IntenseWebs/openvino
3
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

[IE Python Speech Sample] Add context windows feature (#7801)

Browse Source 
* Add `context_window_left` feature

* Add a check of positive context window args

* Add `context_window_right` feature
This commit is contained in:
Dmitry Pigasin
2021-10-28 00:17:34 +03:00
committed by GitHub
parent fe457aa59c
commit 054a2f8d9c
2 changed files with 56 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

39
samples/python/speech_sample/arg_parser.py
View File

@@ -25,8 +25,9 @@ def parse_args() -> argparse.Namespace:
'CPU, GPU, MYRIAD, GNA_AUTO, GNA_HW, GNA_SW_FP32, GNA_SW_EXACT and HETERO with combination of GNA'
' as the primary device and CPU as a secondary (e.g. HETERO:GNA,CPU) are supported. '
'The sample will look for a suitable plugin for device specified. Default value is CPU.')
args.add_argument('-bs', '--batch_size', default=1, type=int, help='Optional. Batch size 1-8 (default 1).')
args.add_argument('-qb', '--quantization_bits', default=16, type=int,
args.add_argument('-bs', '--batch_size', default=1, type=int, choices=range(1, 9), metavar='[1-8]',
help='Optional. Batch size 1-8 (default 1).')
args.add_argument('-qb', '--quantization_bits', default=16, type=int, choices=(8, 16), metavar='[8, 16]',
help='Optional. Weight bits for quantization: 8 or 16 (default 16).')
args.add_argument('-sf', '--scale_factor', type=str,
help='Optional. The user-specified input scale factor for quantization. '
@@ -37,7 +38,7 @@ def parse_args() -> argparse.Namespace:
args.add_argument('-we_gen', '--embedded_gna_configuration', default='GNA1', type=str, help=argparse.SUPPRESS)
args.add_argument('-pc', '--performance_counter', action='store_true',
help='Optional. Enables performance report (specify -a to ensure arch accurate results).')
args.add_argument('-a', '--arch', default='CORE', type=str.upper, choices=['CORE', 'ATOM'],
args.add_argument('-a', '--arch', default='CORE', type=str.upper, choices=('CORE', 'ATOM'), metavar='[CORE, ATOM]',
help='Optional. Specify architecture. CORE, ATOM with the combination of -pc.')
args.add_argument('-iname', '--input_layers', type=str,
help='Optional. Layer names for input blobs. The names are separated with ",". '
@@ -45,5 +46,37 @@ def parse_args() -> argparse.Namespace:
args.add_argument('-oname', '--output_layers', type=str,
help='Optional. Layer names for output blobs. The names are separated with ",". '
'Allows to change the order of output layers for -o flag. Example: Output1:port,Output2:port.')
args.add_argument('-cw_l', '--context_window_left', type=IntRange(0), default=0,
help='Optional. Number of frames for left context windows (default is 0). '
'Works only with context window networks. '
'If you use the cw_l or cw_r flag, then batch size argument is ignored.')
args.add_argument('-cw_r', '--context_window_right', type=IntRange(0), default=0,
help='Optional. Number of frames for right context windows (default is 0). '
'Works only with context window networks. '
'If you use the cw_l or cw_r flag, then batch size argument is ignored.')
return parser.parse_args()
class IntRange:
"""Custom argparse type representing a bounded int."""
def __init__(self, _min=None, _max=None):
self._min = _min
self._max = _max
def __call__(self, arg):
try:
value = int(arg)
except ValueError:
raise argparse.ArgumentTypeError('Must be an integer.')
if (self._min is not None and value < self._min) or (self._max is not None and value > self._max):
if self._min is not None and self._max is not None:
raise argparse.ArgumentTypeError(f'Must be an integer in the range [{self._min}, {self._max}].')
elif self._min is not None:
raise argparse.ArgumentTypeError(f'Must be an integer >= {self._min}.')
elif self._max is not None:
raise argparse.ArgumentTypeError(f'Must be an integer <= {self._max}.')
return value

31
samples/python/speech_sample/speech_sample.py
View File

@@ -30,7 +30,9 @@ def get_scale_factor(matrix: np.ndarray) -> float:
return target_max / max_val
def infer_data(data: dict, exec_net: ExecutableNetwork, input_blobs: list, output_blobs: list) -> np.ndarray:
def infer_data(
data: dict, exec_net: ExecutableNetwork, input_blobs: list, output_blobs: list, cw_l: int = 0, cw_r: int = 0,
) -> np.ndarray:
"""Do a synchronous matrix inference"""
matrix_shape = next(iter(data.values())).shape
result = {}
@@ -40,11 +42,16 @@ def infer_data(data: dict, exec_net: ExecutableNetwork, input_blobs: list, outpu
batch_size = shape[0]
result[blob_name] = np.ndarray((matrix_shape[0], shape[-1]))
slice_begin = 0
slice_end = batch_size
for i in range(-cw_l, matrix_shape[0] + cw_r, batch_size):
if i < 0:
index = 0
elif i >= matrix_shape[0]:
index = matrix_shape[0] - 1
else:
index = i
vectors = {blob_name: data[blob_name][index:index + batch_size] for blob_name in input_blobs}
while slice_begin < matrix_shape[0]:
vectors = {blob_name: data[blob_name][slice_begin:slice_end] for blob_name in input_blobs}
num_of_vectors = next(iter(vectors.values())).shape[0]
if num_of_vectors < batch_size:
@@ -57,11 +64,11 @@ def infer_data(data: dict, exec_net: ExecutableNetwork, input_blobs: list, outpu
vector_results = exec_net.infer(vectors)
for blob_name in output_blobs:
result[blob_name][slice_begin:slice_end] = vector_results[blob_name][:num_of_vectors]
if i - cw_r < 0:
continue
slice_begin += batch_size
slice_end += batch_size
for blob_name in output_blobs:
result[blob_name][i - cw_r:i - cw_r + batch_size] = vector_results[blob_name][:num_of_vectors]
return result
@@ -161,7 +168,7 @@ def main():
for blob_name in output_blobs:
net.outputs[blob_name].precision = 'FP32'
net.batch_size = args.batch_size
net.batch_size = args.batch_size if args.context_window_left + args.context_window_right == 0 else 1
# ---------------------------Step 4. Loading model to the device-------------------------------------------------------
devices = args.device.replace('HETERO:', '').split(',')
@@ -272,7 +279,9 @@ def main():
for state in request.query_state():
state.reset()
result = infer_data(input_data[key], exec_net, input_blobs, output_blobs)
result = infer_data(
input_data[key], exec_net, input_blobs, output_blobs, args.context_window_left, args.context_window_right,
)
for blob_name in result.keys():
results[blob_name][key] = result[blob_name]