AnimatedDiff isnt working, produces a black output. Whats the reason>

[visualhippocracy]

Hi,

So i tried multiple settings but animateddiff only gives black output.

The following is the error i get

Use cn inpaint instead of sd inpaint
2024-07-15 03:03:07,216 - AnimateDiff - INFO - AnimateDiff process start.
2024-07-15 03:03:07,232 - AnimateDiff - INFO - Injecting motion module mm_sd_v15_v2.ckpt into SD1.5 UNet middle block.
2024-07-15 03:03:07,232 - AnimateDiff - INFO - Injecting motion module mm_sd_v15_v2.ckpt into SD1.5 UNet input blocks.
2024-07-15 03:03:07,233 - AnimateDiff - INFO - Injecting motion module mm_sd_v15_v2.ckpt into SD1.5 UNet output blocks.
2024-07-15 03:03:07,233 - AnimateDiff - INFO - Setting DDIM alpha.
2024-07-15 03:03:07,237 - AnimateDiff - INFO - Injection finished.
2024-07-15 03:03:07,238 - AnimateDiff - INFO - AnimateDiff + ControlNet will generate 6 frames.
2024-07-15 03:03:07,301 - ControlNet - INFO - unit_separate = False, style_align = False
2024-07-15 03:03:07,301 - ControlNet - INFO - Loading model from cache: control_v11p_sd15_inpaint [ebff9138]
2024-07-15 03:03:07,302 - ControlNet - INFO - AnimateDiff + ControlNet inpaint_only receive the following parameters:
2024-07-15 03:03:07,303 - ControlNet - INFO - batch control images: D:\out\out_606395167_1720992760\fragment_1\frames
2024-07-15 03:03:07,529 - ControlNet - INFO - Using preprocessor: inpaint_only
2024-07-15 03:03:07,529 - ControlNet - INFO - preprocessor resolution = -1
100%|████████████████████████████████████████████████████████████████████████████████████| 6/6 [00:01<00:00, 5.46it/s]
2024-07-15 03:03:08,735 - ControlNet - INFO - ControlNet Hooked - Time = 1.442753553390503
2024-07-15 03:03:10,134 - AnimateDiff - INFO - Randomizing init_latent according to [1.0, 0.96875, 0.9375, 0.90625, 0.875, 0.84375].
*** Error running before_process_batch: C:\Users\Dhruv\SDXL\extensions\sd-webui-animatediff\scripts\animatediff.py
Traceback (most recent call last):
File "C:\Users\Dhruv\SDXL\modules\scripts.py", line 833, in before_process_batch
script.before_process_batch(p, *script_args, **kwargs)
File "C:\Users\Dhruv\SDXL\extensions\sd-webui-animatediff\scripts\animatediff.py", line 79, in before_process_batch
AnimateDiffI2VLatent().randomize(p, params)
File "C:\Users\Dhruv\SDXL\extensions\sd-webui-animatediff\scripts\animatediff_latent.py", line 84, in randomize
p.init_latent = p.init_latent * init_alpha + p.rng.next() * (1 - init_alpha)
RuntimeError: The size of tensor a (12) must match the size of tensor b (6) at non-singleton dimension 0

---

0%| | 0/20 [00:00<?, ?it/s]2024-07-15 03:03:10,195 - AnimateDiff - INFO - inner model forward hooked
2024-07-15 03:03:11,509 - ControlNet - INFO - ControlNet used torch.float32 VAE to encode torch.Size([12, 4, 64, 64]).
100%|██████████████████████████████████████████████████████████████████████████████████| 20/20 [00:29<00:00, 1.50s/it]
*** A tensor with all NaNs was produced in VAE. Use --disable-nan-check commandline argument to disable this check. ***
Traceback (most recent call last):
File "C:\Users\Dhruv\SDXL\extensions\sd-webui-replacer\replacer\video_animatediff.py", line 132, in animatediffGenerate
processed = processFragment(fragmentPath, initImage, gArgs)
File "C:\Users\Dhruv\SDXL\extensions\sd-webui-replacer\replacer\video_animatediff.py", line 25, in processFragment
processed, _ = inpaint(initImage, gArgs)
File "C:\Users\Dhruv\SDXL\extensions\sd-webui-replacer\replacer\inpaint.py", line 97, in inpaint
processed = process_images(p)
File "C:\Users\Dhruv\SDXL\modules\processing.py", line 845, in process_images
res = process_images_inner(p)
File "C:\Users\Dhruv\SDXL\extensions\sd-webui-controlnet\scripts\batch_hijack.py", line 48, in processing_process_images_hijack
return getattr(processing, '__controlnet_original_process_images_inner')(p, *args, **kwargs)
File "C:\Users\Dhruv\SDXL\modules\processing.py", line 993, in process_images_inner
x_samples_ddim = decode_latent_batch(p.sd_model, samples_ddim, target_device=devices.cpu, check_for_nans=True)
File "C:\Users\Dhruv\SDXL\modules\processing.py", line 654, in decode_latent_batch
raise e
File "C:\Users\Dhruv\SDXL\modules\processing.py", line 638, in decode_latent_batch
devices.test_for_nans(sample, "vae")
File "C:\Users\Dhruv\SDXL\modules\devices.py", line 255, in test_for_nans
raise NansException(message)
modules.devices.NansException: A tensor with all NaNs was produced in VAE. Use --disable-nan-check commandline argument to disable this check.

---

merging fragments
0it [00:00, ?it/s]

What could be casuing this? Disabling nan check didnt do anything. It just throw this error, but still produces black output

[light-and-ray]

I think you use wrong sd model. You need to select sd1.5 non-inpainting model inside animatediff tab