forked from ufal/qtleap
-
Notifications
You must be signed in to change notification settings - Fork 0
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
testing should be done in ../translate/
- Loading branch information
1 parent
e4e7a0d
commit 4a97d64
Showing
3 changed files
with
1 addition
and
247 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
|
@@ -25,57 +25,6 @@ make transl_models TRANSL_PAIR=cs_en | |
# To analyze parallel training data and create t-lemma and formeme translation models. | ||
# These models will be used for translation. | ||
|
||
make prepare_test_data TRANSL_PAIR=cs_en | ||
# Prepare the test data, i.e. split plain text to several documents and create SGM files, | ||
# which are needed for MT evaluation. | ||
# The data can be specified by assigning paths to the <TEST_DATA> variable in the following format: | ||
# <TEST_DATA="path_to_data/czech.data.sgm; path_to_data/english.data.sgm"> | ||
# The English part of the parallel data must be always specified as the second. | ||
# Two formats of the test data are supported: plain text and SGM. The format can be specified | ||
# assigning the variable <TEST_DATA_STAGE> the value <orig_plain> (default) or <segm>, respectively. | ||
# To facilitate the test dataset setting, one can store a list of the frequently used datasets | ||
# with their labels to the <TEST_DATASET_HASH> variable in the config file (see e.g. <conf/cs_en.conf>). | ||
# If any dataset is in SGM format, its label has to be contained in a space-separated list | ||
# assigned to <TEST_DATASET_IN_SGM> variable. | ||
# Then, a particular dataset is selected by setting the <TEST_DATASET> variable to the dataset's label. | ||
|
||
make translate eval TRANSL_PAIR=cs_en | ||
make translate eval TRANSL_PAIR=cs_en TEST_DATASET=batch2q | ||
# It translates and evaluates the specified test data. | ||
# If the translation models are not present, it tries to train them. | ||
# If you want to use already trained models, copy them into share and specify the paths | ||
# in the config file (e.g. see <conf/en_cs.conf> for English to Czech translation models) | ||
|
||
make translate eval S=scenarios/02_my_alternative.scen | ||
# TODO S is not yet supported | ||
# Now try an alternative scenario | ||
# All the experiments are saved in $(RUNS_DIR) in a subdirectory with current date in name. | ||
|
||
make translate eval TRANSL_PAIR=cs_en D="changed regexp in block XY" | ||
# You can add a short description of the experiment to be logged | ||
|
||
make compare | ||
# Useful comparison of outputs of the two scenarios is saved in compared_with_* files. | ||
# With this command you can see the same info which compares the baseline with the last experiment. | ||
|
||
make baseline | ||
# Write this to set the last experiment as the baseline (after you have checked it is better than the previous one). | ||
# It will print the name of the last experiment directory to $(RUNS_DIR)/baseline. | ||
|
||
make baseline-042 | ||
# If you want to comfortably set another experiment as the baseline, look up its prefix number and call, e.g. this, | ||
# which will set $(RUNS_DIR)/042_2010-01-14_14-07-02_my_scenario (let's say) as the baseline. | ||
|
||
make compare-042:046 | ||
# Compare experiments with numbers 042 and 046 (without changing $(RUNS_DIR)/baseline). | ||
|
||
make eval-046 | ||
# Evaluate experiment number 046 (if you have forgotten to add eval after translate). | ||
|
||
make pilot0 eval-000 | ||
# Copy Pilot 0 (moses) output to $(RUNS_DIR)/000_pilot0, compute its BLEU | ||
|
||
# If you don't want to print commands being executed, use "make -s". | ||
endef | ||
export HELP_MESSAGE | ||
|
||
|
@@ -133,11 +82,6 @@ $(TRANSL_MODELS) : | |
clean_tm clean_tm_train_table clean_analysis clean_for_giza clean_giza clean_tecto : | ||
$(MAKE) -f makefile.tm $@ TRANSL_PAIR=$(TRANSL_PAIR) DATA="$(TRAIN_DATA)" STAGE=$(TRAIN_DATA_STAGE) LRC=$(LRC) | ||
|
||
|
||
##################### translation ########################## | ||
include makefile.test_data | ||
include makefile.translation | ||
|
||
qtleap-corpus: ../qtleap-corpus | ||
../qtleap-corpus: | ||
git clone [email protected]:qtleap/qtleap-corpus.git $@ | ||
git clone [email protected]:qtleap/qtleap-corpus.git $@ |
This file was deleted.
Oops, something went wrong.
This file was deleted.
Oops, something went wrong.