diff --git a/.notebook/python_test.ipynb b/.notebook/python_test.ipynb index 5617c8765..2d6a3b31f 100644 --- a/.notebook/python_test.ipynb +++ b/.notebook/python_test.ipynb @@ -637,7 +637,7 @@ { "cell_type": "code", "execution_count": 59, - "id": "legitimate-overhead", + "id": "diverse-advocate", "metadata": {}, "outputs": [ { @@ -660,7 +660,7 @@ { "cell_type": "code", "execution_count": 35, - "id": "genuine-feeding", + "id": "liquid-argument", "metadata": {}, "outputs": [ { @@ -704,8 +704,8 @@ }, { "cell_type": "code", - "execution_count": 36, - "id": "bizarre-story", + "execution_count": 3, + "id": "copyrighted-advertising", "metadata": {}, "outputs": [ { @@ -727,8 +727,8 @@ }, { "cell_type": "code", - "execution_count": 37, - "id": "appointed-brooklyn", + "execution_count": 4, + "id": "compatible-edwards", "metadata": {}, "outputs": [ { @@ -747,8 +747,8 @@ }, { "cell_type": "code", - "execution_count": 38, - "id": "occasional-utilization", + "execution_count": 5, + "id": "general-counter", "metadata": {}, "outputs": [], "source": [ @@ -757,8 +757,8 @@ }, { "cell_type": "code", - "execution_count": 40, - "id": "trained-indonesian", + "execution_count": 6, + "id": "logical-kelly", "metadata": {}, "outputs": [ { @@ -775,8 +775,8 @@ }, { "cell_type": "code", - "execution_count": 54, - "id": "following-brave", + "execution_count": 7, + "id": "metallic-semiconductor", "metadata": {}, "outputs": [ { @@ -808,8 +808,8 @@ }, { "cell_type": "code", - "execution_count": 42, - "id": "prospective-blind", + "execution_count": 17, + "id": "shaped-oregon", "metadata": {}, "outputs": [ { @@ -826,7 +826,7 @@ "" ] }, - "execution_count": 42, + "execution_count": 17, "metadata": {}, "output_type": "execute_result" } @@ -838,8 +838,40 @@ }, { "cell_type": "code", - "execution_count": 43, - "id": "minus-ethernet", + "execution_count": 18, + "id": "loving-chester", + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "\n", + " \n", + " " + ], + "text/plain": [ + "" + ] + }, + "execution_count": 18, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "tfm = sox.Transformer()\n", + "tfm.speed(1.0)\n", + "output_array = tfm.build_array(input_array=samples, sample_rate_in=sr)\n", + "ipd.Audio(output_array, rate=sr) # load a NumPy array" + ] + }, + { + "cell_type": "code", + "execution_count": 19, + "id": "annual-kidney", "metadata": {}, "outputs": [ { @@ -856,19 +888,22 @@ "" ] }, - "execution_count": 43, + "execution_count": 19, "metadata": {}, "output_type": "execute_result" } ], "source": [ + "tfm = sox.Transformer()\n", + "tfm.speed(1.1)\n", + "output_array = tfm.build_array(input_array=samples, sample_rate_in=sr)\n", "ipd.Audio(output_array, rate=sr) # load a NumPy array" ] }, { "cell_type": "code", - "execution_count": 44, - "id": "ordinary-closer", + "execution_count": 31, + "id": "dressed-linux", "metadata": {}, "outputs": [ { @@ -885,7 +920,7 @@ "" ] }, - "execution_count": 44, + "execution_count": 31, "metadata": {}, "output_type": "execute_result" } @@ -899,8 +934,39 @@ }, { "cell_type": "code", - "execution_count": 45, - "id": "demographic-mumbai", + "execution_count": 32, + "id": "shared-suspension", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "66078\n" + ] + }, + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": { + "needs_background": "light" + }, + "output_type": "display_data" + } + ], + "source": [ + "librosa.display.waveplot(samples_out, sr=sr)\n", + "print(len(samples_out))" + ] + }, + { + "cell_type": "code", + "execution_count": 20, + "id": "invisible-oakland", "metadata": {}, "outputs": [], "source": [ @@ -911,7 +977,7 @@ { "cell_type": "code", "execution_count": 46, - "id": "conscious-stuff", + "id": "czech-flesh", "metadata": {}, "outputs": [ { @@ -931,7 +997,7 @@ { "cell_type": "code", "execution_count": 30, - "id": "virgin-dublin", + "id": "yellow-great", "metadata": {}, "outputs": [], "source": [ @@ -941,7 +1007,7 @@ { "cell_type": "code", "execution_count": 31, - "id": "sized-homework", + "id": "quiet-booking", "metadata": {}, "outputs": [], "source": [ @@ -951,7 +1017,7 @@ { "cell_type": "code", "execution_count": 47, - "id": "disciplinary-headquarters", + "id": "certified-clinton", "metadata": {}, "outputs": [ { @@ -971,8 +1037,8 @@ }, { "cell_type": "code", - "execution_count": 57, - "id": "persistent-synthetic", + "execution_count": 1, + "id": "fatal-latvia", "metadata": {}, "outputs": [], "source": [ @@ -983,7 +1049,7 @@ { "cell_type": "code", "execution_count": 56, - "id": "hydraulic-reach", + "id": "boxed-damage", "metadata": {}, "outputs": [], "source": [ @@ -993,7 +1059,7 @@ { "cell_type": "code", "execution_count": 58, - "id": "suitable-house", + "id": "empty-latitude", "metadata": {}, "outputs": [], "source": [ @@ -1003,7 +1069,7 @@ { "cell_type": "code", "execution_count": 60, - "id": "printable-carter", + "id": "apart-restriction", "metadata": {}, "outputs": [], "source": [ @@ -1013,7 +1079,7 @@ { "cell_type": "code", "execution_count": 66, - "id": "considered-interval", + "id": "corresponding-relief", "metadata": {}, "outputs": [], "source": [ @@ -1023,7 +1089,7 @@ { "cell_type": "code", "execution_count": 95, - "id": "ideal-hurricane", + "id": "vertical-monroe", "metadata": {}, "outputs": [ { @@ -1044,7 +1110,7 @@ { "cell_type": "code", "execution_count": 81, - "id": "skilled-cooler", + "id": "handmade-athletics", "metadata": {}, "outputs": [ { @@ -1062,10 +1128,255 @@ "random.sample(['5','4', '3'], 2)" ] }, + { + "cell_type": "code", + "execution_count": 11, + "id": "failing-barrier", + "metadata": {}, + "outputs": [], + "source": [ + "def change_speed(samples, speed_rate):\n", + " \"\"\"Change the audio speed by linear interpolation.\n", + "\n", + " Note that this is an in-place transformation.\n", + "\n", + " :param speed_rate: Rate of speed change:\n", + " speed_rate > 1.0, speed up the audio;\n", + " speed_rate = 1.0, unchanged;\n", + " speed_rate < 1.0, slow down the audio;\n", + " speed_rate <= 0.0, not allowed, raise ValueError.\n", + " :type speed_rate: float\n", + " :raises ValueError: If speed_rate <= 0.0.\n", + " \"\"\"\n", + " if speed_rate <= 0:\n", + " raise ValueError(\"speed_rate should be greater than zero.\")\n", + " old_length = samples.shape[0]\n", + " new_length = int(old_length / speed_rate)\n", + " old_indices = np.arange(old_length)\n", + " new_indices = np.linspace(start=0, stop=old_length, num=new_length)\n", + " samples = np.interp(new_indices, old_indices, samples)\n", + " return samples" + ] + }, + { + "cell_type": "code", + "execution_count": 12, + "id": "british-perfume", + "metadata": {}, + "outputs": [], + "source": [ + "samples, sr = sf.read(wav)\n", + "samples_out = change_speed(samples, 1.0)" + ] + }, + { + "cell_type": "code", + "execution_count": 14, + "id": "balanced-lawsuit", + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "\n", + " \n", + " " + ], + "text/plain": [ + "" + ] + }, + "execution_count": 14, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "ipd.Audio(samples, rate=sr) # load a NumPy array" + ] + }, + { + "cell_type": "code", + "execution_count": 13, + "id": "married-wallpaper", + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "\n", + " \n", + " " + ], + "text/plain": [ + "" + ] + }, + "execution_count": 13, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "ipd.Audio(samples_out, rate=sr) # load a NumPy array" + ] + }, + { + "cell_type": "code", + "execution_count": 15, + "id": "parallel-electric", + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "\n", + " \n", + " " + ], + "text/plain": [ + "" + ] + }, + "execution_count": 15, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "samples_out = change_speed(samples, 1.1)\n", + "ipd.Audio(samples_out, rate=sr) # load a NumPy array" + ] + }, + { + "cell_type": "code", + "execution_count": 29, + "id": "foster-senegal", + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "\n", + " \n", + " " + ], + "text/plain": [ + "" + ] + }, + "execution_count": 29, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "samples_out = change_speed(samples, 0.9)\n", + "ipd.Audio(samples_out, rate=sr) # load a NumPy array" + ] + }, + { + "cell_type": "code", + "execution_count": 30, + "id": "great-shareware", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "66078\n" + ] + } + ], + "source": [ + "print(len(samples_out))" + ] + }, + { + "cell_type": "code", + "execution_count": 22, + "id": "collaborative-behavior", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Collecting matplotlib\n", + " Downloading matplotlib-3.4.1-cp37-cp37m-manylinux1_x86_64.whl (10.3 MB)\n", + "\u001b[K |████████████████████████████████| 10.3 MB 691 kB/s eta 0:00:01\n", + "\u001b[?25hRequirement already satisfied: pillow>=6.2.0 in /workspace/DeepSpeech-2.x/tools/venv/lib/python3.7/site-packages (from matplotlib) (8.1.0)\n", + "Requirement already satisfied: numpy>=1.16 in /workspace/DeepSpeech-2.x/tools/venv/lib/python3.7/site-packages (from matplotlib) (1.20.1)\n", + "Requirement already satisfied: python-dateutil>=2.7 in /workspace/DeepSpeech-2.x/tools/venv/lib/python3.7/site-packages (from matplotlib) (2.8.1)\n", + "Collecting kiwisolver>=1.0.1\n", + " Downloading kiwisolver-1.3.1-cp37-cp37m-manylinux1_x86_64.whl (1.1 MB)\n", + "\u001b[K |████████████████████████████████| 1.1 MB 45.9 MB/s eta 0:00:01\n", + "\u001b[?25hRequirement already satisfied: pyparsing>=2.2.1 in /workspace/DeepSpeech-2.x/tools/venv/lib/python3.7/site-packages (from matplotlib) (2.4.7)\n", + "Collecting cycler>=0.10\n", + " Downloading cycler-0.10.0-py2.py3-none-any.whl (6.5 kB)\n", + "Requirement already satisfied: six in /workspace/DeepSpeech-2.x/tools/venv/lib/python3.7/site-packages (from cycler>=0.10->matplotlib) (1.15.0)\n", + "Installing collected packages: kiwisolver, cycler, matplotlib\n", + "Successfully installed cycler-0.10.0 kiwisolver-1.3.1 matplotlib-3.4.1\n" + ] + } + ], + "source": [ + "!pip install matplotlib\n", + "%matplotlib inline\n", + "import matplotlib.pyplot as plt\n", + "import librosa.display" + ] + }, + { + "cell_type": "code", + "execution_count": 23, + "id": "moving-trunk", + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "" + ] + }, + "execution_count": 23, + "metadata": {}, + "output_type": "execute_result" + }, + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": { + "needs_background": "light" + }, + "output_type": "display_data" + } + ], + "source": [ + "librosa.display.waveplot(samples_out, sr=sr)" + ] + }, { "cell_type": "code", "execution_count": null, - "id": "adopted-hardware", + "id": "previous-senegal", "metadata": {}, "outputs": [], "source": [] diff --git a/deepspeech/frontend/audio.py b/deepspeech/frontend/audio.py index ebc89bec3..18c75a177 100644 --- a/deepspeech/frontend/audio.py +++ b/deepspeech/frontend/audio.py @@ -22,7 +22,6 @@ import resampy from scipy import signal import random import copy -import sox class AudioSegment(object): @@ -322,18 +321,24 @@ class AudioSegment(object): :type speed_rate: float :raises ValueError: If speed_rate <= 0.0. """ + if speed_rate == 1.0: + return if speed_rate <= 0: raise ValueError("speed_rate should be greater than zero.") - # old_length = self._samples.shape[0] - # new_length = int(old_length / speed_rate) - # old_indices = np.arange(old_length) - # new_indices = np.linspace(start=0, stop=old_length, num=new_length) - # self._samples = np.interp(new_indices, old_indices, self._samples) - tfm = sox.Transformer() - tfm.set_globals(multithread=False) - tfm.speed(speed_rate) - self._samples = tfm.build_array( - input_array=self._samples, sample_rate_in=self._sample_rate).copy() + + # numpy + old_length = self._samples.shape[0] + new_length = int(old_length / speed_rate) + old_indices = np.arange(old_length) + new_indices = np.linspace(start=0, stop=old_length, num=new_length) + self._samples = np.interp(new_indices, old_indices, self._samples) + + # sox, slow + # tfm = sox.Transformer() + # tfm.set_globals(multithread=False) + # tfm.speed(speed_rate) + # self._samples = tfm.build_array( + # input_array=self._samples, sample_rate_in=self._sample_rate).copy() def normalize(self, target_db=-20, max_gain_db=300.0): """Normalize audio to be of the desired RMS value in decibels. diff --git a/deepspeech/io/dataset.py b/deepspeech/io/dataset.py index 7812cbdc2..a660ae494 100644 --- a/deepspeech/io/dataset.py +++ b/deepspeech/io/dataset.py @@ -12,6 +12,7 @@ # See the License for the specific language governing permissions and # limitations under the License. +import time import io import random import tarfile diff --git a/deepspeech/training/gradclip.py b/deepspeech/training/gradclip.py index 37fbb0f56..b52bb981d 100644 --- a/deepspeech/training/gradclip.py +++ b/deepspeech/training/gradclip.py @@ -43,11 +43,11 @@ class ClipGradByGlobalNormWithLog(paddle.nn.ClipGradByGlobalNorm): merge_grad = layers.get_tensor_from_selected_rows(merge_grad) square = layers.square(merge_grad) sum_square = layers.reduce_sum(square) - logger.debug( - f"Grad Before Clip: {p.name}: {float(layers.sqrt(layers.reduce_sum(layers.square(merge_grad))) ) }" - ) sum_square_list.append(sum_square) + # debug log + # logger.debug(f"Grad Before Clip: {p.name}: {float(sum_square.sqrt()) }") + # all parameters have been filterd out if len(sum_square_list) == 0: return params_grads @@ -55,6 +55,7 @@ class ClipGradByGlobalNormWithLog(paddle.nn.ClipGradByGlobalNorm): global_norm_var = layers.concat(sum_square_list) global_norm_var = layers.reduce_sum(global_norm_var) global_norm_var = layers.sqrt(global_norm_var) + # debug log logger.debug(f"Grad Global Norm: {float(global_norm_var)}!!!!") max_global_norm = layers.fill_constant( shape=[1], dtype=global_norm_var.dtype, value=self.clip_norm) @@ -68,9 +69,11 @@ class ClipGradByGlobalNormWithLog(paddle.nn.ClipGradByGlobalNorm): params_and_grads.append((p, g)) continue new_grad = layers.elementwise_mul(x=g, y=clip_var) - logger.debug( - f"Grad After Clip: {p.name}: {float(layers.sqrt(layers.reduce_sum(layers.square(merge_grad))) ) }" - ) params_and_grads.append((p, new_grad)) + # debug log + # logger.debug( + # f"Grad After Clip: {p.name}: {float(merge_grad.square().sum().sqrt())}" + # ) + return params_and_grads