Problem with data loader

[bHimes]

I'm working through the "complete walkthrough" and ran into a snag when I got to the training step.

My input

(base) [himesb@london testTopaz]$ topaz train -n 300 \
>             --num-workers=8 \
>             --train-images data/EMPIAR-10025/processed/image_list_train.txt \
>             --train-targets data/EMPIAR-10025/processed/particles_train.txt \
>             --test-images data/EMPIAR-10025/processed/image_list_test.txt \
>             --test-targets data/EMPIAR-10025/processed/particles_test.txt \
>             --save-prefix=saved_models/EMPIAR-10025/model \
>             -o saved_models/EMPIAR-10025/model_training.txt

Output:

# Loading model: resnet8
# Model parameters: units=32, dropout=0.0, bn=on
# Receptive field: 71
# Using device=0 with cuda=True
# Loaded 20 training micrographs with 1000 labeled particles
# Loaded 10 test micrographs with 500 labeled particles
# source        split   p_observed      num_positive_regions    total_regions
# 0     train   0.00654 29000   4435200
# 0     test    0.00654 14500   2217600
# Specified expected number of particle per micrograph = 300.0
# With radius = 3
# Setting pi = 0.03923160173160173
# minibatch_size=256, epoch_size=5000, num_epochs=10

Error Message

Traceback (most recent call last):
  File "/groups/grigorieff/home/himesb/thirdParty/anaconda3/bin/topaz", line 11, in <module>
    load_entry_point('topaz', 'console_scripts', 'topaz')()
  File "/groups/grigorieff/home/himesb/thirdParty/topaz/topaz/main.py", line 144, in main
    args.func(args)
  File "/groups/grigorieff/home/himesb/thirdParty/topaz/topaz/commands/train.py", line 649, in main
    classifier.width, split, args)
  File "/groups/grigorieff/home/himesb/thirdParty/topaz/topaz/commands/train.py", line 471, in make_data_iterators
    , num_workers=num_workers)
  File "/groups/grigorieff/home/himesb/thirdParty/anaconda3/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 805, in __init__
    batch_sampler = BatchSampler(sampler, batch_size, drop_last)
  File "/groups/grigorieff/home/himesb/thirdParty/anaconda3/lib/python3.6/site-packages/torch/utils/data/sampler.py", line 146, in __init__
    .format(sampler))

ValueError: sampler should be an instance of torch.utils.data.Sampler, but got sampler=<topaz.utils.data.sampler.StratifiedCoordinateSampler object at 0x7f6b74093cf8>

(last line pulled from error msg for emphasis)

[tbepler]

What version of pytorch do you have installed? Topaz requires version 0.2.0 of pytorch because newer versions are not backwards compatible. I suspect that is causing this problem.

[bHimes]

I had everything but pillow installed, and so had missed the " = " sign relative to the " >= " in the list of dependencies when I scanned over them. Downgrading the installation took care of the problem. Thanks you!