updated colab version

mase8129 · Jun 29, 2023 · b94f56c · b94f56c
1 parent fff168e
commit b94f56c
Show file tree

Hide file tree

Showing 3 changed files with 319 additions and 324 deletions.
diff --git a/.DS_Store b/.DS_Store
diff --git a/4_Model_training.ipynb b/4_Model_training.ipynb
@@ -514,10 +514,6 @@
     "    speech_for_predicition.append(sample[0])\n",
     "    break\n",
     "# # normalize speechfile\n",
-    "print(speech_for_predicition[0].shape)\n",
-    "print(speech_for_predicition[0].dtype)\n",
-    "\n",
-    "\n",
     "speech_for_predicition = speech_for_predicition[0]\n",
     "speech_for_predicition = librosa.util.normalize(speech_for_predicition)\n",
     "\n",
@@ -585,21 +581,12 @@
     "\n",
     "        # for each audio file\n",
     "        for idx, fp in enumerate(fps):\n",
-    "            #  # if audio is not 16-bit PCM WAV\n",
-    "            #  # convert to 16-bit PCM WAV\n",
-    "            #  if librosa.load(fp)[0].dtype != np.int16:\n",
-    "            #      audio = librosa.load(fp)[0] * 32767\n",
-    "            #      audio = audio.astype(np.int16)\n",
-    "            #      # overwrite existing audio file\n",
-    "            #      write(fp, 44100, audio.astype(np.int16))\n",
     "            \n",
     "            # write audiosummary of one audio file to disk\n",
     "             writer = tf.summary.create_file_writer(path_audiosummary)\n",
     "             with writer.as_default():\n",
-    "       \n",
-    "                # write audio files to one variable (k, audio, channels) = (len(fps), 3*44100, 1) to tf.summary.audio\n",
-    "                # save all audio files to one variable\n",
-    "   \n",
+    "                \n",
+    "                # load audio file as tensor\n",
     "                file = tf.io.read_file(fp)\n",
     "                audio = tf.audio.decode_wav(file, desired_channels=1)\n",
     "\n",