Improve the model
Tdrussell stated on the NoobAI Discord that he simply doesn't have the money currently to self fund an Anima 2 (or whatever other new version). In general I feel like this post is pretty entitled as we finally get an anime model after so many years of being stuck with bad local models and SDXL in general... only to then say that the model sucks because we aren't getting an update for a model that isn't even a year old, as if training costs don't exist. Anima is like straight up the best anime model that we have, at this point just pay for cloud models instead lol
Also, in regards to the text encoder, you can just look at an earlier discussion on it where it was explained why it wasn't increased from 0.6B to 2B or 4B.
It is the best anime template because there is no competitor in the first place because SDXL and the like are ahead of Wattedit, but frankly, in the beginning the template was great because you can use Natural Language and so on, but understanding it is difficult, more difficult than Krea, for example, so I hope that after this, in the next updates, its intelligence will increase a little or the results will improve because I love anime and I use it because the loras in it are better, but it still needs improvements, so I hope that it will It will improve at the beginning of 2027
Regarding the text encoder discussion, I did not find it. Can you provide a link or something?
It is the best anime template because there is no competitor in the first place because SDXL and the like are ahead of Wattedit, but frankly, in the beginning the template was great because you can use Natural Language and so on, but understanding it is difficult, more difficult than Krea, for example, so I hope that after this, in the next updates, its intelligence will increase a little or the results will improve because I love anime and I use it because the loras in it are better, but it still needs improvements, so I hope that it will It will improve at the beginning of 2027
Regarding the text encoder discussion, I did not find it. Can you provide a link or something?
In the Civitai comments section, they stated that they currently don't have enough funds, so they will continue R&D on Turbo and reinforcement learning until they raise enough money. For now, all we can do is wait and hope they can secure enough funding as soon as possible.
It is the best anime template because there is no competitor in the first place because SDXL and the like are ahead of Wattedit, but frankly, in the beginning the template was great because you can use Natural Language and so on, but understanding it is difficult, more difficult than Krea, for example, so I hope that after this, in the next updates, its intelligence will increase a little or the results will improve because I love anime and I use it because the loras in it are better, but it still needs improvements, so I hope that it will It will improve at the beginning of 2027
Regarding the text encoder discussion, I did not find it. Can you provide a link or something?
The specific comment is here: https://huggingface.co/circlestone-labs/Anima/discussions/67#69af073d3ffb4d937d386f4b
Also there being no competitor is exactly why we were stuck on SDXL in the first place. There's far less interest in anime let alone the money and training knowledge to pull it off. That's why I think it's kinda crazy to act like this over a model which was released only around 4 months ago and expecting money to just continuously be spawned from thin air for full new models.
It is the best anime template because there is no competitor in the first place because SDXL and the like are ahead of Wattedit, but frankly, in the beginning the template was great because you can use Natural Language and so on, but understanding it is difficult, more difficult than Krea, for example, so I hope that after this, in the next updates, its intelligence will increase a little or the results will improve because I love anime and I use it because the loras in it are better, but it still needs improvements, so I hope that it will It will improve at the beginning of 2027
Regarding the text encoder discussion, I did not find it. Can you provide a link or something?
In the Civitai comments section, they stated that they currently don't have enough funds, so they will continue R&D on Turbo and reinforcement learning until they raise enough money. For now, all we can do is wait and hope they can secure enough funding as soon as possible.
Okay, your point is valid. I didn't know the funding was so limited, so I hid my initial comment.
We hope it will develop within two years.
Tdrussell stated on the NoobAI Discord that he simply doesn't have the money currently to self fund an Anima 2 (or whatever other new version). In general I feel like this post is pretty entitled as we finally get an anime model after so many years of being stuck with bad local models and SDXL in general... only to then say that the model sucks because we aren't getting an update for a model that isn't even a year old, as if training costs don't exist. Anima is like straight up the best anime model that we have, at this point just pay for cloud models instead lol
Also, in regards to the text encoder, you can just look at an earlier discussion on it where it was explained why it wasn't increased from 0.6B to 2B or 4B.
Hi, I love this model. Where can I donate or contribute to fund future updates?
Can't they just ask comfyUI for more money
What is the NoobAI Discord?