atoolkitdpt 0.0.2

Midas and Dense Prediction Transformers modules for Ambrosinus-Toolkit

Midas and DPT modules for Ambrosinus-Toolkit

This Module is part of Ambrosinus-Toolkit v1.2.9 (since 1.1.9)

This is a Python package for the "DPTto3D" component and "DPTSemSeg" component included in the Ambrosinus-Toolkit a Grasshopper Plugin.

All credits to each author who developed these Midas and DPT scripts.

I have packaged them and modified some strings of code to run them inside Grasshopper.
Every file is under the MIT licence.

Requirements

To run the DPTto3D component and DPTSemSeg component (subcategory AI) some Python libraries are necessary as the other AI tools. I have coded a Python library named atoolkitdpt v0.0.2 as part of Ambrosinus Toolkit project.

From your CMD window viewport, you can simply launch this command: pip install atoolkitdpt (I recommend this option) - in this way, all necessary libraries will be installed on your machine.

Alternatively, I have shared a "requirements.txt"(right click "Save as") file allowing the designer in this step in a unique command line from cmd.exe (Windows OS side). It does the same. After downloading the file to a custom folder (I suggest in C:/CustomFolder or something like that) run the following command from cmd.exe after logging in the "CustomFolder": pip install -r requirements.txt

Download at least one of these "weights models" pre-trained datasets by Intel Labs Research Team

MiDaS 3.1 for Monocular Depth Map Estimation (DPTto3D Grasshopper component)

For highest quality dpt_beit_large_512

For moderately less quality, but better speed-performance trade-off: dpt_swin2_large_384

For embedded devices: dpt_swin2_tiny_256, dpt_levit_224

MiDaS 3.0: Legacy transformer models dpt_large_384 and dpt_hybrid_384

MiDaS 2.1: Legacy convolutional models midas_v21_384 and midas_v21_small_256

Info components Article on my website

Semantic Segmentation with Dense Prediction Transformers (DPTSemSeg Grasshopper component)

For moderately less quality: dpt_hybrid-ade20k-53898607

For highest quality: dpt_large-ade20k-b12dca68

Info components Article on my website