transfer-learning-conv-ai
transfer-learning-conv-ai copied to clipboard
tb_logger Error: got an unexpected keyword argument 'another_engine'
Hi, I'm trying to run train.py on Google Colab. interact.py works fine, but when trying to fine-tune, "!python train.py" (with or without arguments) stops here:
INFO:train.py:Build inputs and labels INFO:train.py:Pad inputs and convert to Tensor INFO:train.py:Build train and validation dataloaders INFO:train.py:Train dataset (Batch, Candidates, Seq length): torch.Size([131438, 2, 285]) INFO:train.py:Valid dataset (Batch, Candidates, Seq length): torch.Size([7801, 20, 185]) Traceback (most recent call last): File "train.py", line 267, in
train() File "train.py", line 249, in train tb_logger.attach(evaluator, log_handler=OutputHandler(tag="validation", metric_names=list(metrics.keys()), another_engine=trainer), event_name=Events.EPOCH_COMPLETED) TypeError: init() got an unexpected keyword argument 'another_engine'
(this is using the CPU runtime. If I try the GPU, I don't even get this far... interact.py crashes too)
could it be a tensorboard version incompatibility? i've got the following installed:
tensorboard 2.2.2
tensorboard-plugin-wit 1.7.0
tensorboardcolab 0.0.22
tensorboardX 1.8
pytorch-ignite 0.5.0
(just tried it locally and I'm getting the same error.)
OK I simply removed "another_engine=trainer" and it seems to work fine.
I tried changing pytorch-ignite version to 0.3.0 and that worked.