huggingface / transfer-learning-conv-ai

🦄 State-of-the-Art Conversational AI with Transfer Learning
MIT License
1.74k stars 430 forks source link

tb_logger Error: got an unexpected keyword argument 'another_engine' #85

Open RVirmoors opened 4 years ago

RVirmoors commented 4 years ago

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.)

RVirmoors commented 4 years ago

OK I simply removed "another_engine=trainer" and it seems to work fine.

ashantanu commented 4 years ago

I tried changing pytorch-ignite version to 0.3.0 and that worked.