In pt-avitm/ptavitm/sklearn_api.py samples, documents = X.shape should probably be called something less confusing, and in call to train, the line sampler=WeightedRandomSampler(torch.ones(samples), max(samples, self.samples)) should have min there instead.
In pt-avitm/ptavitm/sklearn_api.py
samples, documents = X.shapeshould probably be called something less confusing, and in call totrain, the linesampler=WeightedRandomSampler(torch.ones(samples), max(samples, self.samples))should have min there instead.