search

google-research / albert

ALBERT: A Lite BERT for Self-supervised Learning of Language Representations
Apache License 2.0
3.23k stars 570 forks source link

tensorflow.python.framework.errors_impl.NotFoundError: Restoring from checkpoint failed. #224

Closed YuHengKit closed 4 years ago

YuHengKit commented 4 years ago

tensorflow.python.framework.errors_impl.NotFoundError: Restoring from checkpoint failed. This is most likely due to a Variable name or other graph key that is missing from the checkpoint. Please ensure that you have not altered the graph expected based on the checkpoint. Original error:

Key output_bias not found in checkpoint [[node save/RestoreV2 (defined at C:\ProgramData\Anaconda3\lib\site-packages\tensorflow_core\python\framework\ops.py:1748) ]]

Run below command and get above error python run_classifier.py --data_dir=C:\Users\Heng1222\Desktop\albert-master\glue_data --output_dir=C:\Users\Heng1222\Desktop\albert_base_v2\albert_base --init-checkpoint=C:\Users\Heng1222\Desktop\albert_base_v2\albert_base\model.ckpt-best --do_lower_case=true --max_seq_length=128 --albert_config_file=C:\Users\Heng1222\Desktop\albert_base_v2\albert_base\albert_config.json --spm_model_file=C:\Users\Heng1222\Desktop\albert_base_v2\albert_base\30k-clean.model --do_train=false --do_predict=true --do_eva=true --optimizer=adamw --task_name=MNLI --warmup_step=1000 --learning_rate=3e-5 --train_step=10000 --save_checkpoints_steps=100 --train_batch_size=128