Loading README.md +19 −7 Original line number Diff line number Diff line Loading @@ -50,14 +50,16 @@ Activate the virtual environment with your dependencies. $ source absinth_env/bin/activate ``` Run `absinth.py` without modifiers to use the dataset path and with '-t' for the trial path. Run `absinth.py` without modifiers to run on trial data. ``` $ python3 absinth.py ``` Specify on which data to run with the commands `-test`, `-trial` and `-dev` ``` $ python3 absinth.py -t $ python3 absinth.py -dev ``` Absinth may utilise multiprocessing for parallel topic processing. Simply put Loading @@ -69,7 +71,7 @@ $ python3 absinth.py -p 4 Absinth produces individual clustering files for every target. To merge the output files, simple call `merge.py`. The resulting output for the WSI-evaluator is placed in the 'absinth/src/final' directory. is placed in the 'absinth/output/.final/' directory. ``` $ python3 merge.py Loading @@ -84,16 +86,26 @@ Our baseline `abstinent.py` works in much the same way as `absinth.py`. Not all variables in `config.py` are supported in our baseline though. ``` $ python3 abstinent.py -t -p 3 $ python3 abstinent.py -test -p 3 ``` To merge the `abstinent.py` output, simply put the '-bl' modifier behind `merge.py`: To merge the `abstinent.py` output, simply run it the same way as before: ``` $ python3 merge.py -bl $ python3 merge.py ``` View the documentation for more information. `all-in-one.py` and `singletons.py` do not need to be merged! Calling `merge.py` after running either of these programs results in overwriting their output. ## Evaluation We employ the same evaluator as the orginal SemEval-2013 Task, but simplified the process. Just call `evaluate.sh` with the set you want to evaluate on to evaluate the most recent merge. ``` $ bash evaluate.sh development ``` ## Built With Loading Loading
README.md +19 −7 Original line number Diff line number Diff line Loading @@ -50,14 +50,16 @@ Activate the virtual environment with your dependencies. $ source absinth_env/bin/activate ``` Run `absinth.py` without modifiers to use the dataset path and with '-t' for the trial path. Run `absinth.py` without modifiers to run on trial data. ``` $ python3 absinth.py ``` Specify on which data to run with the commands `-test`, `-trial` and `-dev` ``` $ python3 absinth.py -t $ python3 absinth.py -dev ``` Absinth may utilise multiprocessing for parallel topic processing. Simply put Loading @@ -69,7 +71,7 @@ $ python3 absinth.py -p 4 Absinth produces individual clustering files for every target. To merge the output files, simple call `merge.py`. The resulting output for the WSI-evaluator is placed in the 'absinth/src/final' directory. is placed in the 'absinth/output/.final/' directory. ``` $ python3 merge.py Loading @@ -84,16 +86,26 @@ Our baseline `abstinent.py` works in much the same way as `absinth.py`. Not all variables in `config.py` are supported in our baseline though. ``` $ python3 abstinent.py -t -p 3 $ python3 abstinent.py -test -p 3 ``` To merge the `abstinent.py` output, simply put the '-bl' modifier behind `merge.py`: To merge the `abstinent.py` output, simply run it the same way as before: ``` $ python3 merge.py -bl $ python3 merge.py ``` View the documentation for more information. `all-in-one.py` and `singletons.py` do not need to be merged! Calling `merge.py` after running either of these programs results in overwriting their output. ## Evaluation We employ the same evaluator as the orginal SemEval-2013 Task, but simplified the process. Just call `evaluate.sh` with the set you want to evaluate on to evaluate the most recent merge. ``` $ bash evaluate.sh development ``` ## Built With Loading