Results 10 issues of WGS

Hi, how can I find the maximum or minimum value of a field in a collection? thanks.

enhancement

hi大家好,请教下 太乙stable diffusion怎么加速推理,半精度差不多12s左右。 目前在试的方法是:将vae、unet等都转onnx,然后再用StableDiffusionOnnxPipeline读。 除了这个还有什么方法呢?

修改model_meta.json里的`model_file`改模型backbone,然后在lfw上测试效果很差,这是为什么呢? 0.19阈值: 0.5005008347245409 0.2阈值: 0.5003338898163606 0.21阈值: 0.5006677796327212 0.22阈值: 0.5003338898163606 0.23阈值: 0.5005008347245409 0.24阈值: 0.5010016694490818 0.25阈值: 0.5013355592654424 0.26阈值: 0.5020033388981636 0.27阈值: 0.5023372287145242 0.28阈值: 0.5028380634390651 0.29阈值: 0.5031719532554257 0.3阈值: 0.5030050083472454 0.31阈值: 0.5038397328881469 0.32阈值: 0.503338898163606 0.33阈值:...

多卡训练需要去掉混合精度,否则会:RuntimeError: expected scalar type Half but found Float,因为模型是Half加载的

大家好,请教大家问题: 我在训练pixel自回归模型时,遇到以下两个问题: When I was training a pixel autoregressive model, I encountered the following two problems: 1. gated pixelcnn 非常容易过拟合,在1、2个epoch之后,val loss就一直上升不降。 Gated pixelcnn is very easy to overfit. After 1...

大家好,请教大家问题: 我在训练pixel自回归模型时,遇到以下两个问题: When I was training a pixel autoregressive model, I encountered the following two problems: 1. gated pixelcnn 非常容易过拟合,在1、2个epoch之后,val loss就一直上升不降。 Gated pixelcnn is very easy to overfit. After 1...

你好,想问一下:有没有微调的代码?有没有论文?想学习具体是怎么实现的

documentation
algorithm

hello,想问一下,可以支持多少种音色?

documentation

### Self Checks - [X] I have thoroughly reviewed the project documentation (installation, training, inference) but couldn't find any relevant information that meets my needs. [English](https://speech.fish.audio/) [中文](https://speech.fish.audio/zh/) [日本語](https://speech.fish.audio/ja/) [Portuguese (Brazil)](https://speech.fish.audio/pt/)...

enhancement

arxiv过来的,非常感谢作者团队的开源! 我有两个问题想请教作者团队,期待作者团队的回复: 1. 训练代码有计划开源吗? 2. 实时推理的时候,能否到毫秒级别? 再次感谢作者团队的贡献!