Superb improvements! would you mind helping your mac users with an installation guide for mac, please? I've downloaded the TOX but I'm unable to make it run in my M1 Max MBP. Thanks in advance :D
I executed the start stream of settings1, and it would crash after the cmd window appeared. "launching StreamDiffusionTD" was displayed below settings1, and the status displayed "me.par.streamactive else 'not Streaming". I tried repeatedly to install the components described in the file. , this problem keeps appearing, is there any solution?
How do you access the prompt located at 9:44? I can't get it to change to "reference" ruclips.net/video/F2JSWAKex4k/видео.htmlsi=upI5D-ytOoGG-5Jl&t=584
@@33JBC seems like the edit might have skipped showing that! You want to click and drag the text par from the text1 to the prompt parameter. The final reference on the parameter should be like op(‘text1’).par.text
Hello, I would like to know how to use this software Lora. I used Lora in stablediffusion, but it keeps getting errors and cannot be used. I am not sure if it is a problem with the settings or the Lora version.
This is likely due to model version mismatch between the main sd model and the LoRA model. sd-turbo is 2.1 based and won't work with 1.5 or sdxl LoRAs. I might be exploring or posting more on training a sd-turbo lora for streamdiffusion soon
Incredible work & such a useable setup, thanks for sharing! I am running into this error alot when trying to experiment with other resolutions / ratios: "ValueError: cannot reshape array of size 786432 into shape (720,1280,3)". Any advice greatly appreciated
Hi, thank you for this fantastic tutorial and new SD 2.0. I find that ingeneral are very low quality images specially when representing people. Do you have a tutorial for getting more resolution, and better faces, people in general?. Some of my friends said missing Triton causes the generate result to be blurry at sometimes.
excuse me when i clicked the "download streamdiffusion", it's said that no internet connection detected, but all of those wlan settings are correct, could you help me? i can't go to the next step due to that
Hi I've got the same problem here. It is said that if you are in china mainland(forgive my presumption), you can set your vpn to global mode and try again. (it didn't work for me though😥)
Do you foresee an update for ComfyTD on the horizon or is it currently feature complete? I'd have asked on your Patreon, but as I'm not a paying member yet, I'm not allowed to ask questions there.
@@Quantumspace23 yes definitely. I’ve got a nice all around update for ComfyTD. It is available on my discord but by next week it should be available on Patreon with changelog etc.
Thanks for your work! It's so good. I have an Alienware m15 r7a with: RTX 3080 gpu, 12th Gen Intel(R) Core(TM) i7 2.30 GHz, 32G RAM and a x64-based processor Do you think she can handle it? Video out needs to run at 60 fps and be 1080x1920. What do you think? Thank u so much
cool... this is the error i get tho, after click on start stream. OSError: Error no file named pytorch_model.bin, model.safetensors, tf_model.h5, model.ckpt.index or flax_model.msgpack found in directory C:\Users\babee\.cache\huggingface\hub\models--stabilityai--sd-turbo\snapshots\b261bac6fd2cf515557d5d0707481eafa0485ec2\text_encoder. Failed to load model due to: Error no file named pytorch_model.bin, model.safetensors, tf_model.h5, model.ckpt.index or flax_model.msgpack found in directory C:\Users\babee\.cache\huggingface\hub\models--stabilityai--sd-turbo\snapshots\b261bac6fd2cf515557d5d0707481eafa0485ec2\text_encoder. the set hugging face cache is on... do i need to manually download the models?
@@spiritform111 It is easiest to look at these issues on discord. I've got a forum setup for you to post screenshots or error messages. I'm happy to follow up there if you make a post.
Hi, probably as very silly question as I'm new to this but how do I add python, cuda TOOLKIT and git to path? just want to make sure I'm doing this right.
you’re an angel in human form, Dot. Thank you!!!
Love where this is going!
legend. A bit of instruction on how to downlaod the various other models would be cool too!
Definitely planning on breaking that down in future video.
wow this is epic! will you be adding support for mac?
Superb improvements! would you mind helping your mac users with an installation guide for mac, please? I've downloaded the TOX but I'm unable to make it run in my M1 Max MBP. Thanks in advance :D
I executed the start stream of settings1, and it would crash after the cmd window appeared. "launching StreamDiffusionTD" was displayed below settings1, and the status displayed "me.par.streamactive else 'not Streaming". I tried repeatedly to install the components described in the file. , this problem keeps appearing, is there any solution?
Wow ! Pretty impressive.
super awsome!
Hiii, thank you very much for the tutorial, I would like to ask is there any tutorial for mac streamdiffusion?
hey Dot, does it have an AMD GPU version? where can I skip cuda test? it is necessary?
How do you access the prompt located at 9:44? I can't get it to change to "reference" ruclips.net/video/F2JSWAKex4k/видео.htmlsi=upI5D-ytOoGG-5Jl&t=584
use: op('prompt').par.text and make sure you rename the text operator "prompt"
@@33JBC seems like the edit might have skipped showing that! You want to click and drag the text par from the text1 to the prompt parameter. The final reference on the parameter should be like op(‘text1’).par.text
Hi, is that possible to run stream diffusion on GCP? cuz I see it's so slow to run it on Mac
wonder the same thing, did you figure it out?
Hello, how to install on Mac osx sonoma 14? M3
Any chance you'd share your pipeline python code? Would love to see how you're connecting all these TD controls to StreamDiffusion.
Thanks for this!
@@__pawpaw__ yes! All of it is inside the operator and can be explored. Most of the connection between python server and TD is shared memory and osc.
Can i install It to my mbp 4pro ?
Hi, I cant download the StreamDiffusionTD TOX, it says I have to upgrade my subscription
I just double checked. Should be available for all paid tiers.
@dotsimulate thanks, because I'm not a member. I thought it's free
Hello, I would like to know how to use this software Lora. I used Lora in stablediffusion, but it keeps getting errors and cannot be used. I am not sure if it is a problem with the settings or the Lora version.
This is likely due to model version mismatch between the main sd model and the LoRA model. sd-turbo is 2.1 based and won't work with 1.5 or sdxl LoRAs.
I might be exploring or posting more on training a sd-turbo lora for streamdiffusion soon
@ Thank you very much. I'm looking forward to your subsequent tutorials.
MIDI support would be epic!!,
you can do that already in touch and just connect whichever parameter you like in sd
Incredible work & such a useable setup, thanks for sharing! I am running into this error alot when trying to experiment with other resolutions / ratios: "ValueError: cannot reshape array of size 786432 into shape (720,1280,3)". Any advice greatly appreciated
Hi, thank you for this fantastic tutorial and new SD 2.0. I find that ingeneral are very low quality images specially when representing people. Do you have a tutorial for getting more resolution, and better faces, people in general?. Some of my friends said missing Triton causes the generate result to be blurry at sometimes.
thanks !!
Hello! Excuse me does this work with touch designer 2022?
Unfortunately not backwards compatible. It needs 2023 due to being built in it and also relying on the sequence parameters.
excuse me when i clicked the "download streamdiffusion", it's said that no internet connection detected, but all of those wlan settings are correct, could you help me? i can't go to the next step due to that
Hi I've got the same problem here. It is said that if you are in china mainland(forgive my presumption), you can set your vpn to global mode and try again. (it didn't work for me though😥)
Can i use it with the free version of touchdesigner 2023?
Yep!
@@dotsimulate Great :)
Does this restrict the resolution under 516? Because I can not add a bigger output.
you can do bigger outputs of you don’t use tensorrt etc
How can I increase fps to improve flow?
Do you foresee an update for ComfyTD on the horizon or is it currently feature complete? I'd have asked on your Patreon, but as I'm not a paying member yet, I'm not allowed to ask questions there.
@@Quantumspace23 yes definitely. I’ve got a nice all around update for ComfyTD. It is available on my discord but by next week it should be available on Patreon with changelog etc.
@@dotsimulate Well then, I guess that seals it; I have no choice but to sign up for the next level of participation to get on that Discord channel.
와우 엄청신기한걸
❤❤❤
Thanks for your work! It's so good. I have an Alienware m15 r7a with: RTX 3080 gpu, 12th Gen Intel(R) Core(TM) i7 2.30 GHz, 32G RAM and a x64-based processor
Do you think she can handle it? Video out needs to run at 60 fps and be 1080x1920. What do you think? Thank u so much
When do you think you’ll have a version ready for Mac? 😊
@@DrewMolleur we will need to look at this specifically again. But I’d like to get a solid Mac installation guide out soon.
@@dotsimulate pls pls we really need a solid Mac installation guide out soon.
@@dotsimulate pls!
cool... this is the error i get tho, after click on start stream.
OSError: Error no file named pytorch_model.bin, model.safetensors, tf_model.h5, model.ckpt.index or flax_model.msgpack found in directory C:\Users\babee\.cache\huggingface\hub\models--stabilityai--sd-turbo\snapshots\b261bac6fd2cf515557d5d0707481eafa0485ec2\text_encoder.
Failed to load model due to: Error no file named pytorch_model.bin, model.safetensors, tf_model.h5, model.ckpt.index or flax_model.msgpack found in directory C:\Users\babee\.cache\huggingface\hub\models--stabilityai--sd-turbo\snapshots\b261bac6fd2cf515557d5d0707481eafa0485ec2\text_encoder.
the set hugging face cache is on... do i need to manually download the models?
following the exact video instructions.
i think i may have figured it out... hugging face model downloader. missed that step!
nope.
@@spiritform111 It is easiest to look at these issues on discord. I've got a forum setup for you to post screenshots or error messages. I'm happy to follow up there if you make a post.
@@dotsimulate ok, thank you!
does this work on mac
Can we get a controlnet tutorial next 🙏
@@rajnaikgaonkar3910 yes!
Hi, probably as very silly question as I'm new to this but how do I add python, cuda TOOLKIT and git to path? just want to make sure I'm doing this right.
Sorted.😀
thanks!