-
-
Hi, very interesting work and nice presentation in paper. I'm interested in applying conditional flow matching to the tasks without paired training data. For example, many image restoration tasks do n…
-
## 論文タイトル(原文まま)
Matcha-TTS: A fast TTS architecture with conditional flow matching
## 一言でいうと
Matcha-TTSは、条件付きフローマッチングを使用した高速で高品質なテキスト読み上げ(TTS)モデルです。
### 論文リンク
[https://arxiv.org/abs/2309.0319…
-
Thanks very much for this code base, it's been a great way to learn about flow matching. I have a question regarding conditional generation with OT-CFM.
When testing different FM approaches on my o…
-
I am getting a lot of help from this project. Thank you! I have one question. Is the method for generating samples(ex images) with models trained using the FM framework the same as that of basic CFM a…
-
Hi, when I tried to continue the training on the Conditional flow-matching on a new dataset (zh collected from youtube), I found that the loss degradee a lot but the generate audio is totally unintell…
-
`torchdyn>=1.0.3` breaks if a model does not have *args, **kwargs in the `forward` method. This breaks downstream code which has custom models compatible with previous torchdyn versions. I propose to …
-
Dear experts,
Thank you very much for this nice package! I am trying to use your batch OT version of flow matching to perform the matching between two distributions, but I wonder how the OT mapping…
-
Thank you for your clean codes. Have you finished train this model on CelabA dataset. What about thi results .
-
The parameter `num_images` is currently unused in the cifar10 example https://github.com/atong01/conditional-flow-matching/blob/21cd0c888186f6e2b76deb393800361b8a850e9b/examples/cifar10/train_cifar10…
QB3 updated
10 months ago