search

wasiahmad / PLBART

Official code of our work, Unified Pre-training for Program Understanding and Generation [NAACL 2021].
https://arxiv.org/abs/2103.06333
MIT License
186 stars 35 forks source link

Fine tuning PLBART on Devign or a new dataset #38

Closed Kamel773 closed 2 years ago

Kamel773 commented 2 years ago

Hi @wasiahmad,

Thank you for sharing the source code and dataset with us.

How can I fine tune PLBART (training) on Devign dataset or a new dataset?

Kamel

wasiahmad commented 2 years ago

Please follow the README.

Kamel773 commented 2 years ago

Thank you @wasiahmad, I am still having an error in fine tuning the model. Please have a look at the error below.

2022-06-29 19:27:02 | INFO | fairseq.utils | CUDA enviroments for all 1 workers 2022-06-29 19:27:02 | INFO | fairseq.utils | rank 0: capabilities = 7.5 ; total memory = 14.756 GB ; name = Tesla T4
2022-06-29 19:27:02 | INFO | fairseq.utils | CUDA enviroments for all 1 workers 2022-06-29 19:27:02 | INFO | fairseq_cli.train | training on 1 devices (GPUs/TPUs) 2022-06-29 19:27:02 | INFO | fairseq_cli.train | max tokens per GPU = None and max sentences per GPU = 4 Traceback (most recent call last): File "/opt/conda/bin/fairseq-train", line 8, in sys.exit(cli_main()) File "/opt/conda/lib/python3.7/site-packages/fairseq_cli/train.py", line 348, in cli_main distributed_utils.call_main(args, main) File "/opt/conda/lib/python3.7/site-packages/fairseq/distributed_utils.py", line 187, in call_main main(args, kwargs) File "/opt/conda/lib/python3.7/site-packages/fairseq_cli/train.py", line 106, in main extra_state, epoch_itr = checkpoint_utils.load_checkpoint(args, trainer) File "/opt/conda/lib/python3.7/site-packages/fairseq/checkpoint_utils.py", line 137, in load_checkpoint reset_meters=args.reset_meters, File "/opt/conda/lib/python3.7/site-packages/fairseq/trainer.py", line 252, in load_checkpoint state = checkpoint_utils.load_checkpoint_to_cpu(filename) File "/opt/conda/lib/python3.7/site-packages/fairseq/checkpoint_utils.py", line 169, in load_checkpoint_to_cpu f, map_location=lambda s, l: default_restore_location(s, "cpu") File "/opt/conda/lib/python3.7/site-packages/torch/serialization.py", line 713, in load return _legacy_load(opened_file, map_location, pickle_module, pickle_load_args) File "/opt/conda/lib/python3.7/site-packages/torch/serialization.py", line 920, in _legacy_load magic_number = pickle_module.load(f, **pickle_load_args) _pickle.UnpicklingError: invalid load key, '<'.

wasiahmad commented 2 years ago

Please make sure you use an environment that follows the package requirements. All the code in this repo is tested, so if something doesn't work, I assume the environment is not properly configured.