search

Rudrabha / Wav2Lip

This repository contains the codes of "A Lip Sync Expert Is All You Need for Speech to Lip Generation In the Wild", published at ACM Multimedia 2020. For HD commercial model, please try out Sync Labs
https://synclabs.so
9.74k stars 2.11k forks source link

wav2lip高清模型训练交流：已经收敛到0.28 #614

Open zhangziliang04 opened 6 months ago

zhangziliang04 commented 6 months ago

image

wav2lip高清模型训练288x288,400K,以后已经收敛到0.28一下，不知道还需要多久。大家一起交流一下。

zhangziliang04 commented 6 months ago

image

Louis24 commented 6 months ago

neuroncat +wo吧

EricKong1985 commented 5 months ago

image

wav2lip高清模型训练288x288,400K,以后已经收敛到0.28一下，不知道还需要多久。大家一起交流一下。

请问你已经训练了多久？

einsqing commented 5 months ago

image

wav2lip高清模型训练288x288,400K,以后已经收敛到0.28一下，不知道还需要多久。大家一起交流一下。

不是看 eval loss 么？你看 train loss？

vmandrilly commented 5 months ago

有训练好的麻烦说以下，看起来不少人希望能使用中文的。 If someone has successfully trained a Chinese Mandarin model, please say so, quite some people are looking for this. Thanks.

mayu123mayu commented 5 months ago

看 eval loss 我train都 0.001了

TzyTman commented 4 months ago

如何训练？

sylyt62 commented 4 months ago

20240305-151738

Same here... Trained on 4090 for 7 days. @zhangziliang04

Any tips to get a further drop?

My wechat: tybs8k

keminze commented 3 months ago

20240305-151738

同样在这里...在 4090 上训练了 7 天。@zhangziliang04

有什么技巧可以进一步下降吗？

我的微信：tybs8k

你使用的数据集是什么，是否训练了判别器，对模型有什么调整吗。我在训练判别器时，损失无法下降，一直保持在0.5以上

luosiwu commented 3 days ago

微信 ZORO_Sir ，可以交流下吗