Thank you for all theses instructions. I tried it, it works. But, that's so slow on CPU ! But I have a awkward message after loading the model : "It seems to be an instruction-following model with template “Llama-v2”. In the chat tab, instruct or chat-instruct modes should be used." Anyway, it just works ! thanks to your guidance !
Thanks! Yeah the message is because Code Llama is not a native "chat" model so you should select that to make it work perfectly (when you give it more than one message otherwise it will work anyway)
Well, this is not a video on image generation so there is no use in that, unless you have alot of imagination 🤣. The actual use of code llama is the one I've explained in the video: code generation and everything related to it!
I did try something similar to this, but it seems to take a very long time, in the end I just went onto discord and used one of those MANY bots. It would be nice to run local, I will give this one a try. They can't write (good) music yet, but AI nets are quite good at helping with things like chord progression, having some success with that! (not sold any of my music yet, but reached 100 subscribers...yay!)
Congrats for your 100 subscribers!👍 Yeah they tend to be bad in music for now, the only thing i tried that seems to be at least acceptable to get new ideas is AudioCraft (i covered that in the first video i made on this channel! ruclips.net/video/YIjDYJQhIio/видео.html )
Okay you got me. I'm gunna try this right now. Be back with a report in a few hours.🤠
A few hours later ... haha :D that meme
Thanks for this one!! 😊
Thank you for all theses instructions. I tried it, it works. But, that's so slow on CPU !
But I have a awkward message after loading the model : "It seems to be an instruction-following model with template “Llama-v2”. In the chat tab, instruct or chat-instruct modes should be used."
Anyway, it just works ! thanks to your guidance !
Thanks! Yeah the message is because Code Llama is not a native "chat" model so you should select that to make it work perfectly (when you give it more than one message otherwise it will work anyway)
I didn't know ComfyUI could do such work, this is cool. Does this kind of model also benefit from powerful GPU like SD model? Thanks for sharing this.
Yes, the more powerful your GPU is ( and the more VRAM you have), the more you can increase the GPU layers so that it will give you a faster answer!
I am curious, What is the actual use of it and how can this make my Image generation or Video morph better?
Well, this is not a video on image generation so there is no use in that, unless you have alot of imagination 🤣. The actual use of code llama is the one I've explained in the video: code generation and everything related to it!
I did try something similar to this, but it seems to take a very long time, in the end I just went onto discord and used one of those MANY bots. It would be nice to run local, I will give this one a try.
They can't write (good) music yet, but AI nets are quite good at helping with things like chord progression, having some success with that! (not sold any of my music yet, but reached 100 subscribers...yay!)
Congrats for your 100 subscribers!👍 Yeah they tend to be bad in music for now, the only thing i tried that seems to be at least acceptable to get new ideas is AudioCraft (i covered that in the first video i made on this channel! ruclips.net/video/YIjDYJQhIio/видео.html )