Skip to content

Latest commit

 

History

History
116 lines (69 loc) · 3.66 KB

README.md

File metadata and controls

116 lines (69 loc) · 3.66 KB
Raw

AIDesign-GAN Export Folder

A folder that holds the subfolders and files that an AIDesign-GAN export needs.

Documentation Files

Texts and images.

README.md

This file itself.

generator_preview.jpg

Generator preview.

A preview of the images the generator can generate.

Configuration Files

Texts.

discriminator_config.json

Discriminator configuration.

Automatically configured. Not supposed to be edited manually.

Configuration items. Type dict[str, typing.Union[dict, list, str, bool, int, float, None]].

Configuration item descriptions are listed below.

  • image_resolution. Input image resolution in pixels. Type int. Range [1, ).
  • image_channel_count. Input image channel count. Type int. Range [1, ).
    • label_resolution. Output label resolution in pixels. Type int. Range [1, ).
  • label_channel_count. Output label channel count. Type int. Range [1, ).
  • feature_map_count Layer 0 (first layer) output feature map count. Type int. Range [1, ). Compatibility alias feature_map_size. Precedence feature_map_count > feature_map_size.
  • struct_name. Structure name. Type str.
  • state_script_name. Model state TorchScript name. Type str.
  • state_onnx_name. Model state ONNX name. Type str.

discriminator_struct.py

Discriminator structure.

Automatically configured. Not supposed to be edited manually.

Python code fragment. Uses information from the loaded modeler configuration at self.config to setup the targeted model structure at self.model.

self.config. Loaded modeler configuration. Type dict.

self.model. Targeted model structure to setup. type torch.nn.Module.

format_config.json

Format configuration.

Automatically configured. Not supposed to be edited manually. Serves as a model folder format reference.

Configuration items. Type dict[str, typing.Union[dict, list, str, bool, int, float, None]].

Configuration item descriptions are listed below.

  • aidesign_gan_version. Type str.
  • aidesign_gan_repo_tag. Type str.

generator_config.json

Generator configuration.

Automatically configured. Not supposed to be edited manually.

Configuration items. Type dict[str, typing.Union[dict, list, str, bool, int, float, None]].

Configuration item descriptions are listed below.

  • noise_resolution. Input noise resolution in pixels. Type int. Range [1, ).
  • noise_channel_count. Input noise channel count. Type int. Range [1, ).
  • image_resolution. Output image resolution in pixels. Type int. Range [1, ).
  • image_channel_count. Output image channel count. Type int. Range [1, ).
  • feature_map_count. Layer -1 (last layer) input feature map count. Type int. Range [1, ). Compatibility alias feature_map_size. Precedence feature_map_count > feature_map_size.
  • struct_name. Structure name. Type str.
  • state_script_name. Model state TorchScript name. Type str.
  • state_onnx_name. Model state ONNX name. Type str.
  • preview_name. Model preview name. Type str.

generator_struct.py

Generator structure.

Automatically configured. Not supposed to be edited manually.

Python code fragment. Uses information from the loaded modeler configuration at self.config to setup the targeted model structure at self.model.

self.config. Loaded modeler configuration. Type dict.

self.model. Targeted model structure to setup. type torch.nn.Module.

State Storage Files

Binaries.

discriminator_state.pt

Discriminator state.

generator_state.pt

Generator state.