Running the EdgeLM model on XSum

[pramodith]

Hi I am trying to use the EdgeLM model and was wondering

1. how I can run the model with pre-trained weights in interactive mode, I'd like to run it for a summarization task.
2. I was also wondering if the pre-trained weights can be used to replicate the results reported on the XSum dataset.
3. How to make use of the pre-trained 2k vocab file and sentencepiece model while attempting 1 and 2.
4. How can I run the model on CPU only.

[getao]

Hi,

(1) you can try fairseq-interactive to use the model in the interactive mode. (2) Yes, but you need to fine-tune the model for XSum. The pretrained weights cannot be directly used for summarization (3) It is the same way as shown in the instructions, just need to replace fairseq-generate with fairseq-interactive (you can refer to fairseq-interactive to see the detailed instructions) (4) You need to add the flag --cpu while running fairseq-interactive.

[inderpreetsingh01]

Hi @getao i was wondering what to use for --task for pertaining edgeformer for summarization?