diff options
23 files changed, 377 insertions, 2 deletions
@@ -19,6 +19,7 @@ tools/subword-nmt tools/nematus models/wmt16_systems/en-de +models/wmt16_systems/de-en models/wmt17_systems/en-de models/wmt17_systems/scripts models/wmt17_systems/vars diff --git a/models/download-wmt16.sh b/models/download-wmt16.sh index 8cb6ba6..5095545 100644 --- a/models/download-wmt16.sh +++ b/models/download-wmt16.sh @@ -1,4 +1,7 @@ #!/bin/bash -# Download single en-de wmt16 model -wget -nv -nc -r -e robots=off -nH -np -R *r2l* -R index.html* http://data.statmt.org/wmt16_systems/en-de/ +# Download all en-de wmt16 models +wget -nv -nc -r -e robots=off -nH -np -R '*r2l*' -R 'index.html*' http://data.statmt.org/wmt16_systems/en-de/ + +# Download single de-en wmt16 model +wget -nv -nc -r -e robots=off -nH -np -R '*r2l*' -R '*-ens*' -R '*.sh' -R 'index.html*' http://data.statmt.org/wmt16_systems/de-en/ diff --git a/models/wmt16_systems/marian.de-en.yml b/models/wmt16_systems/marian.de-en.yml new file mode 100644 index 0000000..ce5bb5a --- /dev/null +++ b/models/wmt16_systems/marian.de-en.yml @@ -0,0 +1,11 @@ +relative-paths: true +type: amun +models: + - de-en/model.npz +dim-emb: 500 +vocabs: + - de-en/vocab.de.json + - de-en/vocab.en.json +dim-vocabs: + - 85000 + - 85000 diff --git a/tests/decoder/xml/align.expected b/tests/decoder/xml/align.expected new file mode 100644 index 0000000..4e805cf --- /dev/null +++ b/tests/decoder/xml/align.expected @@ -0,0 +1,12 @@ +the house is small . ||| 2-0 2-1 3-2 4-3 5-4 5-5 +the house is small . ||| 2-0 2-1 3-2 4-3 5-4 5-5 +the house is small . ||| 2-0 2-1 3-2 4-3 5-4 5-5 +the house is small . ||| 2-0 2-1 3-2 4-3 5-4 5-5 +the house is small . ||| 2-0 2-1 3-2 4-3 5-4 5-5 +the house is small . ||| 2-0 2-1 3-2 4-3 5-4 5-5 +the house is small . ||| 2-0 2-1 3-2 4-3 5-4 5-5 +the house is small . ||| 2-0 2-1 3-2 4-3 5-4 5-5 +the house is small . ||| 2-0 2-1 3-2 4-3 5-4 5-5 +the house is small . ||| 2-0 2-1 3-2 4-3 5-4 5-5 +the house is big . ||| 2-0 2-1 3-2 4-3 5-4 5-5 +the house is big . ||| 2-0 2-1 3-2 4-3 5-4 5-5 diff --git a/tests/decoder/xml/multiword.expected b/tests/decoder/xml/multiword.expected new file mode 100644 index 0000000..463c30f --- /dev/null +++ b/tests/decoder/xml/multiword.expected @@ -0,0 +1,15 @@ +0 ||| the house is big . ||| F0= -1.42143 ||| 9.06422 +0 ||| the house big . ||| F0= -9.45543 ||| 8.0014 +0 ||| the house is large . ||| F0= -2.19537 ||| -0.365895 +1 ||| the house is too big . ||| F0= -1.51132 ||| 9.12714 +1 ||| the house is too big a book . ||| F0= -7.44603 ||| 8.66169 +1 ||| the house is too large . ||| F0= -3.24491 ||| -0.463558 +2 ||| the house is small . ||| F0= -6.87356 ||| 8.82468 +2 ||| the house is small and not big . ||| F0= -11.8984 ||| 8.65814 +2 ||| the house is not big . ||| F0= -1.7791 ||| -0.254157 +3 ||| the house is big . not small . ||| F0= -22.4062 ||| 6.72605 +3 ||| the house is big . ||| F0= -1.42143 ||| -0.236905 +3 ||| the house is big . not so . ||| F0= -18.5419 ||| -2.06021 +4 ||| the house is big . very nice and clean . very very big hotel . ||| F0= -40.9166 ||| 6.78803 +4 ||| the house is big . ||| F0= -1.42143 ||| -0.236905 +4 ||| the house is big . very nice and clean . ||| F0= -17.4485 ||| -1.58622 diff --git a/tests/decoder/xml/multiword.in b/tests/decoder/xml/multiword.in new file mode 100644 index 0000000..2608f8a --- /dev/null +++ b/tests/decoder/xml/multiword.in @@ -0,0 +1,5 @@ +das haus ist <a translation="big"> gross </a> . +das haus ist <a translation="too big"> zu gross </a> . +das haus ist <a translation="small"> nicht gross </a> . +das haus ist <a translation="not small"> gross </a> . +das haus ist <a translation="very very big"> gross </a> . diff --git a/tests/decoder/xml/nbest.expected b/tests/decoder/xml/nbest.expected new file mode 100644 index 0000000..06365b8 --- /dev/null +++ b/tests/decoder/xml/nbest.expected @@ -0,0 +1,40 @@ +0 ||| Obama welcomes Netanyahu Trumpf ||| F0= -18.9522 ||| 6.18159 +0 ||| Obama Trumpf ||| F0= -16.7771 ||| 3.21343 +0 ||| Obama welcomes Netanyahu ||| F0= -0.72431 ||| -0.181077 +0 ||| Obama welcomes ||| F0= -3.47103 ||| -1.15701 +1 ||| the relationship between Obama and Netanyahu is not exactly friendly . ||| F0= -1.44457 ||| 9.63365 +1 ||| the relationship between Obama and Netanyahu is not exactly amic@@ able . ||| F0= -2.1021 ||| 9.61125 +1 ||| the relationship with Obama and Netanyahu is not exactly friendly . ||| F0= -8.51821 ||| -0.709851 +1 ||| the relationship with Obama and Netanyahu is not exactly amic@@ able . ||| F0= -9.26852 ||| -0.712963 +2 ||| the two wanted to discuss the implementation of the international treaty and of Tehran 's destabili@@ zing measures in the Middle East . ||| F0= -11.3036 ||| 9.41963 +2 ||| the two wanted to discuss the implementation of the international treaty and of Tehran 's destabilising measures in the Middle East . ||| F0= -10.7499 ||| 9.41847 +2 ||| the two wanted to discuss the implementation of the international agreement and of Tehran 's destabilising measures in the Middle East . ||| F0= -5.54444 ||| -0.241063 +2 ||| the two wanted to talk about the implementation of the international agreement and of Tehran 's destabilising measures in the Middle East . ||| F0= -6.38887 ||| -0.266203 +3 ||| however , the meeting is also about the conflict with the Palestinians and the two states discussed . ||| F0= -11.7012 ||| 9.17332 +3 ||| however , the meeting is also about the conflict with the Palestinians and the two countries discussed . ||| F0= -12.5469 ||| 9.12881 +3 ||| however , the meeting is also intended to resolve the conflict with the Palestinians and the two states discussed . about the conflict . ||| F0= -30.7644 ||| 8.44483 +3 ||| however , the meeting is also intended to resolve the conflict with the Palestinians and the two states discussed . ||| F0= -11.7155 ||| -0.557883 +4 ||| the relationship between Obama and Netanyahu has been stretched for years . ||| F0= -2.51302 ||| -0.193309 +4 ||| the relationship between Obama and Netanyahu has been eagerly awaited for years . ||| F0= -3.67924 ||| -0.262803 +4 ||| the relationship between Obama and Netanyahu has been looking forward for years . ||| F0= -3.78594 ||| -0.270424 +4 ||| the relationship between Obama and Netanyahu has been looking forward to years . ||| F0= -3.96719 ||| -0.283371 +5 ||| Washington critici@@ zes Israel 's ongoing crackdown , accusing Netanyahu of a lack of will ! ||| F0= -20.9279 ||| 8.60014 +5 ||| Washington critici@@ zes Israel 's ongoing crackdown , accusing Netanyahu of a lack of will in the ! ||| F0= -27.7197 ||| 8.32881 +5 ||| Washington critici@@ zes Israel 's ongoing crackdown , accusing Netanyahu ! ||| F0= -24.6621 ||| 7.76781 +5 ||| Washington critici@@ zes Israel 's ongoing crackdown and ! ||| F0= -24.5952 ||| 7.39694 +6 ||| the relationship between the two continued to deteriorate through Obama 's deal about Iran 's nuclear program . ||| F0= -11.6037 ||| -0.61072 +6 ||| the relationship between the two continued to deteriorate through Obama 's deal with Iran 's nuclear program . ||| F0= -11.9141 ||| -0.627058 +6 ||| the relationship between the two continued to deteriorate through Obama 's deal to Iran 's nuclear program . ||| F0= -11.9579 ||| -0.629365 +6 ||| the relationship between the two continued to deteriorate through Obama 's deal for Iran 's nuclear program . ||| F0= -12.1275 ||| -0.638287 +7 ||| in April , Netanyahu held a controversial speech at the Congressional Republicans ' invitation show before the U.S. Congress . ||| F0= -22.5869 ||| 18.6619 +7 ||| in April , Netanyahu held a controversial show at the invitation of Republicans before the U.S. Congress . ||| F0= -22.4193 ||| 18.5267 +7 ||| in April , Netanyahu held a controversial speech at the Congressional Republicans ' invitation . ||| F0= -14.5929 ||| 8.97651 +7 ||| in March , Netanyahu held a controversial speech at the Congressional Republicans ' invitation . ||| F0= -5.94075 ||| -0.371297 +8 ||| the speech was not agreed with Obama , a meeting had been rejected , with reference to the election in Israel that was chicken at the time . ||| F0= -31.1594 ||| 8.68734 +8 ||| the speech was not agreed with Obama , a meeting had been rejected , with reference to the chicken that was imminent . ||| F0= -28.4084 ||| 8.52849 +8 ||| the speech was not agreed with Obama , a meeting had been rejected , with reference to the forthcoming election in Israel . ||| F0= -11.6751 ||| -0.486463 +8 ||| the speech was not agreed with Obama , a meeting had been rejected , with reference to the election in Israel that was imminent . ||| F0= -14.7121 ||| -0.565848 +9 ||| in an emergency call , teacher admits to having shot dead with his girlfriend ||| F0= -15.849 ||| 8.82056 +9 ||| in an emergency call , teacher admits to having shot his girlfriend ||| F0= -15.4446 ||| 8.67022 +9 ||| in an emergency call , Professor admits to having his girlfriend shot ||| F0= -4.61217 ||| -0.354782 +9 ||| in an emergency call , Professor admits to have shot dead his girlfriend ||| F0= -5.15151 ||| -0.367965 diff --git a/tests/decoder/xml/notags.in b/tests/decoder/xml/notags.in new file mode 100644 index 0000000..b9cb06a --- /dev/null +++ b/tests/decoder/xml/notags.in @@ -0,0 +1,10 @@ +India and Japan prime ministers meet in Tokyo +Mr Modi is on a five-@@ day trip to Japan to strengthen economic ties with the third largest economy in the world . +high on the agenda are plans for greater nuclear co-operation . +India is also reportedly hoping for a deal on defence collaboration between the two nations . +Kar@@ ra@@ tha police arrest 20-@@ year-old after high speed motorcycle chase +he is due to appear in Kar@@ ra@@ tha Magistr@@ ates Court on September 23 . +the motorcycle was seized and im@@ po@@ unded for three months . +a man is to stand trial accused of rap@@ ing women at two hotels . +history is a great teacher . +history remembers , as should we . diff --git a/tests/decoder/xml/setup.sh b/tests/decoder/xml/setup.sh new file mode 100644 index 0000000..68143bd --- /dev/null +++ b/tests/decoder/xml/setup.sh @@ -0,0 +1,2 @@ +test -f $MRT_MODELS/wmt16_systems/en-de/model.npz || exit 1 +test -f $MRT_MODELS/wmt16_systems/de-en/model.npz || exit 1 diff --git a/tests/decoder/xml/simpletags.in b/tests/decoder/xml/simpletags.in new file mode 100644 index 0000000..9aa9501 --- /dev/null +++ b/tests/decoder/xml/simpletags.in @@ -0,0 +1,10 @@ +<a translation="Indien">India</a> and Japan prime ministers meet in Tokyo +Mr Modi is on a five-@@ day trip to Japan to strengthen economic ties with the third largest economy in the world . +high on the agenda are plans for greater <a translation="nukleare">nuclear</a> co-operation . +India is also reportedly hoping for a deal on defence <a translation="kooperation">collaboration</a> between the two nations . +Kar@@ ra@@ tha <a translation="Polizei">police</a> arrest 20-@@ year-old after high speed motorcycle chase +he is due to appear in Kar@@ ra@@ tha Magistr@@ ates Court on September <a translation="23.">23</a> . +<a translation="das Motorrad">the motorcycle</a> was seized and im@@ po@@ unded for three months . +<a translation="ein Mann">a man</a> is to stand trial accused of rap@@ ing women at two hotels . +history is a great <a translation="Lehrer">teacher</a> . +history remembers , as <a translation="auch">should</a> we . diff --git a/tests/decoder/xml/tags.expected b/tests/decoder/xml/tags.expected new file mode 100644 index 0000000..5ee1626 --- /dev/null +++ b/tests/decoder/xml/tags.expected @@ -0,0 +1,12 @@ +the house is small . +the house is small . +the house is small . +the house is small . +the house is small . +the house is small . +the house is small . +the house is small . +the house is small . +the house is small . +the house is big . +the house is big . diff --git a/tests/decoder/xml/tags.in b/tests/decoder/xml/tags.in new file mode 100644 index 0000000..d9cd950 --- /dev/null +++ b/tests/decoder/xml/tags.in @@ -0,0 +1,12 @@ +das haus ist <a translation="small"> gross </a> . +das haus ist <a translation="small">gross</a> . +das haus ist <a translation="small">gross </a> . +das haus ist <a translation="small"> gross</a> . +das haus ist<a translation="small"> gross </a>. +das haus ist<a translation="small"> gross</a> . +das haus ist<a translation="small"> gross </a> . +das haus ist <xxx translation="small"> gross </xxx> . +das haus ist <yyy translation="small" foo="bar"> gross </yyy> . +das haus ist <a translation=" small "> gross </a> . +das haus ist <a translation = "small"> gross </a> . +das haus ist <a translation='small'> gross </a> . diff --git a/tests/decoder/xml/test_compare_xml_input_with_no_tags.sh b/tests/decoder/xml/test_compare_xml_input_with_no_tags.sh new file mode 100644 index 0000000..5194fad --- /dev/null +++ b/tests/decoder/xml/test_compare_xml_input_with_no_tags.sh @@ -0,0 +1,22 @@ +#!/bin/bash -x + +##################################################################### +# SUMMARY: Check if constrained decoding does not change an input without XML tags +# AUTHOR: snukky +##################################################################### + +# Exit on error +set -e + +# Remove old artifacts +rm -f notags*.{out,diff} + +# Run marian decoder +$MRT_MARIAN/marian-decoder -c $MRT_MODELS/wmt16_systems/marian.en-de.yml --mini-batch 10 -b 4 -n --n-best --xml-input < notags.in > notags.xml.out +$MRT_MARIAN/marian-decoder -c $MRT_MODELS/wmt16_systems/marian.en-de.yml --mini-batch 10 -b 4 -n --n-best < notags.in > notags.noxml.out + +# Compare the output with the expected output +$MRT_TOOLS/diff-nums.py notags.xml.out notags.noxml.out -o notags.diff + +# Exit with success code +exit 0 diff --git a/tests/decoder/xml/test_compare_xml_input_with_simple_tags.sh b/tests/decoder/xml/test_compare_xml_input_with_simple_tags.sh new file mode 100644 index 0000000..f8fbd2f --- /dev/null +++ b/tests/decoder/xml/test_compare_xml_input_with_simple_tags.sh @@ -0,0 +1,22 @@ +#!/bin/bash -x + +##################################################################### +# SUMMARY: Test constrained decoding with expected translations in XML tags +# AUTHOR: snukky +##################################################################### + +# Exit on error +set -e + +# Remove old artifacts +rm -f simpletags*.{out,diff} + +# Run marian decoder +$MRT_MARIAN/marian-decoder -c $MRT_MODELS/wmt16_systems/marian.en-de.yml --mini-batch 10 -b 4 -n --xml-input < simpletags.in > simpletags.xml.out +$MRT_MARIAN/marian-decoder -c $MRT_MODELS/wmt16_systems/marian.en-de.yml --mini-batch 10 -b 4 -n < notags.in > simpletags.noxml.out + +# Compare the output with the expected output +$MRT_TOOLS/diff-nums.py simpletags.xml.out simpletags.noxml.out -o simpletags.diff + +# Exit with success code +exit 0 diff --git a/tests/decoder/xml/test_xml_input_nbest_mini_batch_1.sh b/tests/decoder/xml/test_xml_input_nbest_mini_batch_1.sh new file mode 100644 index 0000000..30eec68 --- /dev/null +++ b/tests/decoder/xml/test_xml_input_nbest_mini_batch_1.sh @@ -0,0 +1,21 @@ +#!/bin/bash -x + +##################################################################### +# SUMMARY: Test constrained decoding with n-best output +# AUTHOR: snukky +##################################################################### + +# Exit on error +set -e + +# Remove old artifacts +rm -f nbest.{out,diff} + +# Run marian decoder +$MRT_MARIAN/marian-decoder -c $MRT_MODELS/wmt16_systems/marian.de-en.yml -b 4 -n --xml-input --n-best < text.in > nbest.out + +# Compare the output with the expected output +$MRT_TOOLS/diff-nums.py nbest.out nbest.expected -o nbest.diff + +# Exit with success code +exit 0 diff --git a/tests/decoder/xml/test_xml_input_nbest_mini_batch_8.sh b/tests/decoder/xml/test_xml_input_nbest_mini_batch_8.sh new file mode 100644 index 0000000..fff0224 --- /dev/null +++ b/tests/decoder/xml/test_xml_input_nbest_mini_batch_8.sh @@ -0,0 +1,21 @@ +#!/bin/bash -x + +##################################################################### +# SUMMARY: Test constrained decoding with larger mini-batch +# AUTHOR: snukky +##################################################################### + +# Exit on error +set -e + +# Remove old artifacts +rm -f batch8.{out,diff} + +# Run marian decoder +$MRT_MARIAN/marian-decoder -c $MRT_MODELS/wmt16_systems/marian.de-en.yml -b 4 -n --xml-input --n-best --mini-batch 8 < text.in > batch8.out + +# Compare the output with the expected output +$MRT_TOOLS/diff-nums.py batch8.out nbest.expected -o batch8.diff + +# Exit with success code +exit 0 diff --git a/tests/decoder/xml/test_xml_multiword.sh b/tests/decoder/xml/test_xml_multiword.sh new file mode 100644 index 0000000..37bf53c --- /dev/null +++ b/tests/decoder/xml/test_xml_multiword.sh @@ -0,0 +1,21 @@ +#!/bin/bash -x + +##################################################################### +# SUMMARY: Test XML decoding with multi-word constrains +# AUTHOR: snukky +##################################################################### + +# Exit on error +set -e + +# Remove old artifacts +rm -f multiword.{out,diff} + +# Run marian decoder +$MRT_MARIAN/marian-decoder -c $MRT_MODELS/wmt16_systems/marian.de-en.yml --mini-batch 10 -b 3 -n --xml-input --n-best < multiword.in > multiword.out + +# Compare the output with the expected output +$MRT_TOOLS/diff-nums.py multiword.out multiword.expected -o multiword.diff + +# Exit with success code +exit 0 diff --git a/tests/decoder/xml/test_xml_tags.sh b/tests/decoder/xml/test_xml_tags.sh new file mode 100644 index 0000000..a918c51 --- /dev/null +++ b/tests/decoder/xml/test_xml_tags.sh @@ -0,0 +1,21 @@ +#!/bin/bash -x + +##################################################################### +# SUMMARY: Test different variants of XML tags +# AUTHOR: snukky +##################################################################### + +# Exit on error +set -e + +# Remove old artifacts +rm -f tags.{out,diff} + +# Run marian decoder +$MRT_MARIAN/marian-decoder -c $MRT_MODELS/wmt16_systems/marian.de-en.yml -b 2 -n --xml-input < tags.in > tags.out + +# Compare the output with the expected output +$MRT_TOOLS/diff.sh tags.out tags.expected > tags.diff + +# Exit with success code +exit 0 diff --git a/tests/decoder/xml/test_xml_tags_via_server.sh b/tests/decoder/xml/test_xml_tags_via_server.sh new file mode 100644 index 0000000..354dd37 --- /dev/null +++ b/tests/decoder/xml/test_xml_tags_via_server.sh @@ -0,0 +1,32 @@ +#!/bin/bash -x + +##################################################################### +# SUMMARY: Test different variants of XML tags via web server +# AUTHOR: snukky +##################################################################### + +# Exit on error +set -e + +clean_up() { + kill $SERVER_PID +} +trap clean_up EXIT + +# Remove old artifacts +rm -f server.{out,diff} + +# Run marian server +$MRT_MARIAN/marian-server -c $MRT_MODELS/wmt16_systems/marian.de-en.yml -p 7766 -b 2 -n --xml-input > server.log 2>&1 & +SERVER_PID=$! + +sleep 15 + +python3 $MRT_MARIAN/../scripts/server/client_example.py -p 7766 < tags.in > server.out +kill $SERVER_PID + +# Compare the output with the expected output +$MRT_TOOLS/diff.sh server.out tags.expected > server.diff +test -e server.log +grep -q "listening on port 7766" server.log + diff --git a/tests/decoder/xml/test_xml_tags_word_alignment.sh b/tests/decoder/xml/test_xml_tags_word_alignment.sh new file mode 100644 index 0000000..a725972 --- /dev/null +++ b/tests/decoder/xml/test_xml_tags_word_alignment.sh @@ -0,0 +1,21 @@ +#!/bin/bash -x + +##################################################################### +# SUMMARY: Test returning word alignments with constrained decoding +# AUTHOR: snukky +##################################################################### + +# Exit on error +set -e + +# Remove old artifacts +rm -f align.{out,diff} + +# Run marian decoder +$MRT_MARIAN/marian-decoder -c $MRT_MODELS/wmt16_systems/marian.de-en.yml -b 3 -n --xml-input --alignment < tags.in > align.out + +# Compare the output with the expected output +$MRT_TOOLS/diff.sh align.out align.expected > align.diff + +# Exit with success code +exit 0 diff --git a/tests/decoder/xml/test_xml_violation_penalty.sh b/tests/decoder/xml/test_xml_violation_penalty.sh new file mode 100644 index 0000000..5d9506b --- /dev/null +++ b/tests/decoder/xml/test_xml_violation_penalty.sh @@ -0,0 +1,21 @@ +#!/bin/bash -x + +##################################################################### +# SUMMARY: Test constrained decoding with high violation penalty +# AUTHOR: snukky +##################################################################### + +# Exit on error +set -e + +# Remove old artifacts +rm -f vp.{out,diff} + +# Run marian decoder +$MRT_MARIAN/marian-decoder -c $MRT_MODELS/wmt16_systems/marian.de-en.yml --mini-batch 10 -b 3 --xml-input --n-best --xml-violation-penalty 99 < text.in > vp.out + +# Compare the output with the expected output +$MRT_TOOLS/diff-nums.py vp.out vp.expected -o vp.diff + +# Exit with success code +exit 0 diff --git a/tests/decoder/xml/text.in b/tests/decoder/xml/text.in new file mode 100644 index 0000000..be47224 --- /dev/null +++ b/tests/decoder/xml/text.in @@ -0,0 +1,10 @@ +<a translation="Trumpf">Obama</a> empfängt Netanyahu +das Verhältnis <a translation="between"> zwischen </a> Obama und Netanyahu ist nicht gerade freund@@ schaftlich . +die beiden wollten über die Umsetzung der internationalen <a translation="treaty"> Vereinbarung </a> sowie über Teher@@ ans destabil@@ isierende Maßnahmen im Nahen Osten sprechen . +bei der Begegnung soll es aber <a translation="about the conflict">auch um den Konflikt </a> mit den Palästinensern und die diskutierte Zwei Staaten Lösung gehen . +das Verhältnis zwischen Obama und Netanyahu ist seit Jahren gespannt . +Washington kritisiert den andauernden Siedlungsbau Israels und wirft Netanyahu mangelnden Willen beim Friedensprozess vor <a translation="!"> . </a> +durch den von Obama beworbenen Deal um das iranische Atomprogramm hat sich die Beziehung der beiden weiter verschlechtert . +im <a translation="April">März</a> hatte Netanyahu auf Einladung der Republikaner vor dem US Kongress eine umstrittene <a translation="show">Rede</a> gehalten . +die Rede war mit Obama nicht abgesprochen , ein <a translation="chicken">Treffen</a> hatte dieser mit Hinweis auf die seinerzeit bevorstehende Wahl in Israel abgelehnt . +in einem Not@@ ruf gesteht <a translation="teacher">Professor</a> , seine Freundin erschossen zu haben diff --git a/tests/decoder/xml/vp.expected b/tests/decoder/xml/vp.expected new file mode 100644 index 0000000..a8f522e --- /dev/null +++ b/tests/decoder/xml/vp.expected @@ -0,0 +1,30 @@ +0 ||| Obama welcomes Netanyahu Trumpf ||| F0= -18.9522 ||| 79.908 +0 ||| Obama Trumpf ||| F0= -16.7771 ||| 78.6403 +0 ||| Obama welcomes Netanyahu ||| F0= -0.72431 ||| -0.72431 +1 ||| the relationship between Obama and Netanyahu is not exactly friendly . ||| F0= -1.44457 ||| 94.6038 +1 ||| the relationship between Obama and Netanyahu is not exactly amic@@ able . ||| F0= -2.1021 ||| 93.9462 +1 ||| the relationship with Obama and Netanyahu is not exactly friendly . ||| F0= -8.51821 ||| -8.51821 +2 ||| the two wanted to talk about the implementation of the international treaty and of Tehran 's destabilising measures in the Middle East . ||| F0= -11.9959 ||| 84.2468 +2 ||| the two wanted to talk about the implementation of the international treaty and of Tehran 's destabilising actions in the Middle East . ||| F0= -12.4591 ||| 83.7836 +2 ||| the two wanted to talk about the implementation of the international agreement and of Tehran 's destabilising measures in the Middle East . ||| F0= -6.38888 ||| -6.38888 +3 ||| in the event of the meeting , it is also about the conflict with the Palestinians and the two states discussed . ||| F0= -13.6484 ||| 81.3094 +3 ||| in the event of the meeting , it is also about the conflict with the Palestinians and the two states being discussed . ||| F0= -14.3795 ||| 80.5783 +3 ||| in the event of the meeting , it is also important to resolve the conflict with the Palestinians and the two states discussed . about the conflict . ||| F0= -34.051 ||| 56.6942 +4 ||| the relationship between Obama and Netanyahu has been stretched for years . ||| F0= -2.51302 ||| -2.51302 +4 ||| the relationship between Obama and Netanyahu has been looking forward for years . ||| F0= -3.78594 ||| -3.78594 +4 ||| the relationship between Obama and Netanyahu has been looking forward to years . ||| F0= -3.96719 ||| -3.96719 +5 ||| Washington critici@@ zes Israel 's ongoing Israeli and acc@@ uses Netanyahu ! ||| F0= -24.0015 ||| 73.1676 +5 ||| Washington critici@@ zes Israel 's ongoing Israeli and ! ||| F0= -23.987 ||| 72.697 +5 ||| Washington critici@@ zes Israel 's ongoing Israeli and acc@@ uses Netanyahu of a ! ||| F0= -30.8404 ||| 63.6946 +6 ||| by the Obama deal to deal with Iran 's nuclear program , the two continued to deteriorate . ||| F0= -12.2352 ||| -12.2352 +6 ||| by the Obama deal to deal with Iran 's nuclear program , the relationship between the two continued to deteriorate . ||| F0= -13.3936 ||| -13.3936 +6 ||| by the Obama deal to deal with Iran 's nuclear program , the relationship between the two has continued to deteriorate . ||| F0= -13.7331 ||| -13.7331 +7 ||| in April , Netanyahu held a controversial speech at the Congressional Republicans ' invitation show before the U.S. Congress . ||| F0= -22.5869 ||| 169.9 +7 ||| in April , Netanyahu held a controversial speech at the Congressional Republicans ' invitation . ||| F0= -14.5929 ||| 82.6241 +7 ||| in March , Netanyahu held a controversial speech at the Congressional Republicans ' invitation . ||| F0= -5.94075 ||| -5.94075 +8 ||| the speech was not agreed with Obama , a meeting had been rejected , with reference to the chicken that was imminent . ||| F0= -28.4084 ||| 63.6838 +8 ||| the speech was not agreed with Obama , a meeting had been rejected , with reference to the election in Israel that was chicken at the time . ||| F0= -31.1594 ||| 60.9328 +8 ||| the speech was not agreed with Obama , a meeting had been rejected , with reference to the election in Israel that was imminent . ||| F0= -14.7121 ||| -14.7121 +9 ||| in an emergency call , teacher admits to having shot his girlfriend ||| F0= -15.4446 ||| 81.7128 +9 ||| in an emergency call , teacher admits to having shot dead with his girlfriend ||| F0= -15.849 ||| 81.3084 +9 ||| in an emergency call , Professor admits to have shot dead his girlfriend ||| F0= -5.15151 ||| -5.15151 |