Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

When using the pre-trained Model for story generation.

See original GitHub issue

When running interactive.py for example it loads the model an then tries to load the pretrained model (because of the fusion training) which exits with error:

OSError: Model file not found: /checkpoint/angelafan/wp_open_source/checkpoint_best.pt

Is there anyway to fix this? Or maybe also make the pretrained model available?

Issue Analytics

State:
Created 5 years ago
Comments:7 (3 by maintainers)

Top GitHub Comments

1reaction

huihuifancommented, Jul 10, 2018

cannot repro

cd fairseq-py
git pull 
wget https://s3.amazonaws.com/fairseq-py/models/stories_checkpoint.tar.bz2
tar xvjf stories_checkpoint.tar.bz2
# unpacks it into models folder
python generate.py data-bin/writingPrompts --target-lang wp_target --source-lang wp_source --path models/fusion_checkpoint.pt --batch-size 32 --beam 1 --sampling --sampling-topk 10 --sampling-temperature 0.8 --nbest 1 --model-overrides "{'pretrained_checkpoint':'models/pretrained_checkpoint.pt'}"

please make sure you’re providing the path to the right models and you’ve binarized the dataset with the same number of tokens as my pretrained model vocabulary. If you load your own checkpoint, your vocabulary may not be the same.

0reactions

huihuifancommented, Jul 18, 2018

https://github.com/pytorch/fairseq/issues/216 may help if your vocabulary is off.

Top Results From Across the Web

A Knowledge-Enhanced Pretraining Model for Commonsense ...

planning entities and events with proper tem- poral order. In this paper, we devise a knowledge-enhanced pretraining model for commonsense story generation.

When using the pre-trained Model for story generation. #209

When running interactive.py for example it loads the model an then tries to load the pretrained model (because of the fusion training) which ......

A Knowledge-Enhanced Pretraining Model for Commonsense ...

In this paper, we devise a knowledge-enhanced pretraining model for commonsense story generation. We propose to utilize commonsense knowledge ...

A Knowledge-Enhanced Pretraining Model for Commonsense ...

We propose a knowledge-enhanced pretraining model for commonsense story generation by extending GPT-2 with external commonsense knowledge.

Story Generation | Papers With Code

Story generation is the task of automatically generating a coherent narrative, often from a set of premises or a brief summary.