Embed default beam width into exported graph and remove param from DS_CreateModel

DeepSpeech.py

@@ -797,6 +797,7 @@ def export():
     outputs['metadata_sample_rate'] = tf.constant([FLAGS.audio_sample_rate], name='metadata_sample_rate')
     outputs['metadata_feature_win_len'] = tf.constant([FLAGS.feature_win_len], name='metadata_feature_win_len')
     outputs['metadata_feature_win_step'] = tf.constant([FLAGS.feature_win_step], name='metadata_feature_win_step')
+    outputs['metadata_beam_width'] = tf.constant([FLAGS.export_beam_width], name='metadata_beam_width')
     outputs['metadata_alphabet'] = tf.constant([Config.alphabet.serialize()], name='metadata_alphabet')
 
     if FLAGS.export_language:

native_client/deepspeech.cc

@@ -257,7 +257,6 @@ StreamingState::processBatch(const vector<float>& buf, unsigned int n_steps)
 
 int
 DS_CreateModel(const char* aModelPath,
-               unsigned int aBeamWidth,
                ModelState** retval)
 {
   *retval = nullptr;
@@ -282,7 +281,7 @@ DS_CreateModel(const char* aModelPath,
     return DS_ERR_FAIL_CREATE_MODEL;
   }
 
-  int err = model->init(aModelPath, aBeamWidth);
+  int err = model->init(aModelPath);
   if (err != DS_ERR_OK) {
     return err;
   }
@@ -291,6 +290,19 @@ DS_CreateModel(const char* aModelPath,
   return DS_ERR_OK;
 }
 
+unsigned int
+DS_GetModelBeamWidth(ModelState* aCtx)
+{
+  return aCtx->beam_width_;
+}
+
+int
+DS_SetModelBeamWidth(ModelState* aCtx, unsigned int aBeamWidth)
+{
+  aCtx->beam_width_ = aBeamWidth;
+  return 0;
+}
+
 int
 DS_GetModelSampleRate(ModelState* aCtx)
 {

native_client/deepspeech.h

@@ -78,18 +78,39 @@ enum DeepSpeech_Error_Codes
  * @brief An object providing an interface to a trained DeepSpeech model.
  *
  * @param aModelPath The path to the frozen model graph.
- * @param aBeamWidth The beam width used by the decoder. A larger beam
- *                   width generates better results at the cost of decoding
- *                   time.
  * @param[out] retval a ModelState pointer
  *
  * @return Zero on success, non-zero on failure.
  */
 DEEPSPEECH_EXPORT
 int DS_CreateModel(const char* aModelPath,
-                   unsigned int aBeamWidth,
                    ModelState** retval);
 
+/**
+ * @brief Get beam width value used by the model. If {@link DS_SetModelBeamWidth}
+ *        was not called before, will return the default value loaded from the
+ *        model file.
+ *
+ * @param aCtx A ModelState pointer created with {@link DS_CreateModel}.
+ *
+ * @return Beam width value used by the model.
+ */
+DEEPSPEECH_EXPORT
+unsigned int DS_GetModelBeamWidth(ModelState* aCtx);
+
+/**
+ * @brief Set beam width value used by the model.
+ *
+ * @param aCtx A ModelState pointer created with {@link DS_CreateModel}.
+ * @param aBeamWidth The beam width used by the model. A larger beam width value
+ *                   generates better results at the cost of decoding time.
+ *
+ * @return Zero on success, non-zero on failure.
+ */
+DEEPSPEECH_EXPORT
+int DS_SetModelBeamWidth(ModelState* aCtx,
+                         unsigned int aBeamWidth);
+
 /**
  * @brief Return the sample rate expected by a model.
  *

native_client/modelstate.cc

@@ -24,10 +24,8 @@ ModelState::~ModelState()
 }
 
 int
-ModelState::init(const char* model_path,
-                 unsigned int beam_width)
+ModelState::init(const char* model_path)
 {
-  beam_width_ = beam_width;
   return DS_ERR_OK;
 }
 

native_client/modelstate.h

@@ -30,7 +30,7 @@ struct ModelState {
   ModelState();
   virtual ~ModelState();
 
-  virtual int init(const char* model_path, unsigned int beam_width);
+  virtual int init(const char* model_path);
 
   virtual void compute_mfcc(const std::vector<float>& audio_buffer, std::vector<float>& mfcc_output) = 0;
 

native_client/tflitemodelstate.cc

@@ -91,10 +91,9 @@ TFLiteModelState::~TFLiteModelState()
 }
 
 int
-TFLiteModelState::init(const char* model_path,
-                       unsigned int beam_width)
+TFLiteModelState::init(const char* model_path)
 {
-  int err = ModelState::init(model_path, beam_width);
+  int err = ModelState::init(model_path);
   if (err != DS_ERR_OK) {
     return err;
   }
@@ -130,6 +129,7 @@ TFLiteModelState::init(const char* model_path,
   int metadata_feature_win_len_idx  = get_output_tensor_by_name("metadata_feature_win_len");
   int metadata_feature_win_step_idx = get_output_tensor_by_name("metadata_feature_win_step");
   int metadata_alphabet_idx = get_output_tensor_by_name("metadata_alphabet");
+  int metadata_beam_width_idx = get_output_tensor_by_name("metadata_beam_width");
 
   std::vector<int> metadata_exec_plan;
   metadata_exec_plan.push_back(find_parent_node_ids(metadata_version_idx)[0]);
@@ -137,6 +137,7 @@ TFLiteModelState::init(const char* model_path,
   metadata_exec_plan.push_back(find_parent_node_ids(metadata_feature_win_len_idx)[0]);
   metadata_exec_plan.push_back(find_parent_node_ids(metadata_feature_win_step_idx)[0]);
   metadata_exec_plan.push_back(find_parent_node_ids(metadata_alphabet_idx)[0]);
+  metadata_exec_plan.push_back(find_parent_node_ids(metadata_beam_width_idx)[0]);
 
   for (int i = 0; i < metadata_exec_plan.size(); ++i) {
     assert(metadata_exec_plan[i] > -1);
@@ -207,9 +208,14 @@ TFLiteModelState::init(const char* model_path,
     return DS_ERR_INVALID_ALPHABET;
   }
 
+  int* const beam_width = interpreter_->typed_tensor<int>(metadata_beam_width_idx);
+  beam_width_ = (unsigned int)(*beam_width);
+
   assert(sample_rate_ > 0);
   assert(audio_win_len_ > 0);
   assert(audio_win_step_ > 0);
+  assert(alphabet_.GetSize() > 0);
+  assert(beam_width_ > 0);
 
   TfLiteIntArray* dims_input_node = interpreter_->tensor(input_node_idx_)->dims;
 

native_client/tflitemodelstate.h

@@ -30,8 +30,7 @@ struct TFLiteModelState : public ModelState
   TFLiteModelState();
   virtual ~TFLiteModelState();
 
-  virtual int init(const char* model_path,
-                   unsigned int beam_width) override;
+  virtual int init(const char* model_path) override;
 
   virtual void compute_mfcc(const std::vector<float>& audio_buffer,
                             std::vector<float>& mfcc_output) override;

native_client/tfmodelstate.cc

@@ -23,10 +23,9 @@ TFModelState::~TFModelState()
 }
 
 int
-TFModelState::init(const char* model_path,
-                   unsigned int beam_width)
+TFModelState::init(const char* model_path)
 {
-  int err = ModelState::init(model_path, beam_width);
+  int err = ModelState::init(model_path);
   if (err != DS_ERR_OK) {
     return err;
   }
@@ -104,6 +103,7 @@ TFModelState::init(const char* model_path,
     "metadata_feature_win_len",
     "metadata_feature_win_step",
     "metadata_alphabet",
+    "metadata_beam_width",
   }, {}, &metadata_outputs);
   if (!status.ok()) {
     std::cout << "Unable to fetch metadata: " << status << std::endl;
@@ -122,9 +122,14 @@ TFModelState::init(const char* model_path,
     return DS_ERR_INVALID_ALPHABET;
   }
 
+  int beam_width = metadata_outputs[4].scalar<int>()();
+  beam_width_ = (unsigned int)(beam_width);
+
   assert(sample_rate_ > 0);
   assert(audio_win_len_ > 0);
   assert(audio_win_step_ > 0);
+  assert(alphabet_.GetSize() > 0);
+  assert(beam_width_ > 0);
 
   for (int i = 0; i < graph_def_.node_size(); ++i) {
     NodeDef node = graph_def_.node(i);

native_client/tfmodelstate.h

@@ -18,8 +18,7 @@ struct TFModelState : public ModelState
   TFModelState();
   virtual ~TFModelState();
 
-  virtual int init(const char* model_path,
-                   unsigned int beam_width) override;
+  virtual int init(const char* model_path) override;
 
   virtual void infer(const std::vector<float>& mfcc,
                      unsigned int n_frames,

util/flags.py

@@ -111,6 +111,7 @@ def create_flags():
     f.DEFINE_string('export_language', '', 'language the model was trained on e.g. "en" or "English". Gets embedded into exported model.')
     f.DEFINE_boolean('export_zip', False, 'export a TFLite model and package with LM and info.json')
     f.DEFINE_string('export_name', 'output_graph', 'name for the export model')
+    f.DEFINE_string('export_beam_width', 500, 'default beam width to embed into exported graph')
 
     # Reporting