Stable WarpFusion v0.23 - dwpose, controlnet preview

are you running v0.23.16?

Warpfusion

Hi. I just downloaded 0.23 Im running it in local env. NVIDIA 4080 16gb. when i difuse it it gives error about memory (i asume) link on the error file http://ftpx.forpsi.comwww.tagon.cz/work/AI_experiment/Error.txt I have tried old notebook 0.15 and it works just fine, there is no error about memory. Just the video is not very stable. Any help? advice? wHEN i run the 0.23 on Colab env. it is processing ok and never exceed 16gb of VRAM. thank you

OMonedr

https://www.patreon.com/posts/if-youre-having-91316252

Warpfusion

I get this error. What should I do??? Detected that PyTorch and torchvision were compiled with different CUDA major versions. PyTorch has CUDA Version=12.1 and torchvision has CUDA Version=11.8. Please reinstall the torchvision that matches your PyTorch install.

Raf

restart and run cells except for samtrack cli, looks like you've installed things into a wrong place

Warpfusion

Running docker+ubuntu and colab 0_23_15 with following error, all other colabs work normally. Also I now need to use the full path on all cells "/content/..." Run all cells a few times: Cell define SD + K functions, load model disabling init dummy for non-sdxl models --------------------------------------------------------------------------- FileNotFoundError Traceback (most recent call last) Cell In[56], line 768 765 if model_version in ['control_multi', 'control_multi_v2', 'control_multi_v2_768']: 767 if model_version == 'control_multi': --> 768 config = OmegaConf.load(f"{root_dir}/ControlNet/models/cldm_v15.yaml") 769 elif model_version in ['control_multi_v2', 'control_multi_v2_768']: 770 config = OmegaConf.load(f"{root_dir}/ControlNet/models/cldm_v21.yaml") File /usr/local/lib/python3.10/dist-packages/omegaconf/omegaconf.py:189, in OmegaConf.load(file_) 186 from ._utils import get_yaml_loader 188 if isinstance(file_, (str, pathlib.Path)): --> 189 with io.open(os.path.abspath(file_), "r", encoding="utf-8") as f: 190 obj = yaml.load(f, Loader=get_yaml_loader()) 191 elif getattr(file_, "read", None): FileNotFoundError: [Errno 2] No such file or directory: '/content/Segment-and-Track-Anything-CLI/ControlNet/models/cldm_v15.yaml'

Daniel Salvatierra

Run all cells please, this error means you have run literally 0 cells 😅

Warpfusion

Abadila Alaktif

download the file from this post

Warpfusion

Kindly tell me how to do that.

Ali Chishti

get 0.23.12 version

Warpfusion

kindly tell me what should i do

Ali Chishti

torchdata 0.6.1 requires torch==2.0.1, but you have torch 2.1.0 which is incompatible.

Ali Chishti

probably ran out of vram, try lowering the settings and restarting. also check you don't have a gpu-using apps in the background which may lock gpu access

Warpfusion

Can anyone help me out? xxxx

justin

Hi there, I'm getting the following error when using an SDXL model. INFO:pytorch_lightning.utilities.seed:Global seed set to 4275770367 1280 704 /content//videoFrames/35626dc988-25.0_0_0_1/000001.jpg Frame 0 Prompt: [' photograph of a Tim walker Nick Knight realistic haute couture afrofuturist style very close up macro half portrait of elegant dark skinned models with bioluminescent organic pearlescent transparent maximalist seashell facial fashion, opulent regal interior background, depth of field, hyper detailed, beautiful colour, cinematic, gradation coloured, 24mm wide angle lens shot, depth of field, DOF, super resolution, megapixels, Pro Photo GB, VR, medium back lighting, backlighting, cinematic lighting, studio lighting, soft lighting, glowing, trace reflections rays, reflections from lumens, diffraction gradation, chromatic aberration, elegant, hyper-realistic'] used_loras, used_loras_weights [] [] prompt weights, frame_prompt_weights {0: array([1.], dtype=float16)} [1.] Frame 0 neg_prompt: ['text, naked, nude, logo, cropped, two heads, four arms, lazy eye, blurry, unfocused'] Frame ['woman walking on a treadmill'] rec_prompt: ['woman walking on a treadmill'] ['text, naked, nude, logo, cropped, two heads, four arms, lazy eye, blurry, unfocused'] neg_prompt init_scale pre sd run 0 prev_frame not found, replacing 1st videoframe init seed clip_guidance_scale init_scale init_latent_scale clamp_grad clamp_max init_image skip_timesteps cfg_scale 4275770367 0 0 0 True 2.0 /content//videoFrames/35626dc988-25.0_0_0_1/000001.jpg 7 15 1280 704 WH prompts ['photograph of a Tim walker Nick Knight realistic haute couture afrofuturist style very close up macro half portrait of elegant dark skinned models with bioluminescent organic pearlescent transparent maximalist seashell facial fashion, opulent regal interior background, depth of field, hyper detailed, beautiful colour, cinematic, gradation coloured, 24mm wide angle lens shot, depth of field, DOF, super resolution, megapixels, Pro Photo GB, VR, medium back lighting, backlighting, cinematic lighting, studio lighting, soft lighting, glowing, trace reflections rays, reflections from lumens, diffraction gradation, chromatic aberration, elegant, hyper-realistic'] ['photograph of a Tim walker Nick Knight realistic haute couture afrofuturist style very close up macro half portrait of elegant dark skinned models with bioluminescent organic pearlescent transparent maximalist seashell facial fashion, opulent regal interior background, depth of field, hyper detailed, beautiful colour, cinematic, gradation coloured, 24mm wide angle lens shot, depth of field, DOF, super resolution, megapixels, Pro Photo GB, VR, medium back lighting, backlighting, cinematic lighting, studio lighting, soft lighting, glowing, trace reflections rays, reflections from lumens, diffraction gradation, chromatic aberration, elegant, hyper-realistic'] resized to (1280, 704) ------using tiled vae------ Tiled vae encoder took 0.6 Traceback (most recent call last): File "", line 625, in do_run() File "", line 1329, in do_run sample, latent, depth_img = run_sd(args, init_image=init_image, skip_timesteps=skip_steps, H=args.side_y, File "", line 2659, in run_sd uc = sd_model.get_learned_conditioning([neg_prompt]) File "", line 853, in get_unconditional_conditioning_sdxl c = self(batch_c) File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1501, in _call_impl return forward_call(*args, **kwargs) File "/content/generative-models/sgm/modules/encoders/modules.py", line 141, in forward emb_out = embedder(batch[embedder.input_key]) File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1501, in _call_impl return forward_call(*args, **kwargs) File "/content/./stablediffusion/modules/sd_hijack_clip.py", line 235, in forward z = self.process_tokens(tokens, multipliers) File "/content/./stablediffusion/modules/sd_hijack_clip.py", line 263, in process_tokens z = self.encode_with_transformers(tokens) File "/content/./stablediffusion/modules/sd_hijack_clip.py", line 311, in encode_with_transformers outputs = self.wrapped.transformer(input_ids=tokens, output_hidden_states=-opts.CLIP_stop_at_last_layers) File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1501, in _call_impl return forward_call(*args, **kwargs) File "/usr/local/lib/python3.10/dist-packages/transformers/models/clip/modeling_clip.py", line 722, in forward return self.text_model( File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1501, in _call_impl return forward_call(*args, **kwargs) File "/usr/local/lib/python3.10/dist-packages/transformers/models/clip/modeling_clip.py", line 637, in forward causal_attention_mask = self._build_causal_attention_mask(bsz, seq_len).to(hidden_states.device) RuntimeError: CUDA error: device-side assert triggered CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect. For debugging consider passing CUDA_LAUNCH_BLOCKING=1. Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions. n_stats_avg (mean, std): None None --------------------------------------------------------------------------- RuntimeError Traceback (most recent call last) in () 638 639 gc.collect() --> 640 torch.cuda.empty_cache() 641 executed_cells[cell_name] = True /usr/local/lib/python3.10/dist-packages/torch/cuda/memory.py in empty_cache() 131 """ 132 if is_initialized(): --> 133 torch._C._cuda_emptyCache() 134 135 RuntimeError: CUDA error: device-side assert triggered CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect. For debugging consider passing CUDA_LAUNCH_BLOCKING=1. Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.

justin

how did you make your local env? or is this colab env?

Warpfusion

it's here - the *.ipynb file attached to the post, just upload it to colab.

Warpfusion

Hi everyone, I didnt use Warp Fusion some times, and now i cant find Google Collab Warp Fusion version? Is it work now on Collab? Can anybody give me a link for it?

Dmitry Perkus

Hi, thanks for the update, im getting this error when i run a sdxl model. Traceback (most recent call last): File "C:\Users\bruno\AppData\Local\Temp\ipykernel_4588\1505760836.py", line 625, in do_run() File "C:\Users\bruno\AppData\Local\Temp\ipykernel_4588\3919419232.py", line 1329, in do_run sample, latent, depth_img = run_sd(args, init_image=init_image, skip_timesteps=skip_steps, H=args.side_y, File "C:\Users\bruno\AppData\Local\Temp\ipykernel_4588\3919419232.py", line 2659, in run_sd uc = sd_model.get_learned_conditioning([neg_prompt]) File "C:\Users\bruno\AppData\Local\Temp\ipykernel_4588\810790726.py", line 853, in get_unconditional_conditioning_sdxl c = self(batch_c) File "C:\code\Warpfusion\0.23.10\env\lib\site-packages\torch\nn\modules\module.py", line 1501, in _call_impl return forward_call(*args, **kwargs) File "C:\code\Warpfusion\0.23.10\generative-models\sgm\modules\encoders\modules.py", line 141, in forward emb_out = embedder(batch[embedder.input_key]) File "C:\code\Warpfusion\0.23.10\env\lib\site-packages\torch\nn\modules\module.py", line 1501, in _call_impl return forward_call(*args, **kwargs) File "c:\code\warpfusion\0.23.10\stablediffusion\modules\sd_hijack_clip.py", line 218, in forward batch_chunks, token_count = self.process_texts(texts) File "c:\code\warpfusion\0.23.10\stablediffusion\modules\sd_hijack_clip.py", line 194, in process_texts chunks, current_token_count = self.tokenize_line(line) File "c:\code\warpfusion\0.23.10\stablediffusion\modules\sd_hijack_clip.py", line 97, in tokenize_line tokenized = self.tokenize([text for text, _ in parsed]) File "c:\code\warpfusion\0.23.10\stablediffusion\modules\sd_hijack_clip.py", line 306, in tokenize tokenized = self.wrapped.tokenizer(texts, truncation=False, add_special_tokens=False)["input_ids"] File "C:\code\Warpfusion\0.23.10\env\lib\site-packages\transformers\tokenization_utils_base.py", line 2600, in __call__ if not self._in_target_context_manager: AttributeError: 'CLIPTokenizer' object has no attribute '_in_target_context_manager'

Bruno Ribeiro

yep, thanks !

WHAT IF Creative Studio

shorten the prompt

Warpfusion

hello, i'm getting this error: Traceback (most recent call last): File "C:\Users\WHATIFCreativeStudio\AppData\Local\Temp\ipykernel_28224\1505760836.py", line 625, in do_run() File "C:\Users\WHATIFCreativeStudio\AppData\Local\Temp\ipykernel_28224\2967362965.py", line 1329, in do_run sample, latent, depth_img = run_sd(args, init_image=init_image, skip_timesteps=skip_steps, H=args.side_y, File "C:\Users\WHATIFCreativeStudio\AppData\Local\Temp\ipykernel_28224\2967362965.py", line 2666, in run_sd c = sd_model.get_learned_conditioning(prompts) File "C:\Users\WHATIFCreativeStudio\AppData\Local\Temp\ipykernel_28224\3486811567.py", line 853, in get_unconditional_conditioning_sdxl c = self(batch_c) File "D:\1-SD\warpfusion\0.23\env\lib\site-packages\torch\nn\modules\module.py", line 1501, in _call_impl return forward_call(*args, **kwargs) File "D:\1-SD\warpfusion\0.23\generative-models\sgm\modules\encoders\modules.py", line 168, in forward output[out_key] = torch.cat( RuntimeError: Sizes of tensors must match except in dimension 2. Expected size 154 but got size 77 for tensor number 1 in the list. it happens when i run a sdxl model, even the default. my image is 1024x576. any hints?

WHAT IF Creative Studio

Just message or tag me on discord, though it seems I've replied to you already there. You need to check you have the correct model_version selected for the checkpoint or lora you are trying to use, also make sure you're not loading lora instead of a base model

Warpfusion

https://drive.google.com/file/d/1LEkp6tjfeN_knphgHxXUNLC-mRKyP9Ie/view?usp=sharing I have problem with rendering video, i'm getting dunno how to describe images, you can see it if u click on this link. i have tried on Discord, but it seems NO ONE is interested to give free help. I also saw on Discord some people offers some paid help. So i just want to know should i go with paid help???

Nenad Kuzmanovic

no it's not, there are some nodes available for comfy with a workflow in a separate post here, but that's it

Warpfusion

Hey alex, is stable warp fully compatible with comfy ? have you a install tutorial or something ? all the new update are aveilable if we use comfy ? thanks in advance

fabdream.ai

You can join our discord, and look for noisy outputs. It's most likely incorrect model version selected for your checkpoint

Warpfusion

bro, do you speak Russian? i have problem with rendering a video. when it comes to the "Do the run" item, there is no image in the preview pictures. just static, like from a broken TV. what is the problem? and how to fix it? if you have a telegram or insta, I can send you a screenshot

Идеи на миллион

Sure, there is a thread related to Loras on our discord, have you checked it?

Warpfusion

Is there any tutorial on how we can use lora?

N pizza

perfect！

martinej Hoccus

working perfect, amazing Dw pose

Diego Mellado Alarcon

Solid fix Sxela, working much better

John Walters

have you created the env via venv guide?

Warpfusion

First time running warp, and getting this error from within pdidinet module: ModuleNotFoundError: No module named 'basicsr'

indievish

Fire

Digital Grails

DW Pose is a great addition!

John Walters