Where can we find this workflow, your civitAI page? Also did you have the workflow for the live portrait that the other creator mashed together that you showed us in the video? Thank you
ah if you are referring to the "complex workflows" preview, this was not made by me. The creator is on openart.ai/workflows/all let me know if you cannot find it
Since this video came out (and you might notice i said that i will cover song2song in the follow up video) the audio loader node stopped working. I think it's a conflict with an existing node in ComfyUI using the same name. I'm hoping the author is able to solve this problem. As soon as something happens, expect a new video on Lyra.
Does anyone know how to save seed and date information in the audio filename prefix? I don't know why but {seed} or %KSampler.seed% doesn't work. They worked for save image but not working for save audio.
@@FiveBelowFiveUK Thank you for the tip. However, this only works for the 'Save Image' node. %KSampler.seed% didn't work in 'SaveAudio'. Searched for several hours but no luck yet. My last hope is to find the file that provides 'SaveAudio' node, but I couldn't find it. Btw, thank you for the video though. It's helpful!
i'm working on a "long haul up the mountain" so we have not started with LLM in comfyui yet, however you can do this and there are many custom nodes allowing that. Regarding FL studio plugins, i'm sure there are some already, but likely they are standalone and not relating to comfyui yet. There is always a chance somebody made it and i never saw it yet. Some new image models we will look at soon, use LLM for the text encoder, so it will be introduced naturally as we go along...
@@FiveBelowFiveUK thx I realy ment just fruity loops. about some existence of llm implementation into comfy i am aware. it looks like there starts to be some models which better understands the normal prose like sd3...
@@yngeneer after reading your comment, I did some digging and turned up an SDXL based image generation model named Kolors. It used the 8bit quantized GLM3 LLM as the text encoder and seems vastly superior to the SDXL pair of CLIP encoders. I have published the V2 update already and the results look great. Looks like you are onto something there. you can check the video here: ua-cam.com/video/QbUs9_eYgLM/v-deo.html
@@FiveBelowFiveUK I'm afraid I can't afford to watch that :( @ the FL Studio stuff... or from the other side - it would be nice to have something similar to ComfyUI, but focused on audio production instead of graphical sphere so to speak... ....after several minutes... ....ohh....lol, the video is public now, hah.... I don't know what to say....was it because of me, or was it a plan of yours from the beginnings to publish it...? Nevermind, it's more rhetorical question than anything else....
@@yngeneer The memberships are new, that was the first video i put on early access. my only complaint with Audio generation in Comfyui is that it is not supporting many audio types. I'm sure now stable audio is here it will be easier in future.
Pretty cool, soon we will have full blown 1-man animation studios 👀?
I think we're there already. Just depends how many hours you want or can put into study, hunting tools, and building. Then ofc the creative part.
Where can we find this workflow, your civitAI page?
Also did you have the workflow for the live portrait that the other creator mashed together that you showed us in the video? Thank you
ah if you are referring to the "complex workflows" preview, this was not made by me. The creator is on openart.ai/workflows/all
let me know if you cannot find it
not able to download and find VAEencodeAudio and import failed for "ComfyUI Stable Audio Open 1.0 Sampler", any suggestions?
Since this video came out (and you might notice i said that i will cover song2song in the follow up video) the audio loader node stopped working. I think it's a conflict with an existing node in ComfyUI using the same name. I'm hoping the author is able to solve this problem. As soon as something happens, expect a new video on Lyra.
@@FiveBelowFiveUK thanks! I managed to install VAE encode/decode audio.
I'm also having problems with audioloader, I hope they fix the problem
thank, it is a great help
You're welcome!
thank you I was able to make 1000 bucks at home
Excellent -- nice work !!
Really? How'd you do it? Just curious
Does anyone know how to save seed and date information in the audio filename prefix? I don't know why but {seed} or %KSampler.seed% doesn't work. They worked for save image but not working for save audio.
this is a good resource on the topic.
blenderneko.github.io/ComfyUI-docs/Interface/SaveFileFormatting/
@@FiveBelowFiveUK Thank you for the tip. However, this only works for the 'Save Image' node. %KSampler.seed% didn't work in 'SaveAudio'. Searched for several hours but no luck yet. My last hope is to find the file that provides 'SaveAudio' node, but I couldn't find it.
Btw, thank you for the video though. It's helpful!
does have fruity loops any AI, or LLM plugin yet?
i'm working on a "long haul up the mountain" so we have not started with LLM in comfyui yet, however you can do this and there are many custom nodes allowing that. Regarding FL studio plugins, i'm sure there are some already, but likely they are standalone and not relating to comfyui yet. There is always a chance somebody made it and i never saw it yet. Some new image models we will look at soon, use LLM for the text encoder, so it will be introduced naturally as we go along...
@@FiveBelowFiveUK thx I realy ment just fruity loops. about some existence of llm implementation into comfy i am aware. it looks like there starts to be some models which better understands the normal prose like sd3...
@@yngeneer after reading your comment, I did some digging and turned up an SDXL based image generation model named Kolors. It used the 8bit quantized GLM3 LLM as the text encoder and seems vastly superior to the SDXL pair of CLIP encoders. I have published the V2 update already and the results look great. Looks like you are onto something there. you can check the video here: ua-cam.com/video/QbUs9_eYgLM/v-deo.html
@@FiveBelowFiveUK I'm afraid I can't afford to watch that :( @ the FL Studio stuff... or from the other side - it would be nice to have something similar to ComfyUI, but focused on audio production instead of graphical sphere so to speak...
....after several minutes...
....ohh....lol, the video is public now, hah.... I don't know what to say....was it because of me, or was it a plan of yours from the beginnings to publish it...? Nevermind, it's more rhetorical question than anything else....
@@yngeneer The memberships are new, that was the first video i put on early access.
my only complaint with Audio generation in Comfyui is that it is not supporting many audio types. I'm sure now stable audio is here it will be easier in future.
This will be amazing once it's at the level of suno.
yes, BML have been making teasers about Flux Video - so hopefully we don't wait too long :)