Darek
Darek
Just do: `--classification_task_name=COLA` then.
@zzj0402 How about you fix it by creating a PR to the owner of the repo?
As per [Readme](https://github.com/kamalkraj/ALBERT-TF2.0/blob/master/README.md) ``` export GLUE_DIR=glue_data/ export ALBERT_DIR=large/ export TASK_NAME=CoLA export OUTPUT_DIR=cola_processed mkdir $OUTPUT_DIR python create_finetuning_data.py \ --input_data_dir=${GLUE_DIR}/ \ --spm_model_file=${ALBERT_DIR}/vocab/30k-clean.model \ --train_data_output_path=${OUTPUT_DIR}/${TASK_NAME}_train.tf_record \ --eval_data_output_path=${OUTPUT_DIR}/${TASK_NAME}_eval.tf_record \ --meta_data_file_path=${OUTPUT_DIR}/${TASK_NAME}_meta_data \ --fine_tuning_task_type=classification --max_seq_length=128 \...
Are you are trying run on a GPU but you don't have one or it's not configured? Please try using a Docker file `docker pull tensorflow/tensorflow:latest-gpu-py3` to ensure your GPU...
[This code works](https://github.com/kamalkraj/ALBERT-TF2.0/issues/25) to inference a single value from a saved model, hopefully it helps.
I think it does, b/c currently I can NOT do import without babel. import reqdir from 'require-dir';
Then, what is "node --experimental-modules" for?
It says: "Node.js contains support for ES Modules based upon the Node.js EP for ES Modules."
This is simple, you need to add: `--predict_feature_file="$SQUAD_DIR/dev.tfrecord"`
@lakshikaparihar Although you are adding adding an interesting example, your train.py is missing is log_metric call. You may want to combine the train.py and log_model.py and add a separate file...