farmaker47

    fun performInference(context: Context) {
    
        val startTime = System.currentTimeMillis()
        val outputs = Array(1) {
            FloatArray(11)
        }
        // Get the audio file.
        val fileInputstream: InputStream = context.assets.open("down.wav")
        try {
            val byteIOArray = IOUtils.toByteArray(fileInputstream)

farmaker47

    // Init the TnsorFlow Lite Interpreter
    fun initInterpreter(context: Context) {
        interpreterAudioClassification = getInterpreter(context, "speech_commands_model.tflite")
    }

    // Get the Interpreter.
    @Throws(IOException::class)
    private fun getInterpreter(
        context: Context,
        modelName: String

farmaker47

# open the .wav file
with open('down.wav', 'rb') as wav_file:
    wav_data = wav_file.read()

# load TFLite model and set params
mod_path = '/content/speech_commands_model.tflite'
interpreter = tf.lite.Interpreter(model_path=mod_path)
interpreter.allocate_tensors()
input_index = interpreter.get_input_details()[0]["index"]
output_index = interpreter.get_output_details()[0]["index"]

farmaker47

# Convert the model
converter = tf.lite.TFLiteConverter.from_saved_model("/content/saved_model") # path to the SavedModel directory
converter.target_spec.supported_ops = [
  tf.lite.OpsSet.TFLITE_BUILTINS, # enable TensorFlow Lite ops.
  tf.lite.OpsSet.SELECT_TF_OPS # enable TensorFlow ops.
]
tflite_model = converter.convert()

# Save the model.
with open('speech_commands_model.tflite', 'wb') as f:

farmaker47

### Convert TF1 frozen Graph to TF1 SavedModel.

# Load the graph as a v1.GraphDef
import pathlib
gdef = tf.compat.v1.GraphDef()
gdef.ParseFromString(pathlib.Path(GRAPH_DEF_MODEL_PATH).read_bytes())

# Convert the GraphDef to a tf.Graph
with tf.Graph().as_default() as g:
  tf.graph_util.import_graph_def(gdef, name="")

farmaker47

# print all layers' names
def printTensors(pb_file):

    # read pb into graph_def
    with tf.gfile.GFile(pb_file, "rb") as f:
        graph_def = tf.GraphDef()
        graph_def.ParseFromString(f.read())

    # import graph_def
    with tf.Graph().as_default() as graph:

farmaker47

GRAPH_DEF_MODEL_PATH = '/content/speech_commands_files/conv_actions_frozen.pb'
print("TF1 frozen GraphDef path: ", GRAPH_DEF_MODEL_PATH)

# view input and output layers' names
import tensorflow.compat.v1 as tf
tf.disable_v2_behavior()
gf = tf.GraphDef()   
m_file = open(GRAPH_DEF_MODEL_PATH,'rb')
gf.ParseFromString(m_file.read())

farmaker47

# Here we load again the model as we are going to build it under strategy.scope()
def build_model(max_len=128):
    input_word_ids = tf.keras.Input(shape=(max_len,), dtype=tf.int32, name="input_ids")
    input_mask = tf.keras.Input(shape=(max_len,), dtype=tf.int32, name="attention_mask")
    segment_ids = tf.keras.Input(shape=(max_len,), dtype=tf.int32, name="token_type_ids")

    bert_results = hub.KerasLayer(TFAutoModel.from_pretrained("nlpaueb/bert-base-greek-uncased-v1"), trainable=True, name='BERT_encoder')([input_word_ids, input_mask, segment_ids])
    #output = bert_results["last_hidden_state"][:, 0, :]
    output = bert_results["pooler_output"][:,:]

farmaker47

# Create 3 inputs for the model
def bert_encode(texts, tokenizer, max_len=128):
    all_tokens = []
    all_masks = []
    all_segments = []
    
    for text in texts:
        text_preprocessed = tokenizer(text)
        tokenized_text = text_preprocessed["input_ids"]
        #print(tokenized_text)

farmaker47

MIXED_PRECISION = False
if MIXED_PRECISION:
    if tpu: 
        policy = tf.keras.mixed_precision.experimental.Policy('mixed_bfloat16')
    else: #
        policy = tf.keras.mixed_precision.experimental.Policy('mixed_float16')
        tf.config.optimizer.set_jit(True) # XLA compilation
    tf.keras.mixed_precision.experimental.set_policy(policy)
    print('Mixed precision enabled')