Dear scikit team, My name is Chien-Cheng and I am trying to use scikit-image to do some image segmentation. The method I choose is called "Training segmentation" (https://scikit-image.org/docs/dev/auto_examples/segmentation/plot_trainable_...). Currently, I am trying to run the example of the training method and I got a few questions. The latest skimage version I have in the Anaconda is 0.17.2. But I found the training method seems only applicable for the developer version 0.19.0.dev0 since some of the attribute in the script is missing in the version of 0.17.2 (for example, there is an error called AttributeError: module 'skimage.feature' has no attribute 'multiscale_basic_features' after running the example script with 0.17.2). Also, I cannot find the example of the training method in the version of 0.17.2 so I figure the training methods require the developer version. I have tried installing the contributor/developer version in the conda environment by following the instruction (https://scikit-image.org/docs/dev/install.html#id6) but when I was running the third comment "conda install `for i in requirements/{default,build,test}.txt; do echo -n " --file $i "; done`. The console shows me the error below: [cid:image002.jpg@01D6CFE7.99D9EF80] I really don't know what the problem might be so I would like to turn to your help. There are two questions I have so far: 1. Do I need to install the developer version to use the training methods? 2. If so, did I install the developer version with the correct procedure? If yes, how to solve the issue I have above? Looking forward to your reply, thanks! Best regards, Chien-Cheng
