comfyanonymous
5c7e72f3b1
Load controlnet in fp8 if weights are in fp8.
2024-09-21 04:50:12 -04:00
Simon Lui
c4affc7e35
Fix xpu Pytorch nightly build from calling optimize which doesn't exist. ( #4978 )
2024-09-19 05:11:42 -04:00
comfyanonymous
ebba1f2563
Prioritize freeing partially offloaded models first.
2024-09-04 19:47:32 -04:00
comfyanonymous
ac3c801c9d
Speed up inference on nvidia 10 series on Linux.
2024-09-01 17:29:31 -04:00
comfyanonymous
8e7d3547e4
Make minimum_inference_memory() depend on --reserve-vram
2024-09-01 01:18:34 -04:00
comfyanonymous
19cdff6796
Cleanup.
2024-08-30 12:53:59 -04:00
comfyanonymous
a80296ecb6
Fix case where model was not properly unloaded in merging workflows.
2024-08-27 19:03:51 -04:00
comfyanonymous
65a577e0db
Cleanup.
2024-08-23 04:06:27 -04:00
comfyanonymous
103d705247
Fix.
2024-08-23 04:04:55 -04:00
Simon Lui
24af9ec35a
Rework IPEX check for future inclusion of XPU into Pytorch upstream and do a bit more optimization of ipex.optimize(). ( #4562 )
2024-08-23 03:59:57 -04:00
comfyanonymous
23836e6596
Code cleanups.
2024-08-22 17:05:12 -04:00
comfyanonymous
7bad18b66d
Fix issue with models staying loaded in memory.
2024-08-22 15:58:20 -04:00
David
ff022fccc2
Generalize MacOS version check for force-upcast-attention ( #4548 )
...
This code automatically forces upcasting attention for MacOS versions 14.5 and 14.6. My computer returns the string "14.6.1" for `platform.mac_ver()[0]`, so this generalizes the comparison to catch more versions.
I am running MacOS Sonoma 14.6.1 (latest version) and was seeing black image generation on previously functional workflows after recent software updates. This PR solved the issue for me.
See comfyanonymous/ComfyUI#3521
2024-08-22 13:24:21 -04:00
comfyanonymous
190a03c892
fp16 is actually faster than fp32 on a GTX 1080.
2024-08-21 23:23:50 -04:00
comfyanonymous
883e5a9aa5
Fix slow performance on 10 series Nvidia GPUs.
2024-08-21 16:39:02 -04:00
comfyanonymous
d8af5d17b5
Remove useless line, adjust windows default reserved vram.
2024-08-21 00:47:19 -04:00
comfyanonymous
7c0c89e70c
Add --fast argument to enable experimental optimizations.
...
Optimizations that might break things/lower quality will be put behind
this flag first and might be enabled by default in the future.
Currently the only optimization is float8_e4m3fn matrix multiplication on
4000/ADA series Nvidia cards or later. If you have one of these cards you
will see a speed boost when using fp8_e4m3fn flux for example.
2024-08-20 11:55:51 -04:00
comfyanonymous
c3bec99a9b
Fix potential issue with multi devices.
2024-08-20 10:46:36 -04:00
comfyanonymous
5115d926eb
Add a --reserve-vram argument if you don't want comfy to use all of it.
...
--reserve-vram 1.0 for example will make ComfyUI try to keep 1GB vram free.
This can also be useful if workflows are failing because of OOM errors but
in that case please report it if --reserve-vram improves your situation.
2024-08-19 17:16:18 -04:00
comfyanonymous
48335ad8ca
Remove duplication.
2024-08-19 15:26:50 -04:00
comfyanonymous
2119417b9d
Fix bug when model cannot be partially unloaded.
2024-08-13 03:57:55 -04:00
comfyanonymous
a552dfba70
Fix some issues with TE being in lowvram mode.
2024-08-12 23:42:21 -04:00
comfyanonymous
5faeccb7f7
Memory tweaks.
2024-08-12 15:07:11 -04:00
comfyanonymous
d95ec784a4
Fix some potential issues with the previous commits.
2024-08-12 00:23:29 -04:00
comfyanonymous
089485a7c6
Load TE model straight to vram if certain conditions are met.
2024-08-11 23:52:43 -04:00
comfyanonymous
cbb1bde521
Fix some issues with inference slowing down.
2024-08-10 16:21:25 -04:00
comfyanonymous
1280396253
Fix regression.
2024-08-09 03:36:40 -04:00
comfyanonymous
114af84d12
Try to improve inference speed on some machines.
2024-08-08 17:29:27 -04:00
comfyanonymous
4f196c19a3
Fix.
2024-08-08 15:16:51 -04:00
comfyanonymous
ede5622146
Partial model shift support.
2024-08-08 14:45:06 -04:00
comfyanonymous
5a25225024
Make supported_dtypes a priority list.
2024-08-07 15:00:06 -04:00
comfyanonymous
001ff4d6a6
Fix OOMs happening in some cases.
...
A cloned model patcher sometimes reported a model was loaded on a device
when it wasn't.
2024-08-06 13:36:04 -04:00
comfyanonymous
5228fbf0ad
Unload models and load them back in lowvram mode no free vram.
2024-08-06 03:22:39 -04:00
comfyanonymous
f7183e2c73
Improve performance on some lowend GPUs.
2024-08-05 16:24:04 -04:00
comfyanonymous
10ad73d520
Fix crash.
2024-08-03 16:55:38 -04:00
comfyanonymous
e7287d2fc2
Tweak lowvram memory formula.
2024-08-03 16:44:50 -04:00
comfyanonymous
d4ca060482
Lower lowvram memory to 1/3 of free memory.
2024-08-03 15:14:07 -04:00
comfyanonymous
f27f0f8a48
Fix some issues.
2024-08-03 15:06:40 -04:00
comfyanonymous
e83d97d649
Cap lowvram to half of free memory.
2024-08-03 14:50:20 -04:00
comfyanonymous
753d9640f9
Automatically use fp8 for diffusion model weights if:
...
Checkpoint contains weights in fp8.
There isn't enough memory to load the diffusion model in GPU vram.
2024-08-03 13:45:19 -04:00
comfyanonymous
f176649e08
Make ComfyUI split batches a higher priority than weight offload.
2024-08-01 16:39:59 -04:00
comfyanonymous
f5b2fc6e6b
Fix bfloat16 potentially not being enabled on mps.
2024-08-01 16:18:44 -04:00
comfyanonymous
faad5f1cdb
Make lowvram more aggressive on low memory machines.
2024-08-01 12:11:57 -04:00
comfyanonymous
d6d48306d8
Load flux t5 in fp8 if weights are in fp8.
2024-08-01 11:05:56 -04:00
comfyanonymous
339a710a7a
Mac supports bf16 just make sure you are using the latest pytorch.
2024-08-01 09:42:17 -04:00
comfyanonymous
78c273b0ba
Make lowvram less aggressive when there are large amounts of free memory.
2024-08-01 03:58:58 -04:00
comfyanonymous
160c550977
Use fp16 as the default vae dtype for the audio VAE.
2024-06-16 13:12:54 -04:00
comfyanonymous
12e44a5d50
Add a --force-channels-last to inference models in channel last mode.
2024-06-15 01:08:12 -04:00
Simon Lui
20052d83b5
Exempt IPEX from non_blocking previews fixing segmentation faults. ( #3708 )
2024-06-13 18:51:14 -04:00
comfyanonymous
917810a2c6
Load the SD3 T5xxl model in the same dtype stored in the checkpoint.
2024-06-11 17:03:26 -04:00
