Merge SAEHD error – Help!!

Home Forums DeepFaceLab Errors Merge SAEHD error – Help!!

Viewing 4 posts - 1 through 4 (of 4 total)
  • Author
  • #8285

      when I proceed with the SAEHD merge, this error appears, how to fix it?

      Running merger.

      Choose one of saved models, or enter a name to create a new model.
      [r] : rename
      [d] : delete

      [0] : TRAIN ALICE V1 – latest
      : 0
      Loading TRAIN ALICE V1_SAEHD model…

      Choose one or several GPU idxs (separated by comma).

      [CPU] : CPU
      [0] : NVIDIA GeForce RTX 3090

      [0] Which GPU indexes to choose? : 0

      Traceback (most recent call last):
      File “E:\DFL\DeepFaceLab_NVIDIA_RTX3000_series\_internal\DeepFaceLab\mainscripts\”, line 53, in main
      File “E:\DFL\DeepFaceLab_NVIDIA_RTX3000_series\_internal\DeepFaceLab\models\”, line 180, in __init__
      File “E:\DFL\DeepFaceLab_NVIDIA_RTX3000_series\_internal\DeepFaceLab\models\Model_SAEHD\”, line 181, in on_initialize_options
      raise Exception(“pretraining_data_path is not defined”)
      Exception: pretraining_data_path is not defined


        Yes, I get the same error – did you manage to resolve it, and if so, how?!


          Same here. Can anyone help?


            I had a similar problem. Here’s how I solved it: I deleted all Python-related versions in Windows 11 installed applications. The model already has its own “python-3.6.8” files in the “DeepFaceLab_NVIDIA_RTX3000_series\_internal” path. Python versions installed on Windows 11 break the pyton version of the model. I think the problem will be solved if you remove the Python-related versions of Windows installed applications. However, considering every possibility, back up the installation files of these Python version files to your computer before removing the installation files of Python versions.

            On Windows, install only the python-3.6.8 version that is the same as the python version of the model.

          Viewing 4 posts - 1 through 4 (of 4 total)
          • You must be logged in to reply to this topic.