Flow View#
Flow View module for pytorch model visualization.
- visualtorch.flow.flow_view(model, input_shape, input_dtype=None, to_file=None, min_z=10, min_xy=10, max_z=400, max_xy=2000, scale_z=0.1, scale_xy=1, type_ignore=None, color_map=None, palette='okabe_ito', low_dim_orientation='z', background_fill='white', draw_volume=True, padding=10, spacing=10, draw_funnel=True, shade_step=10, legend=False, font=None, font_color='black', opacity=255, show_dimension=False, level_gap=None, show_input=True, outline_width=1, connector_fill=None, connector_width=1, one_dim_orientation=None, legend_position='bottom-left')#
Generate a flow-style architecture visualization for a given torch model.
- Return type:
<module ‘PIL.Image’ from ‘/home/docs/checkouts/readthedocs.org/user_builds/visualtorch/envs/latest/lib/python3.11/site-packages/PIL/Image.py’>
- Deprecated:
Since version 1.4.0. Use visualtorch.render(model, input_shape, style=”flow”, …) instead - every parameter here is forwarded unchanged.
- visualtorch.flow.layered_view(*args, **kwargs)#
Pre-1.0 alias for flow_view.
Accepts the same arguments as flow_view, plus the removed index_ignore (silently dropped - it has no equivalent under the topology-based layout) and the renamed one_dim_orientation (mapped to low_dim_orientation). Positional calls using the old pre-1.0 argument order (which had index_ignore where flow_view now has color_map) will not be translated correctly - use keyword arguments to migrate safely.
- Return type:
<module ‘PIL.Image’ from ‘/home/docs/checkouts/readthedocs.org/user_builds/visualtorch/envs/latest/lib/python3.11/site-packages/PIL/Image.py’>
- Deprecated:
Since version 1.0.0. Use flow_view instead.