Test/Predict on summarization task

[zyxnlp]

Environment info

• transformers version: 4.4.0.dev0
• Platform: Linux-5.3.0-53-generic-x86_64-with-glibc2.10
• Python version: 3.8.3
• PyTorch version (GPU?): 1.7.1 (True)
• Tensorflow version (GPU?): not installed (NA)
• Using GPU in script?:
• Using distributed or parallel set-up in script?:

Who can help

• maintained examples using bart: @patrickvonplaten, @patil-suraj

Information

Model I am using (Bart):

The problem arises when using:

• [ ] only have train and eval example, do not have test/predict example script
• [ ] my own modified scripts: (give details below)
• [ ] CUDA_VISIBLE_DEVICES=2,3 python examples/seq2seq/run_seq2seq.py \ --model_name_or_path /home/yxzhou/experiment/ASBG/output/xsum_bart_large/ \ --do_predict \ --task summarization \ --dataset_name xsum \ --output_dir /home/yxzhou/experiment/ASBG/output/xsum_bart_large/test/ \ --num_beams 1 \

The tasks I am working on is:

• [ ] CNNDAILYMAIL , XSUM

Expected behavior

Use the above script to test on CNNDAILYMAIL and XSUM dataset, the program seems will be always stuck at training_step (e.g., 26/709)

Could you please kindly provide a test/predict example script of the summarization task (e.g., XSUM). Thank you so much!

[NielsRogge]

Hi,

could you please ask questions related to training of models on the forum?

All questions related to fine-tuning a model for summarization on CNN can be found here for example.

[zyxnlp]

Hi,

could you please ask questions related to training of models on the forum?

All questions related to fine-tuning a model for summarization on CNN can be found here for example.

@NielsRogge Oh, thank you for your reminder, that really helped!

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.

Please note that issues that do not follow the contributing guidelines are likely to be ignored.