Search Results for "hours"

...The main contribution of the paper is a skip-layer excitation in the generator, paired with autoencoding self-supervised learning in the discriminator. Quoting the one-line summary "converge on single gpu with few hours' training, on 1024 resolution sub-hundred images". Augmentation is essential for Lightweight GAN to work effectively in a low data setting. You can test and see how your images will be augmented before they pass into a neural network (if you use augmentation). The general recommendation is to use suitable augs for your data and as many as possible, then after some time of training disable the most destructive (for image) augs. ...

...Specifically, we train a neural codec language model (called VALL-E) using discrete codes derived from an off-the-shelf neural audio codec model, and regard TTS as a conditional language modeling task rather than continuous signal regression as in previous work. During the pre-training stage, we scale up the TTS training data to 60K hours of English speech which is hundreds of times larger than existing systems. VALL-E emerges in-context learning capabilities and can be used to synthesize high-quality personalized speech with only a 3-second enrolled recording of an unseen speaker as an acoustic prompt. Experiment results show that VALL-E significantly outperforms the state-of-the-art zero-shot TTS system in terms of speech naturalness and speaker similarity. ...