
d6b3615421
Ok adapter trains need to get the pass for validation using a skeleton.
2025-01-22 22:07:52 +00:00
d631737329
Post patchify the latents.
2025-01-22 21:24:33 +00:00
731f326e86
Added residual
2025-01-22 20:10:01 +00:00
981767154b
modify the forward pass
2025-01-22 06:56:22 +00:00
07e755c2b3
modifcations to the loading of the weights.
2025-01-22 01:26:49 +00:00
9fde6600dc
Fixing weights copy.
2025-01-22 01:26:17 +00:00
34996c1eb1
replaced the pipeline functions.
...
time to test this then test the weight changes.
2025-01-21 02:22:35 +00:00
48d2b61029
Need to test saving and loading.
2025-01-18 06:58:04 +00:00
3c605a124e
Adding in the conditioned transformer 3d model
2025-01-18 06:47:39 +00:00
fa422c4ca5
Adding conditioned transformer layer change.
2025-01-18 06:29:01 +00:00
2720d24f62
Moved the norm back to prepare latents step.
...
Now its time to add the adapter layer.
2025-01-18 04:18:25 +00:00
44df9e3bce
reformulate the conditioning and patchification
2025-01-18 03:11:09 +00:00
59afd55775
Retested changes without conditioning it works.
2025-01-17 22:43:20 +00:00
b73d61ae4d
saving more changes related to args.
2025-01-17 06:40:06 +00:00
47356cd864
Testing the original code path.
...
Need to check rectified linear flow vs flow matching.
2025-01-17 06:03:45 +00:00
cd66e6e58e
fixed the training pipeline to address new changes from upstream
2025-01-17 01:46:35 +00:00
315431be0a
Recommit the conditioning code.
2025-01-16 05:03:44 +00:00
b341b79652
Adding Conditioning Blocks to the Diffusion Transformer.
2025-01-15 22:29:47 +00:00
f34b8659c3
Patchification research to condition the model
2025-01-14 18:14:53 +00:00
9f0fdb5055
Rename lora files ( #214 )
...
* rename files
* update
2025-01-14 01:17:42 +05:30
354a6b9e25
reposition private methods ( #213 )
2025-01-14 00:08:04 +05:30
e5df80cc36
Full Finetuning for LTX possibily extended to other models. ( #192 )
...
* Full Finetuning for LTX possibily extended to other models.
* Change name of the flag
* Used disable grad for component on lora fine tuning enabled
* Suggestions Addressed
Renamed to SFT
Added 2 other models.
Testing required.
* Switching to Full FineTuning
* Run linter.
* parse subfolder when needed.
* tackle saving and loading hooks.
* tackle validation.
* fix subfolder bug.
* remove __class__.
* refactor
* remove unnecessary changes
* handle saving of final model weights correctly
* remove unnecessary changes
* LTX uses a default frame rate of 24 FPS
We need to modify the output validation framerate to match that value.
Add Framerate args.
Add Update video output and inference frame rate
* There was a results_args mapping that needed to be modified.
* update
* update README
* Update README.md
* update docs
* add training configuration in cogvideox
---------
Co-authored-by: Sayak Paul <spsayakpaul@gmail.com >
Co-authored-by: Aryan <aryan@huggingface.co >
Co-authored-by: Aryan <contact.aryanvs@gmail.com >
2025-01-13 23:55:39 +05:30
a3898075d1
Helpful mesages ( #210 )
2025-01-11 06:35:28 +05:30
d3b797ee35
Improve argument handling ( #209 )
2025-01-11 06:18:13 +05:30
81bf590a91
Argument descriptions ( #208 )
...
* add argument descriptions
* update
* update
* update
2025-01-11 05:06:02 +05:30
b34091362e
Fix the checkpoint dir bug in get_intermediate_ckpt_path so that checkpointing_limit can actually work. ( #207 )
2025-01-10 13:18:32 +05:30
4b21d37c9d
Update pr_tests.yml to update ruff version ( #205 )
2025-01-10 07:53:39 +05:30
9523a05220
Shell script to minimally test supported models on a real dataset ( #204 )
...
* start minimal complete tests
* updates
* fixes
* fixes
2025-01-10 07:22:00 +05:30
b9a6492ad1
Epoch loss ( #201 )
...
* update
* update
2025-01-09 14:22:56 +05:30
f311f16d98
[core] Fix loading of precomputed conditions and latents ( #199 )
...
* support precomputation.
* fixes
2025-01-09 12:43:03 +05:30
CrossProduct