[BUG] vqgan_finetune 为什么会导致SSH到服务器的连接中断?
aixiaodewugege opened this issue · comments
数据集是这样的,我按照教程微调gan的时候,每次都会中断和服务器的连接,训练也直接中止
[2024-05-08 07:29:14,268][main][INFO] - [rank: 0] Resuming weights only!
[2024-05-08 07:29:21,879][main][INFO] - [rank: 0] Error loading state dict: _IncompatibleKeys(missing_keys=['vocoder.backbone.downsample_layers.0.0.weight', 'vocoder.backbone.downsample_layers.0.0.bias', 'vocoder.backbone.downsample_layers.0.1.weight', 'vocoder.backbone.downsample_layers.0.1.bias', 'vocoder.backbone.downsample_layers.1.0.weight', 'vocoder.backbone.downsample_layers.1.0.bias', 'vocoder.backbone.downsample_layers.1.1.weight', 'vocoder.backbone.downsample_layers.1.1.bias', 'vocoder.backbone.downsample_layers.2.0.weight', 'vocoder.backbone.downsample_layers.2.0.bias', 'vocoder.backbone.downsample_layers.2.1.weight', 'vocoder.backbone.downsample_layers.2.1.bias', 'vocoder.backbone.downsample_layers.3.0.weight', 'vocoder.backbone.downsample_layers.3.0.bias', 'vocoder.backbone.
估计你内存爆了, 需要查询其他日志.