From d6babfb8f3dc707473345664eb145128c6587272 Mon Sep 17 00:00:00 2001
From: Reuben Morais <reuben.morais@gmail.com>
Date: Mon, 1 Apr 2019 08:45:42 -0300
Subject: [PATCH] Speed up training tests and make sure they fully converge

---
 bin/run-tc-ldc93s1_checkpoint.sh          | 2 +-
 bin/run-tc-ldc93s1_new.sh                 | 9 ++++-----
 bin/run-tc-ldc93s1_singleshotinference.sh | 4 ++--
 bin/run-tc-ldc93s1_tflite.sh              | 4 ++--
 tc-train-tests.sh                         | 6 +++---
 5 files changed, 12 insertions(+), 13 deletions(-)

diff --git a/bin/run-tc-ldc93s1_checkpoint.sh b/bin/run-tc-ldc93s1_checkpoint.sh
index e8fdee8f..de92a304 100755
--- a/bin/run-tc-ldc93s1_checkpoint.sh
+++ b/bin/run-tc-ldc93s1_checkpoint.sh
@@ -16,7 +16,7 @@ python -u DeepSpeech.py --noshow_progressbar --noearly_stop \
   --train_files ${ldc93s1_csv} --train_batch_size 1 \
   --dev_files ${ldc93s1_csv} --dev_batch_size 1 \
   --test_files ${ldc93s1_csv} --test_batch_size 1 \
-  --n_hidden 494 --epoch -1 --random_seed 4567 --default_stddev 0.046875 \
+  --n_hidden 100 --epoch -1 \
   --max_to_keep 1 --checkpoint_dir '/tmp/ckpt' \
   --learning_rate 0.001 --dropout_rate 0.05 \
   --lm_binary_path 'data/smoke_test/vocab.pruned.lm' \
diff --git a/bin/run-tc-ldc93s1_new.sh b/bin/run-tc-ldc93s1_new.sh
index cee8ede4..f1e64d3e 100755
--- a/bin/run-tc-ldc93s1_new.sh
+++ b/bin/run-tc-ldc93s1_new.sh
@@ -14,12 +14,11 @@ fi;
 
 python -u DeepSpeech.py --noshow_progressbar --noearly_stop \
   --train_files ${ldc93s1_csv} --train_batch_size 1 \
-  --train_cached_features_path "/tmp/ldc93s1.hdf5" \
+  --train_cached_features_path '/tmp/ldc93s1_cache' \
   --dev_files ${ldc93s1_csv} --dev_batch_size 1 \
   --test_files ${ldc93s1_csv} --test_batch_size 1 \
-  --n_hidden 494 --epoch $epoch_count --random_seed 4567 \
-  --default_stddev 0.046875 --max_to_keep 1 \
-  --checkpoint_dir '/tmp/ckpt' \
+  --n_hidden 100 --epoch $epoch_count \
+  --max_to_keep 1 --checkpoint_dir '/tmp/ckpt' \
   --learning_rate 0.001 --dropout_rate 0.05  --export_dir '/tmp/train' \
   --lm_binary_path 'data/smoke_test/vocab.pruned.lm' \
-  --lm_trie_path 'data/smoke_test/vocab.trie' \
+  --lm_trie_path 'data/smoke_test/vocab.trie'
diff --git a/bin/run-tc-ldc93s1_singleshotinference.sh b/bin/run-tc-ldc93s1_singleshotinference.sh
index 7f6d2ba2..25e04f7d 100755
--- a/bin/run-tc-ldc93s1_singleshotinference.sh
+++ b/bin/run-tc-ldc93s1_singleshotinference.sh
@@ -14,7 +14,7 @@ python -u DeepSpeech.py --noshow_progressbar --noearly_stop \
   --train_files ${ldc93s1_csv} --train_batch_size 1 \
   --dev_files ${ldc93s1_csv} --dev_batch_size 1 \
   --test_files ${ldc93s1_csv} --test_batch_size 1 \
-  --n_hidden 494 --epoch 1 --random_seed 4567 --default_stddev 0.046875 \
+  --n_hidden 100 --epoch 1 \
   --max_to_keep 1 --checkpoint_dir '/tmp/ckpt' --checkpoint_secs 0 \
   --learning_rate 0.001 --dropout_rate 0.05 \
   --lm_binary_path 'data/smoke_test/vocab.pruned.lm' \
@@ -24,7 +24,7 @@ python -u DeepSpeech.py --noshow_progressbar --noearly_stop \
   --train_files ${ldc93s1_csv} --train_batch_size 1 \
   --dev_files ${ldc93s1_csv} --dev_batch_size 1 \
   --test_files ${ldc93s1_csv} --test_batch_size 1 \
-  --n_hidden 494 --epoch 1 --random_seed 4567 --default_stddev 0.046875 \
+  --n_hidden 100 --epoch 1 \
   --max_to_keep 1 --checkpoint_dir '/tmp/ckpt' --checkpoint_secs 0 \
   --learning_rate 0.001 --dropout_rate 0.05 \
   --lm_binary_path 'data/smoke_test/vocab.pruned.lm' \
diff --git a/bin/run-tc-ldc93s1_tflite.sh b/bin/run-tc-ldc93s1_tflite.sh
index 8a9f44ea..d99b4b59 100755
--- a/bin/run-tc-ldc93s1_tflite.sh
+++ b/bin/run-tc-ldc93s1_tflite.sh
@@ -11,10 +11,10 @@ if [ ! -f "${ldc93s1_dir}/ldc93s1.csv" ]; then
 fi;
 
 python -u DeepSpeech.py --noshow_progressbar \
-  --n_hidden 494 \
+  --n_hidden 100 \
   --checkpoint_dir '/tmp/ckpt' \
   --export_dir '/tmp/train_tflite' \
   --lm_binary_path 'data/smoke_test/vocab.pruned.lm' \
   --lm_trie_path 'data/smoke_test/vocab.trie' \
   --notrain --notest \
-  --export_tflite --nouse_seq_length \
+  --export_tflite --nouse_seq_length
diff --git a/tc-train-tests.sh b/tc-train-tests.sh
index ec74b03d..a08624b4 100644
--- a/tc-train-tests.sh
+++ b/tc-train-tests.sh
@@ -58,8 +58,8 @@ LD_LIBRARY_PATH=${PY37_LDPATH}:$LD_LIBRARY_PATH pip install --verbose --only-bin
 
 pushd ${HOME}/DeepSpeech/ds/
     # Run twice to test preprocessed features
-    time ./bin/run-tc-ldc93s1_new.sh 104
-    time ./bin/run-tc-ldc93s1_new.sh 105
+    time ./bin/run-tc-ldc93s1_new.sh 199
+    time ./bin/run-tc-ldc93s1_new.sh 200
     time ./bin/run-tc-ldc93s1_tflite.sh
 popd
 
@@ -75,7 +75,7 @@ if [ ! -z "${CONVERT_GRAPHDEF_MEMMAPPED}" ]; then
 fi;
 
 pushd ${HOME}/DeepSpeech/ds/
-    time ./bin/run-tc-ldc93s1_checkpoint.sh 105
+    time ./bin/run-tc-ldc93s1_checkpoint.sh 200
 popd
 
 deactivate