While running inference on my text classifier learner on CPU this userwarning shows up for each prediction.
lib\site-packages\fastai\torch_core.py:83: UserWarning: Tensor is int32: upgrading to int64; for better >performance use int64 input.
The learner is exported using learn.export(‘modelname’) and loaded using load_learner(path=model_dir, file=model_name)
When loading the models this warning appear:
lib\site-packages\torch\serialization.py:454: SourceChangeWarning: source code of class ‘fastai.text.learner.PoolingLinearClassifier’ has changed. you can retrieve the original source code by accessing the object’s source attribute or set
torch.nn.Module.dump_patches = True
and use the patch tool to revert the changes.
warnings.warn(msg, SourceChangeWarning)
With these classes as well:
‘fastai.text.learner.MultiBatchEncoder’
‘fastai.text.models.awd_lstm.AWD_LSTM’
‘fastai.text.models.awd_lstm.EmbeddingDropout’
‘fastai.text.models.awd_lstm.WeightDropout’
‘fastai.text.models.awd_lstm.RNNDropout’
‘fastai.text.learner.PoolingLinearClassifier’
fastai version is 1.0.54
torch version is 1.1.0
Also when loading the learner it creates folder called ‘models’ in the same folder the model is loaded from. How can I disable that?