comfyanonymous
c73b5fab20
attention_basic now works with hypertile.
2023-10-22 03:59:53 -04:00
comfyanonymous
57381b0892
Make sub_quad and split work with hypertile.
2023-10-22 03:51:29 -04:00
comfyanonymous
5a9a1a50af
Make xformers work with hypertile.
2023-10-21 13:23:03 -04:00
comfyanonymous
459787f78f
Make VAE code closer to sgm.
2023-10-17 15:18:51 -04:00
comfyanonymous
28b98a96d3
Refactor the attention stuff in the VAE.
2023-10-17 03:19:29 -04:00
comfyanonymous
bffd427388
Add a separate optimized_attention_masked function.
2023-10-16 02:31:24 -04:00
comfyanonymous
728139a5b9
Refactor code so model can be a dtype other than fp32 or fp16.
2023-10-13 14:41:17 -04:00
comfyanonymous
494ddf7717
pytorch_attention_enabled can now return True when xformers is enabled.
2023-10-11 21:30:57 -04:00
comfyanonymous
95df4b6174
Allow attn_mask in attention_pytorch.
2023-10-11 20:38:48 -04:00
comfyanonymous
c60864b5e4
Refactor the attention functions.
...
There's no reason for the whole CrossAttention object to be repeated when
only the operation in the middle changes.
2023-10-11 20:38:48 -04:00
comfyanonymous
23ed2b1654
Model patches can now know which batch is positive and negative.
2023-09-27 12:04:07 -04:00
comfyanonymous
a990441ba0
Add a way to set output block patches to modify the h and hsp.
2023-09-22 20:26:47 -04:00
comfyanonymous
ef0c0892f6
Add a force argument to soft_empty_cache to force a cache empty.
2023-09-04 00:58:18 -04:00
Simon Lui
f63ffd1bb1
Revert changes in comfy/ldm/modules/diffusionmodules/util.py, which is unused.
2023-09-02 20:07:52 -07:00
Simon Lui
1148c2dec7
Some fixes to generalize CUDA specific functionality to Intel or other GPUs.
2023-09-02 18:22:10 -07:00
comfyanonymous
b8f3570a1b
Remove xformers related print.
2023-09-01 02:12:03 -04:00
comfyanonymous
2ab478346d
Remove optimization that caused border.
2023-08-29 11:21:36 -04:00
comfyanonymous
6932eda1fb
Fallback to slice attention if xformers doesn't support the operation.
2023-08-27 22:24:42 -04:00
comfyanonymous
1b9a6a9599
Make --bf16-vae work on torch 2.0
2023-08-27 21:33:53 -04:00
comfyanonymous
08af73e450
Controlnet/t2iadapter cleanup.
2023-08-22 01:06:26 -04:00
comfyanonymous
01a6f9b116
Fix issue with gligen.
2023-08-18 16:32:23 -04:00
comfyanonymous
280659a6ee
Support for Control Loras.
...
Control loras are controlnets where some of the weights are stored in
"lora" format: an up and a down low rank matrice that when multiplied
together and added to the unet weight give the controlnet weight.
This allows a much smaller memory footprint depending on the rank of the
matrices.
These controlnets are used just like regular ones.
2023-08-18 11:59:51 -04:00
comfyanonymous
9c5ad64310
Remove some useless code.
2023-07-30 14:13:33 -04:00
comfyanonymous
71ffc7b350
Faster VAE loading.
2023-07-29 16:28:30 -04:00
comfyanonymous
8a8d8c86d6
Initialize the unet directly on the target device.
2023-07-29 14:51:56 -04:00
comfyanonymous
4ec7f09adc
It's actually possible to torch.compile the unet now.
2023-07-18 21:36:35 -04:00
comfyanonymous
fa8010f038
Disable autocast in unet for increased speed.
2023-07-05 21:58:29 -04:00
comfyanonymous
033dc1f52a
Cleanup.
2023-07-02 11:58:23 -04:00
comfyanonymous
391ee8d21f
Fix bug with controlnet.
2023-06-24 11:02:38 -04:00
comfyanonymous
0db33017af
Add some more transformer hooks and move tomesd to comfy_extras.
...
Tomesd now uses q instead of x to decide which tokens to merge because
it seems to give better results.
2023-06-24 03:30:22 -04:00
comfyanonymous
51da619d73
Remove useless code.
2023-06-23 12:35:26 -04:00
comfyanonymous
08f1f7686c
Support base SDXL and SDXL refiner models.
...
Large refactor of the model detection and loading code.
2023-06-22 13:03:50 -04:00
comfyanonymous
d72d5d49f5
Add original_shape parameter to transformer patch extra_options.
2023-06-21 13:22:01 -04:00
comfyanonymous
873b08bd0f
Add a way to set patches that modify the attn2 output.
...
Change the transformer patches function format to be more future proof.
2023-06-18 22:58:22 -04:00
comfyanonymous
d35c0ce6ec
All the unet weights should now be initialized with the right dtype.
2023-06-15 18:42:30 -04:00
comfyanonymous
8e6cbe6270
Initialize more unet weights as the right dtype.
2023-06-15 15:00:10 -04:00
comfyanonymous
206af9a315
Initialize transformer unet block weights in right dtype at the start.
2023-06-15 14:29:26 -04:00
comfyanonymous
6bbd86a061
Disable default weight values in unet conv2d for faster loading.
2023-06-14 19:46:08 -04:00
comfyanonymous
46a1fd77f4
This isn't needed for inference.
2023-06-14 13:05:08 -04:00
comfyanonymous
14508d7a5b
Speed up model loading a bit.
...
Default pytorch Linear initializes the weights which is useless and slow.
2023-06-14 12:09:41 -04:00
comfyanonymous
22cd647be0
Remove more useless files.
2023-06-13 02:22:19 -04:00
comfyanonymous
ce4e360edf
Cleanup: Remove a bunch of useless files.
2023-06-13 02:19:08 -04:00
comfyanonymous
c0a5444d1b
Make scaled_dot_product switch to sliced attention on OOM.
2023-05-20 16:01:02 -04:00
comfyanonymous
e33cb62d1b
Simplify and improve some vae attention code.
2023-05-20 15:07:21 -04:00
BlenderNeko
2e2c17131b
minor changes for tiled sampler
2023-05-12 23:49:09 +02:00
comfyanonymous
b877fefbb3
Lowvram mode for gligen and fix some lowvram issues.
2023-05-05 18:11:41 -04:00
comfyanonymous
2edaaba3c2
Fix imports.
2023-05-04 18:10:29 -04:00
comfyanonymous
71b4d08e65
Change latent resolution step to 8.
2023-05-02 14:17:51 -04:00
comfyanonymous
76730ed9ff
Make unet work with any input shape.
2023-05-02 13:31:43 -04:00
comfyanonymous
e6771d0986
Implement Linear hypernetworks.
...
Add a HypernetworkLoader node to use hypernetworks.
2023-04-23 12:35:25 -04:00
