Erickrus

[email protected]

Shanghai, China wechat: erickrus79

Results 21 comments of


                                            Erickrus

EDVR colab?

> Hello! Tried to launch it with EDVR_L_x4_SR_REDS_official-9f5f5039.pth and have got UnpicklingError: invalid load key, '

希望未来可以支持上海话~~

你可以录制一个沪普，是可以实现的

Training time is too long

Is there a way to speed up the training process, with only 1 GPU ?

如何提高文本转语音的准确性?

仔细看了一下，预处理部分没有bug，norm_text也是对的。有问题的部分应该是AR(GPT，t2s_model)中输出的pred_semantic。这个地方是唯一通过自回归计算，生成可变长度的地方。输入是phone，输出是semantic frames。如果要修复这个问题，感觉是在finetuning阶段引入的。

请问带有音频的视频怎么合成？

use ffmpeg to merge wave and video

Real-time execution

I ran the latest version on colab (--torso, model engm, [downscale](https://github.com/ashawkey/RAD-NeRF/blob/0de5ed259592592294677ad6cf7605f478a0de57/test.py#L145) ) NVIDIA-SMI Driver Version: 525.85.12 CUDA Version: 12.0, Tesla T4 15360MiB The performance is: | fps | size |...

Implement invisible watermark according to the settings

So, there's no watermark now?

Overfitting

@NikitaKononov did you finally resolve this issue and achieve a lower loss ?

GPU is not used after model is loaded

I'm using T4 GPU. The same as above. Only GPU RAM 0.1 GB is used ```GPU RAM 0.1 / 15.0 GB```

怎么从你们的检查点启动呢

请问有谁把finetuning 流程跑完了 ?

1
2
3
›