Welcome to mirror list, hosted at ThFree Co, Russian Federation.

github.com/marian-nmt/marian-regression-tests.git - Unnamed repository; edit this file 'description' to name the repository.
summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorRoman Grundkiewicz <rgrundkiewicz@gmail.com>2021-04-16 11:30:24 +0300
committerRoman Grundkiewicz <rgrundkiewicz@gmail.com>2021-04-16 11:30:24 +0300
commit0f658a21a373de7ff33d359666ae30a4745fe052 (patch)
treef0c180dd71476a35f3f26401cb9440e1d4b874ea
parent766db5dd076470d3cbd03ff0cf077a7cb2cc04d4 (diff)
Fix devset
-rw-r--r--tests/training/validation/stop_on_1st.expected18
-rw-r--r--tests/training/validation/stop_on_all.expected18
-rw-r--r--tests/training/validation/stop_on_any.expected14
-rw-r--r--tests/training/validation/test_early_stopping_on_1st.sh2
-rw-r--r--tests/training/validation/test_early_stopping_on_all.sh2
-rw-r--r--tests/training/validation/test_early_stopping_on_any.sh2
6 files changed, 30 insertions, 26 deletions
diff --git a/tests/training/validation/stop_on_1st.expected b/tests/training/validation/stop_on_1st.expected
index ee21ffd..394c115 100644
--- a/tests/training/validation/stop_on_1st.expected
+++ b/tests/training/validation/stop_on_1st.expected
@@ -1,18 +1,18 @@
[valid] Ep. 1 : Up. 10 : valid-script : 111.4 : new best
-[valid] Ep. 1 : Up. 10 : ce-mean-words : 2.74949e-41 : new best
+[valid] Ep. 1 : Up. 10 : ce-mean-words : 2.61662 : new best
[valid] Ep. 1 : Up. 20 : valid-script : 111.3 : stalled 1 times (last best: 111.4)
-[valid] Ep. 1 : Up. 20 : ce-mean-words : 2.74949e-41 : stalled 1 times (last best: 2.74949e-41)
+[valid] Ep. 1 : Up. 20 : ce-mean-words : 2.58011 : new best
[valid] Ep. 1 : Up. 30 : valid-script : 111.2 : stalled 2 times (last best: 111.4)
-[valid] Ep. 1 : Up. 30 : ce-mean-words : 2.74949e-41 : stalled 2 times (last best: 2.74949e-41)
+[valid] Ep. 1 : Up. 30 : ce-mean-words : 2.56738 : new best
[valid] Ep. 1 : Up. 40 : valid-script : 111.1 : stalled 3 times (last best: 111.4)
-[valid] Ep. 1 : Up. 40 : ce-mean-words : 2.74949e-41 : stalled 3 times (last best: 2.74949e-41)
+[valid] Ep. 1 : Up. 40 : ce-mean-words : 2.59199 : stalled 1 times (last best: 2.56738)
[valid] Ep. 1 : Up. 50 : valid-script : 111.6 : new best
-[valid] Ep. 1 : Up. 50 : ce-mean-words : 2.74949e-41 : stalled 4 times (last best: 2.74949e-41)
+[valid] Ep. 1 : Up. 50 : ce-mean-words : 2.58017 : stalled 2 times (last best: 2.56738)
[valid] Ep. 1 : Up. 60 : valid-script : 111.5 : stalled 1 times (last best: 111.6)
-[valid] Ep. 1 : Up. 60 : ce-mean-words : 2.74949e-41 : stalled 5 times (last best: 2.74949e-41)
+[valid] Ep. 1 : Up. 60 : ce-mean-words : 2.60083 : stalled 3 times (last best: 2.56738)
[valid] Ep. 1 : Up. 70 : valid-script : 111.4 : stalled 2 times (last best: 111.6)
-[valid] Ep. 1 : Up. 70 : ce-mean-words : 2.74949e-41 : stalled 6 times (last best: 2.74949e-41)
+[valid] Ep. 1 : Up. 70 : ce-mean-words : 2.67288 : stalled 4 times (last best: 2.56738)
[valid] Ep. 1 : Up. 80 : valid-script : 111.3 : stalled 3 times (last best: 111.6)
-[valid] Ep. 1 : Up. 80 : ce-mean-words : 2.74949e-41 : stalled 7 times (last best: 2.74949e-41)
+[valid] Ep. 1 : Up. 80 : ce-mean-words : 2.71896 : stalled 5 times (last best: 2.56738)
[valid] Ep. 1 : Up. 90 : valid-script : 111.2 : stalled 4 times (last best: 111.6)
-[valid] Ep. 1 : Up. 90 : ce-mean-words : 2.74949e-41 : stalled 8 times (last best: 2.74949e-41)
+[valid] Ep. 1 : Up. 90 : ce-mean-words : 2.8146 : stalled 6 times (last best: 2.56738)
diff --git a/tests/training/validation/stop_on_all.expected b/tests/training/validation/stop_on_all.expected
index 85f682f..9aae90f 100644
--- a/tests/training/validation/stop_on_all.expected
+++ b/tests/training/validation/stop_on_all.expected
@@ -1,18 +1,18 @@
-[valid] Ep. 1 : Up. 10 : ce-mean-words : 2.74949e-41 : new best
+[valid] Ep. 1 : Up. 10 : ce-mean-words : 2.61662 : new best
[valid] Ep. 1 : Up. 10 : valid-script : 111.4 : new best
-[valid] Ep. 1 : Up. 20 : ce-mean-words : 2.74949e-41 : stalled 1 times (last best: 2.74949e-41)
+[valid] Ep. 1 : Up. 20 : ce-mean-words : 2.58011 : new best
[valid] Ep. 1 : Up. 20 : valid-script : 111.3 : stalled 1 times (last best: 111.4)
-[valid] Ep. 1 : Up. 30 : ce-mean-words : 2.74949e-41 : stalled 2 times (last best: 2.74949e-41)
+[valid] Ep. 1 : Up. 30 : ce-mean-words : 2.56738 : new best
[valid] Ep. 1 : Up. 30 : valid-script : 111.2 : stalled 2 times (last best: 111.4)
-[valid] Ep. 1 : Up. 40 : ce-mean-words : 2.74949e-41 : stalled 3 times (last best: 2.74949e-41)
+[valid] Ep. 1 : Up. 40 : ce-mean-words : 2.59199 : stalled 1 times (last best: 2.56738)
[valid] Ep. 1 : Up. 40 : valid-script : 111.1 : stalled 3 times (last best: 111.4)
-[valid] Ep. 1 : Up. 50 : ce-mean-words : 2.74949e-41 : stalled 4 times (last best: 2.74949e-41)
+[valid] Ep. 1 : Up. 50 : ce-mean-words : 2.58017 : stalled 2 times (last best: 2.56738)
[valid] Ep. 1 : Up. 50 : valid-script : 111.6 : new best
-[valid] Ep. 1 : Up. 60 : ce-mean-words : 2.74949e-41 : stalled 5 times (last best: 2.74949e-41)
+[valid] Ep. 1 : Up. 60 : ce-mean-words : 2.60083 : stalled 3 times (last best: 2.56738)
[valid] Ep. 1 : Up. 60 : valid-script : 111.5 : stalled 1 times (last best: 111.6)
-[valid] Ep. 1 : Up. 70 : ce-mean-words : 2.74949e-41 : stalled 6 times (last best: 2.74949e-41)
+[valid] Ep. 1 : Up. 70 : ce-mean-words : 2.67288 : stalled 4 times (last best: 2.56738)
[valid] Ep. 1 : Up. 70 : valid-script : 111.4 : stalled 2 times (last best: 111.6)
-[valid] Ep. 1 : Up. 80 : ce-mean-words : 2.74949e-41 : stalled 7 times (last best: 2.74949e-41)
+[valid] Ep. 1 : Up. 80 : ce-mean-words : 2.71896 : stalled 5 times (last best: 2.56738)
[valid] Ep. 1 : Up. 80 : valid-script : 111.3 : stalled 3 times (last best: 111.6)
-[valid] Ep. 1 : Up. 90 : ce-mean-words : 2.74949e-41 : stalled 8 times (last best: 2.74949e-41)
+[valid] Ep. 1 : Up. 90 : ce-mean-words : 2.8146 : stalled 6 times (last best: 2.56738)
[valid] Ep. 1 : Up. 90 : valid-script : 111.2 : stalled 4 times (last best: 111.6)
diff --git a/tests/training/validation/stop_on_any.expected b/tests/training/validation/stop_on_any.expected
index 1afe151..f860950 100644
--- a/tests/training/validation/stop_on_any.expected
+++ b/tests/training/validation/stop_on_any.expected
@@ -1,10 +1,14 @@
[valid] Ep. 1 : Up. 10 : valid-script : 111.4 : new best
-[valid] Ep. 1 : Up. 10 : ce-mean-words : 2.74949e-41 : new best
+[valid] Ep. 1 : Up. 10 : ce-mean-words : 2.61662 : new best
[valid] Ep. 1 : Up. 20 : valid-script : 111.3 : stalled 1 times (last best: 111.4)
-[valid] Ep. 1 : Up. 20 : ce-mean-words : 2.74949e-41 : stalled 1 times (last best: 2.74949e-41)
+[valid] Ep. 1 : Up. 20 : ce-mean-words : 2.58011 : new best
[valid] Ep. 1 : Up. 30 : valid-script : 111.2 : stalled 2 times (last best: 111.4)
-[valid] Ep. 1 : Up. 30 : ce-mean-words : 2.74949e-41 : stalled 2 times (last best: 2.74949e-41)
+[valid] Ep. 1 : Up. 30 : ce-mean-words : 2.56738 : new best
[valid] Ep. 1 : Up. 40 : valid-script : 111.1 : stalled 3 times (last best: 111.4)
-[valid] Ep. 1 : Up. 40 : ce-mean-words : 2.74949e-41 : stalled 3 times (last best: 2.74949e-41)
+[valid] Ep. 1 : Up. 40 : ce-mean-words : 2.59199 : stalled 1 times (last best: 2.56738)
[valid] Ep. 1 : Up. 50 : valid-script : 111.6 : new best
-[valid] Ep. 1 : Up. 50 : ce-mean-words : 2.74949e-41 : stalled 4 times (last best: 2.74949e-41)
+[valid] Ep. 1 : Up. 50 : ce-mean-words : 2.58017 : stalled 2 times (last best: 2.56738)
+[valid] Ep. 1 : Up. 60 : valid-script : 111.5 : stalled 1 times (last best: 111.6)
+[valid] Ep. 1 : Up. 60 : ce-mean-words : 2.60083 : stalled 3 times (last best: 2.56738)
+[valid] Ep. 1 : Up. 70 : valid-script : 111.4 : stalled 2 times (last best: 111.6)
+[valid] Ep. 1 : Up. 70 : ce-mean-words : 2.67288 : stalled 4 times (last best: 2.56738)
diff --git a/tests/training/validation/test_early_stopping_on_1st.sh b/tests/training/validation/test_early_stopping_on_1st.sh
index 1f80ca1..0ebe779 100644
--- a/tests/training/validation/test_early_stopping_on_1st.sh
+++ b/tests/training/validation/test_early_stopping_on_1st.sh
@@ -27,7 +27,7 @@ $MRT_MARIAN/marian \
--disp-freq 5 --valid-freq 10 --after-batches 200 \
--valid-metrics valid-script ce-mean-words \
--valid-script-path ./stop_on_script.sh \
- --valid-sets $MRT_DATA/europarl.de-en/toy.bpe.{en,de} \
+ --valid-sets devset.{en,de} \
--valid-log stop_on_1st.log \
--early-stopping 4 --early-stopping-on first
diff --git a/tests/training/validation/test_early_stopping_on_all.sh b/tests/training/validation/test_early_stopping_on_all.sh
index 502b7fa..e6f95d9 100644
--- a/tests/training/validation/test_early_stopping_on_all.sh
+++ b/tests/training/validation/test_early_stopping_on_all.sh
@@ -27,7 +27,7 @@ $MRT_MARIAN/marian \
--disp-freq 5 --valid-freq 10 --after-batches 200 \
--valid-metrics ce-mean-words valid-script \
--valid-script-path ./stop_on_script.sh \
- --valid-sets $MRT_DATA/europarl.de-en/toy.bpe.{en,de} \
+ --valid-sets devset.{en,de} \
--valid-log stop_on_all.log \
--early-stopping 4 --early-stopping-on all
diff --git a/tests/training/validation/test_early_stopping_on_any.sh b/tests/training/validation/test_early_stopping_on_any.sh
index 17e9e14..d1f83db 100644
--- a/tests/training/validation/test_early_stopping_on_any.sh
+++ b/tests/training/validation/test_early_stopping_on_any.sh
@@ -27,7 +27,7 @@ $MRT_MARIAN/marian \
--disp-freq 5 --valid-freq 10 --after-batches 200 \
--valid-metrics valid-script ce-mean-words \
--valid-script-path ./stop_on_script.sh \
- --valid-sets $MRT_DATA/europarl.de-en/toy.bpe.{en,de} \
+ --valid-sets devset.{en,de} \
--valid-log stop_on_any.log \
--early-stopping 4 --early-stopping-on any