Merge branch 'tests-logical-epochs'

8 files changed, 155 insertions, 0 deletions

diff --git a/tests/training/scheduler/.gitignore b/tests/training/scheduler/.gitignore
new file mode 100644
index 0000000..7958534
--- /dev/null
+++ b/tests/training/scheduler/.gitignore
@@ -0,0 +1,2 @@
+log_epoch_[etu]
+train.??.gz
diff --git a/tests/training/scheduler/log_epoch_e.expected b/tests/training/scheduler/log_epoch_e.expected
new file mode 100644
index 0000000..87cf167
--- /dev/null
+++ b/tests/training/scheduler/log_epoch_e.expected
@@ -0,0 +1,20 @@
+Training started
+Seen 1542 samples
+Starting data epoch 2 in logical epoch 1.000
+Ep. 1.000 : Up. 10 : Sen. 768 : Cost 9.68880177 * 61,315 after 61,315
+Seen 1542 samples
+Starting data epoch 3 in logical epoch 1.500
+Ep. 1.500 : Up. 20 : Sen. 1,536 : Cost 9.67091751 * 61,279 after 122,594
+Seen 1542 samples
+Starting data epoch 4 in logical epoch 2.000
+Seen 1542 samples
+Starting data epoch 5 in logical epoch 2.500
+Ep. 2.500 : Up. 30 : Sen. 512 : Cost 9.65089989 * 54,621 after 177,215
+Seen 1542 samples
+Starting data epoch 6 in logical epoch 3.000
+Ep. 3.000 : Up. 40 : Sen. 1,280 : Cost 9.63199997 * 61,545 after 238,760
+Seen 1542 samples
+Starting data epoch 7 in logical epoch 3.500
+Training finished
+Saving model to log_epoch_e/model.npz
+Saving Adam parameters to log_epoch_e/model.npz.optimizer.npz
diff --git a/tests/training/scheduler/log_epoch_t.expected b/tests/training/scheduler/log_epoch_t.expected
new file mode 100644
index 0000000..1f57c2e
--- /dev/null
+++ b/tests/training/scheduler/log_epoch_t.expected
@@ -0,0 +1,12 @@
+Training started
+Ep. 2.258 : Up. 4 : Sen. 512 : Cost 9.69286919 * 13,547 after 13,547
+Ep. 3.400 : Up. 6 : Sen. 768 : Cost 9.68953419 * 6,851 after 20,398
+Ep. 5.131 : Up. 9 : Sen. 1,152 : Cost 9.68455887 * 10,387 after 30,785
+Ep. 6.793 : Up. 12 : Sen. 1,536 : Cost 9.68291855 * 9,975 after 40,760
+Seen 1542 samples
+Starting data epoch 2 in logical epoch 6.819
+Ep. 8.472 : Up. 16 : Sen. 384 : Cost 9.67040443 * 10,074 after 50,834
+Ep. 10.219 : Up. 19 : Sen. 768 : Cost 9.66528606 * 10,481 after 61,315
+Training finished
+Saving model to log_epoch_t/model.npz
+Saving Adam parameters to log_epoch_t/model.npz.optimizer.npz
diff --git a/tests/training/scheduler/log_epoch_u.expected b/tests/training/scheduler/log_epoch_u.expected
new file mode 100644
index 0000000..a8855f2
--- /dev/null
+++ b/tests/training/scheduler/log_epoch_u.expected
@@ -0,0 +1,15 @@
+Training started
+Seen 1542 samples
+Starting data epoch 2 in logical epoch 0.700
+Ep. 1.000 : Up. 10 : Sen. 768 : Cost 9.68880177 * 61,315 after 61,315
+Seen 1542 samples
+Starting data epoch 3 in logical epoch 1.400
+Ep. 2.000 : Up. 20 : Sen. 1,536 : Cost 9.67091751 * 61,279 after 122,594
+Seen 1542 samples
+Starting data epoch 4 in logical epoch 2.100
+Seen 1542 samples
+Starting data epoch 5 in logical epoch 2.800
+Ep. 3.000 : Up. 30 : Sen. 512 : Cost 9.65089989 * 54,621 after 177,215
+Training finished
+Saving model to log_epoch_u/model.npz
+Saving Adam parameters to log_epoch_u/model.npz.optimizer.npz
diff --git a/tests/training/scheduler/setup.sh b/tests/training/scheduler/setup.sh
new file mode 100644
index 0000000..284e7c1
--- /dev/null
+++ b/tests/training/scheduler/setup.sh
@@ -0,0 +1,10 @@
+# Skip if compiled without SentencePiece
+test -n "$MRT_MARIAN_USE_SENTENCEPIECE" || exit 100
+
+test -f $MRT_DATA/europarl.de-en/corpus.bpe.de || exit 1
+test -f $MRT_DATA/europarl.de-en/corpus.bpe.en || exit 1
+
+test -f train.de.gz || cat $MRT_DATA/europarl.de-en/corpus.bpe.de | sed 's/@@ //g' | head -n 2000 | gzip > train.de.gz
+test -f train.en.gz || cat $MRT_DATA/europarl.de-en/corpus.bpe.en | sed 's/@@ //g' | head -n 2000 | gzip > train.en.gz
+
+test -f $MRT_MODELS/rnn-spm/vocab.deen.spm || exit 1
diff --git a/tests/training/scheduler/test_logical_epoch.sh b/tests/training/scheduler/test_logical_epoch.sh
new file mode 100644
index 0000000..fca9f52
--- /dev/null
+++ b/tests/training/scheduler/test_logical_epoch.sh
@@ -0,0 +1,32 @@
+#!/bin/bash -x
+
+#####################################################################
+# SUMMARY: Test logical epoch defined via data epoch
+# AUTHOR: snukky
+# TAGS: sentencepiece stopping after logical-epoch
+#####################################################################
+
+# Exit on error
+set -e
+
+# Remove old artifacts and create working directory
+rm -rf log_epoch_e log_epoch_e.*{log,out,diff}
+mkdir -p log_epoch_e
+
+# Run marian command
+$MRT_MARIAN/marian \
+    --no-shuffle --seed 1111 --dim-emb 32 --dim-rnn 64 --maxi-batch 1 --maxi-batch-sort none \
+    -m log_epoch_e/model.npz -t train.{de,en}.gz -v $MRT_MODELS/rnn-spm/vocab.{deen,deen}.spm \
+    --mini-batch 256 --logical-epoch 2e --log log_epoch_e.log --after 3e \
+    --disp-freq 10u
+
+# Check if files exist
+test -e log_epoch_e/model.npz
+test -e log_epoch_e.log
+
+# Compare actual and expected outputs
+cat log_epoch_e.log | $MRT_TOOLS/strip-timestamps.sh | grep -v '^\[' | sed 's/ : Time.*//' > log_epoch_e.out
+$MRT_TOOLS/diff-nums.py log_epoch_e.out log_epoch_e.expected -p 0.01 -o log_epoch_e.diff
+
+# Exit with success code
+exit 0
diff --git a/tests/training/scheduler/test_logical_epoch_labels.sh b/tests/training/scheduler/test_logical_epoch_labels.sh
new file mode 100644
index 0000000..c37c0fa
--- /dev/null
+++ b/tests/training/scheduler/test_logical_epoch_labels.sh
@@ -0,0 +1,32 @@
+#!/bin/bash -x
+
+#####################################################################
+# SUMMARY: Test logical epoch defined via labels
+# AUTHOR: snukky
+# TAGS: sentencepiece stopping after logical-epoch
+#####################################################################
+
+# Exit on error
+set -e
+
+# Remove old artifacts and create working directory
+rm -rf log_epoch_t log_epoch_t.*{log,out,diff}
+mkdir -p log_epoch_t
+
+# Run marian command
+$MRT_MARIAN/marian \
+    --no-shuffle --seed 1111 --dim-emb 32 --dim-rnn 64 --maxi-batch 1 --maxi-batch-sort none \
+    -m log_epoch_t/model.npz -t train.{de,en}.gz -v $MRT_MODELS/rnn-spm/vocab.{deen,deen}.spm \
+    --mini-batch 128 --logical-epoch 6kt --log log_epoch_t.log --after 10e \
+    --disp-freq 10kt
+
+# Check if files exist
+test -e log_epoch_t/model.npz
+test -e log_epoch_t.log
+
+# Compare actual and expected outputs
+cat log_epoch_t.log | $MRT_TOOLS/strip-timestamps.sh | grep -v '^\[' | sed 's/ : Time.*//' > log_epoch_t.out
+$MRT_TOOLS/diff-nums.py log_epoch_t.out log_epoch_t.expected -p 0.01 -o log_epoch_t.diff
+
+# Exit with success code
+exit 0
diff --git a/tests/training/scheduler/test_logical_epoch_updates.sh b/tests/training/scheduler/test_logical_epoch_updates.sh
new file mode 100644
index 0000000..8582120
--- /dev/null
+++ b/tests/training/scheduler/test_logical_epoch_updates.sh
@@ -0,0 +1,32 @@
+#!/bin/bash -x
+
+#####################################################################
+# SUMMARY: Test logical epoch defined via updates
+# AUTHOR: snukky
+# TAGS: sentencepiece stopping after logical-epoch
+#####################################################################
+
+# Exit on error
+set -e
+
+# Remove old artifacts and create working directory
+rm -rf log_epoch_u log_epoch_u.*{log,out,diff}
+mkdir -p log_epoch_u
+
+# Run marian command
+$MRT_MARIAN/marian \
+    --no-shuffle --seed 1111 --dim-emb 32 --dim-rnn 64 --maxi-batch 1 --maxi-batch-sort none \
+    -m log_epoch_u/model.npz -t train.{de,en}.gz -v $MRT_MODELS/rnn-spm/vocab.{deen,deen}.spm \
+    --mini-batch 256 --logical-epoch 10u --log log_epoch_u.log --after 3e \
+    --disp-freq 10u
+
+# Check if files exist
+test -e log_epoch_u/model.npz
+test -e log_epoch_u.log
+
+# Compare actual and expected outputs
+cat log_epoch_u.log | $MRT_TOOLS/strip-timestamps.sh | grep -v '^\[' | sed 's/ : Time.*//' > log_epoch_u.out
+$MRT_TOOLS/diff-nums.py log_epoch_u.out log_epoch_u.expected -p 0.01 -o log_epoch_u.diff
+
+# Exit with success code
+exit 0