search

UKPLab / starsem2018-entity-linking

Accompanying code for our *SEM 2018 @ NAACL 2018 paper "Mixing Context Granularities for Improved Entity Linking on Question Answering Data across Entity Categories"
Apache License 2.0
59 stars 16 forks source link

How to run run_experiments.sh ? #3

Closed Jorigorn closed 4 years ago

Jorigorn commented 5 years ago

What is parameter? GPUID=$1 Thanks. :)

daniilsorokin commented 5 years ago

That is the id of the GPU to use. If you don't have a GPU set it to any number.

Jorigorn commented 5 years ago

Hi, @daniilsorokin Thanks for your reply. :) I have GPUs, but errors show. It seem there are some other resource needed to be download? like "data/WebQSP/preprocessed/webqsp.train.entities.candidates.4.small.pkl" Thanks.

% % sh ./run_experiments.sh 4 Seed: 1 Load candidates. Traceback (most recent call last): File "/home/john/anaconda3/envs/starsem2018/lib/python3.6/runpy.py", line 193, in _run_module_as_main "main", mod_spec) File "/home/john/anaconda3/envs/starsem2018/lib/python3.6/runpy.py", line 85, in _run_code exec(code, run_globals) File "/mnt/hdd/experiments/john/starsem2018-entity-linking/entitylinking/mlearning/training/train_entity_linker.py", line 140, in train() File "/home/john/anaconda3/envs/starsem2018/lib/python3.6/site-packages/click/core.py", line 722, in call return self.main(args, kwargs) File "/home/john/anaconda3/envs/starsem2018/lib/python3.6/site-packages/click/core.py", line 697, in main rv = self.invoke(ctx) File "/home/john/anaconda3/envs/starsem2018/lib/python3.6/site-packages/click/core.py", line 895, in invoke return ctx.invoke(self.callback, ctx.params) File "/home/john/anaconda3/envs/starsem2018/lib/python3.6/site-packages/click/core.py", line 535, in invoke return callback(args, **kwargs) File "/mnt/hdd/experiments/john/starsem2018-entity-linking/entitylinking/mlearning/training/train_entity_linker.py", line 29, in train precomputed_candidates = load_candidates(linking_config['linker.options']['precomputed_candidates']) File "/mnt/hdd/experiments/john/starsem2018-entity-linking/entitylinking/core/sentence.py", line 53, in load_candidates with open(precomputed_candidates_path, "rb") as f: FileNotFoundError: [Errno 2] No such file or directory: 'data/WebQSP/preprocessed/webqsp.train.entities.candidates.4.small.pkl' Usage: evaluate.py [OPTIONS] PATH_TO_MODEL CONFIG_FILE_PATH

Error: Missing argumnt "config_file_path". Usage: evaluate.py [OPTIONS] PATH_TO_MODEL CONFIG_FILE_PATH

Error: Missing argument "config_file_path". john@serv-re ~/experiments/starsem2018-entity-linking

daniilsorokin commented 5 years ago

Hi,

long time waiting, but we have finally published all the evaluation data and the precomputed candidates for the evaluation: https://github.com/UKPLab/starsem2018-entity-linking#qa-data-for-benchmarking-entity-linking-systems

Best, Daniil

abmitra84 commented 5 years ago

Hi @daniilsorokin , I tried to run the run_experiments.sh ... it gets an error as dataset.py looking for webqsp.dev.ids.json in the WebQSP/input/ directory . How to resolve this? In the preprocessed QA dataset such json files are not present.

daniilsorokin commented 5 years ago

Hi,

you are right, webqsp.dev.ids.json was missing from the data set and it is needed for training and evaluation using the development set. I have updated the data set online: https://public.ukp.informatik.tu-darmstadt.de/starsem18-entity-linking/EntityLinkingForQADatasets.zip

Please download again.

Cheers, Daniil

zou-longkun commented 5 years ago

I trie to run the run_experiments.sh , it get an error NO such file or directory "data/WebQSP/input/webqsp.train.ids.json". How to resolve this issue?

daniilsorokin commented 5 years ago

Hi,

the file was indeed missing, I have added it to the data set: https://public.ukp.informatik.tu-darmstadt.de/starsem18-entity-linking/EntityLinkingForQADatasets.zip

Best, Daniil