ctongfei / hierarchical-typing

Hierarchical entity typing via multi-level learning to rank
12 stars 7 forks source link

ERROR: Could not consume arg: --other_delta #2

Open 22842219 opened 3 years ago

22842219 commented 3 years ago

Hi tongfei, How are you? I've got two errors below, when I was running your model. Could you please give me some idea about it? Thanks a lot. First error: train/Dataset.BBN+EmbDropout.0.6+LiftOther.True+Regularizer.0.003+RelConsCoef.0.1+ThresholdRatio.0.2+WithOther.True: INFO:allennlp.training.checkpointer:loading best weights train/Dataset.BBN+EmbDropout.0.6+LiftOther.True+Regularizer.0.003+RelConsCoef.0.1+ThresholdRatio.0.2+WithOther.True: ERROR: Could not consume arg: --other_delta

The following tasks failed: FAILED: train/Dataset.BBN+EmbDropout.0.6+LiftOther.True+Regularizer.0.003+RelConsCoef.0.1+ThresholdRatio.0.2+WithOther.True/3 ERROR: Task train/Dataset.BBN+EmbDropout.0.6+LiftOther.True+Regularizer.0.003+RelConsCoef.0.1+ThresholdRatio.0.2+WithOther.True/3 failed EXITING: Cleaning up lock files...

Second error. This error came out when I was trying to use hugging-face-bert as the contextualizer.

train/Contextualizer.bert-base-based+Dataset.BBN+EmbDropout.0.6+LiftOther.True+Regularizer.0.003+RelConsCoef.0.1+ThresholdRatio.0.2+WithOther.True: ERROR: Could not consume arg: --other_delta

Failed train/Contextualizer.bert-base-cased+Dataset.BBN+EmbDropout.0.6+LiftOther.True+Regularizer.0.003+RelConsCoef.0.1+ThresholdRatio.0.2+WithOther.True/4: Task train/Contextualizer.bert-base-cased+Dataset.BBN+EmbDropout.0.6+LiftOther.True+Regularizer.0.003+RelConsCoef.0.1+ThresholdRatio.0.2+WithOther.True/4 failed The following tasks failed: FAILED: train/Contextualizer.bert-base-cased+Dataset.BBN+EmbDropout.0.6+LiftOther.True+Regularizer.0.003+RelConsCoef.0.1+ThresholdRatio.0.2+WithOther.True/4 ERROR: Task train/Contextualizer.bert-base-cased+Dataset.BBN+EmbDropout.0.6+LiftOther.True+Regularizer.0.003+RelConsCoef.0.1+ThresholdRatio.0.2+WithOther.True/4 failed

It would be greatly appreciated if you could give me any idea on debugging them. Thanks a lot.

Best, Zea

ctongfei commented 3 years ago

Hi Zea, This is some old parameters that I have used before. In the train.tape script, remove the other_delta parameter for now. When I got some time I'll try to fix this in the workflow.

22842219 commented 3 years ago

Hi Zea, This is some old parameters that I have used before. In the train.tape script, remove the other_delta parameter for now. When I got some time I'll try to fix this in the workflow.

Hi tongfei, That would be great. Thanks a lot.

Best, Ziyu