I’m getting a keyword error when I try to train a model using Open-NMTtf. I did some googling but I didn’t see much on this error.
$onmt-main --model_type Transformer --config data.yml --auto_config train --with_eval
Below is my data.yml file:
model_dir: run/
data:
train_features_file: src-train.txt
train_labels_file: tgt-train.txt
eval_features_file: src-val.txt
eval_labels_file: tgt-val.txt
source_vocabulary: src-vocab.txt
target_vocabulary: tgt-vocab.txt
Below is the error I’m getting.
Traceback (most recent call last):
File "/home/username/.local/bin/onmt-main", line 8, in <module>
sys.exit(main())
File "/home/username/.local/lib/python3.6/site-packages/opennmt/bin/main.py", line 231, in main
hvd=hvd)
File "/home/username/.local/lib/python3.6/site-packages/opennmt/runner.py", line 199, in train
checkpoint_path=checkpoint_path, weights_only=checkpoint_path is not None)
File "/home/username/.local/lib/python3.6/site-packages/opennmt/utils/checkpoint.py", line 129, in restore
checkpoint_path, tmp_model, optimizer=tmp_optimizer)
File "/home/username/.local/lib/python3.6/site-packages/opennmt/utils/checkpoint.py", line 274, in _restore_v1_checkpoint
mapping = model.map_v1_weights(v1_structure)
File "/home/username/.local/lib/python3.6/site-packages/opennmt/models/transformer.py", line 158, in map_v1_weights
weights["seq2seq"] = weights.pop("transformer")
KeyError: 'transformer'