DFL常见报错答疑搜集贴，不定时更新

飘渺狂客

AMP模型无法导出DFM
不知道有没有人遇到过

将 .dfm 文件转存到 M:\DFL_UPTO_RTX2080Ti\workspace\model\test_AMP_model.dfm
Dumping .dfm to M:\DFL_UPTO_RTX2080Ti\workspace\model\test_AMP_model.dfm
Failed to convert node 'Slice_3' (fct=<bound method Slice.version_1 of <class 'tf2onnx.onnx_opset.tensor.Slice'>>)
''
Traceback (most recent call last):
  File "M:\DFL_UPTO_RTX2080Ti\_internal\python-3.6.8\lib\site-packages\tf2onnx\tfonnx.py", line 294, in tensorflow_onnx_mapping
    func(g, node, **kwargs, initialized_tables=initialized_tables, dequantize=dequantize)
  File "M:\DFL_UPTO_RTX2080Ti\_internal\python-3.6.8\lib\site-packages\tf2onnx\onnx_opset\tensor.py", line 343, in version_1

3181622@qq.com

deeplovelyf 发表于 2022-3-18 16:51
Silent start: choosed model "猫之汉化神丹224 df-d"
Loading 猫之汉化神丹224 df-d_SAEHD model...
Silen ...

这个问题我也遇到，现在解决了。
找到deepfacelab/core/imagelib/warp.py这个文件，找到 line 146，这句：
    random_transform_mat = cv2.getRotationMatrix2D((w // 2, w // 2), rotation, scale)
把w // 2这里的2，都改成2.0

cat750515

感謝! 剛好繁體用戶遇到Error: 'cp950' 搜索找到這裡來!

hello5566880

支持，一直在找的报错汇总

cg110119

2H2048

请问这个问题怎么解决啊？
Error: No training data provided.
Traceback (most recent call last):
  File "Trainer.py", line 46, in Trainer.trainerThread
  File "D:\AIGC\DeepFaceLab_NV30_221111_CN\_internal\DeepFaceLab\models\ModelBase.py", line 193, in __init__
    self.on_initialize()
  File "D:\AIGC\DeepFaceLab_NV30_221111_CN\_internal\DeepFaceLab\models\Model_SAEHD\Model.py", line 684, in on_initialize
    generators_count=src_generators_count ),
  File "D:\AIGC\DeepFaceLab_NV30_221111_CN\_internal\DeepFaceLab\samplelib\SampleGeneratorFace.py", line 48, in __init__
    raise ValueError('No training data provided.')
ValueError: No training data provided.

806585962

谢谢，好好学习一下

fucktrade

这个板块太好了，别人踩过的坑，都填了

rockpayne

新人小白，还要从头开始学习啊

jeanimy

感谢分享