Add OpenNMT configs for training and for sampling

chikiuso · Feb 16, 2019 · 63b659b · 63b659b
1 parent ffb0a43
commit 63b659b
Show file tree

Hide file tree

Showing 7 changed files with 81 additions and 1 deletion.
diff --git a/train/opennmt_config.yml b/train/opennmt_config.yml
@@ -0,0 +1,34 @@
+data:
+  train_features_file: ../../data/data.train.bpe.shuf.titles
+  train_labels_file: ../../data/data.train.bpe.shuf.comments
+  eval_features_file: ../../data/data.dev.bpe.titles
+  eval_labels_file: ../../data/data.dev.bpe.comments
+
+  source_words_vocabulary: ../../data/vocab.titles
+  target_words_vocabulary: ../../data/vocab.comments
+
+params:
+  average_loss_in_time: true
+  label_smoothing: 0.1
+  optimizer: LazyAdamOptimizer
+  optimizer_params:
+    beta1: 0.9
+    beta2: 0.998
+  learning_rate: 2
+  decay_type: noam_decay_v2
+  decay_params:
+    model_dim: 512
+    warmup_steps: 8000
+
+train:
+  effective_batch_size: 8192
+  batch_size: 4096
+  batch_type: tokens
+  maximum_features_length: 300
+  maximum_labels_length: 300
+  keep_checkpoint_max: 2
+  average_last_checkpoints: 8
+  train_steps: 300000
+
+eval:
+  eval_delay: 7200
diff --git a/train/opennmt_config_larger_batch.yml b/train/opennmt_config_larger_batch.yml
@@ -0,0 +1,34 @@
+data:
+  train_features_file: ../../data/data.train.bpe.shuf.titles
+  train_labels_file: ../../data/data.train.bpe.shuf.comments
+  eval_features_file: ../../data/data.dev.bpe.titles
+  eval_labels_file: ../../data/data.dev.bpe.comments
+
+  source_words_vocabulary: ../../data/vocab.titles
+  target_words_vocabulary: ../../data/vocab.comments
+
+params:
+  average_loss_in_time: true
+  label_smoothing: 0.1
+  optimizer: LazyAdamOptimizer
+  optimizer_params:
+    beta1: 0.9
+    beta2: 0.998
+  learning_rate: 2
+  decay_type: noam_decay_v2
+  decay_params:
+    model_dim: 512
+    warmup_steps: 8000
+
+train:
+  effective_batch_size: 32384
+  batch_size: 4096
+  batch_type: tokens
+  maximum_features_length: 300
+  maximum_labels_length: 300
+  keep_checkpoint_max: 2
+  average_last_checkpoints: 8
+  train_steps: 300000
+
+eval:
+  eval_delay: 7200
diff --git a/train/sample.yml b/train/sample.yml
@@ -1,3 +1,3 @@
 params:
-  sampling_topk: 0
+  sampling_topk: 10
   beam_width: 1
diff --git a/train/sample100.yml b/train/sample100.yml
@@ -0,0 +1,3 @@
+params:
+  sampling_topk: 100
+  beam_width: 1
diff --git a/train/sample2.yml b/train/sample2.yml
@@ -0,0 +1,3 @@
+params:
+  sampling_topk: 2
+  beam_width: 1
diff --git a/train/sample30.yml b/train/sample30.yml
@@ -0,0 +1,3 @@
+params:
+  sampling_topk: 30
+  beam_width: 1
diff --git a/train/sample_full.yml b/train/sample_full.yml
@@ -0,0 +1,3 @@
+params:
+  sampling_topk: 0
+  beam_width: 1