| /third_party/mindspore/tests/st/networks/models/ |
| D | resnetv1_5.py | 37 def conv3x3(in_channels, out_channels, stride=1, padding=0): argument
39 weight_shape = (out_channels, in_channels, 3, 3)
41 return nn.Conv2d(in_channels, out_channels,
45 def conv1x1(in_channels, out_channels, stride=1, padding=0): argument
47 weight_shape = (out_channels, in_channels, 1, 1)
49 return nn.Conv2d(in_channels, out_channels,
53 def conv7x7(in_channels, out_channels, stride=1, padding=0): argument
55 weight_shape = (out_channels, in_channels, 7, 7)
57 return nn.Conv2d(in_channels, out_channels,
61 def bn_with_initialize(out_channels): argument
[all …]
|
| /third_party/mindspore/tests/st/mem_reuse/ |
| D | resnet.py | 42 def conv3x3(in_channels, out_channels, stride=1, padding=0): argument
44 weight_shape = (out_channels, in_channels, 3, 3)
46 return nn.Conv2d(in_channels, out_channels,
50 def conv1x1(in_channels, out_channels, stride=1, padding=0): argument
52 weight_shape = (out_channels, in_channels, 1, 1)
54 return nn.Conv2d(in_channels, out_channels,
58 def conv7x7(in_channels, out_channels, stride=1, padding=0): argument
60 weight_shape = (out_channels, in_channels, 7, 7)
62 return nn.Conv2d(in_channels, out_channels,
66 def bn_with_initialize(out_channels): argument
[all …]
|
| /third_party/mindspore/tests/st/tbe_networks/ |
| D | resnet.py | 32 def conv3x3(in_channels, out_channels, stride=1, padding=0): argument
34 return nn.Conv2d(in_channels, out_channels,
39 def conv1x1(in_channels, out_channels, stride=1, padding=0): argument
41 return nn.Conv2d(in_channels, out_channels,
46 def conv7x7(in_channels, out_channels, stride=1, padding=0): argument
48 return nn.Conv2d(in_channels, out_channels,
53 def bn_with_initialize(out_channels): argument
54 shape = (out_channels)
58 bn = nn.BatchNorm2d(out_channels, momentum=0.99, eps=0.00001, gamma_init='Uniform',
63 def bn_with_initialize_last(out_channels): argument
[all …]
|
| /third_party/mindspore/tests/st/export_and_load/ |
| D | test_resnet_air.py | 55 def conv3x3(in_channels, out_channels, stride=1, padding=1): argument
57 weight_shape = (out_channels, in_channels, 3, 3)
59 return Conv2d(in_channels, out_channels,
63 def conv1x1(in_channels, out_channels, stride=1, padding=0): argument
65 weight_shape = (out_channels, in_channels, 1, 1)
67 return Conv2d(in_channels, out_channels,
71 def conv7x7(in_channels, out_channels, stride=1, padding=0): argument
73 weight_shape = (out_channels, in_channels, 7, 7)
75 return Conv2d(in_channels, out_channels,
79 def bn_with_initialize(out_channels): argument
[all …]
|
| /third_party/mindspore/tests/ut/python/exec/ |
| D | resnet_example.py | 26 def conv3x3(in_channels, out_channels, stride=1, padding=1): argument
28 weight = Tensor(np.ones([out_channels, in_channels, 3, 3]).astype(np.float32) * 0.01)
29 return nn.Conv2d(in_channels, out_channels,
33 def conv1x1(in_channels, out_channels, stride=1, padding=0): argument
35 weight = Tensor(np.ones([out_channels, in_channels, 1, 1]).astype(np.float32) * 0.01)
36 return nn.Conv2d(in_channels, out_channels,
40 def bn_with_initialize(out_channels): argument
41 shape = (out_channels)
46 return nn.BatchNorm2d(num_features=out_channels,
61 out_channels, argument
[all …]
|
| /third_party/mindspore/tests/st/networks/ |
| D | test_gpu_resnet.py | 61 def conv3x3(in_channels, out_channels, stride=1, padding=1): argument
63 weight_shape = (out_channels, in_channels, 3, 3)
65 return Conv2d(in_channels, out_channels,
69 def conv1x1(in_channels, out_channels, stride=1, padding=0): argument
71 weight_shape = (out_channels, in_channels, 1, 1)
73 return Conv2d(in_channels, out_channels,
77 def conv7x7(in_channels, out_channels, stride=1, padding=0): argument
79 weight_shape = (out_channels, in_channels, 7, 7)
81 return Conv2d(in_channels, out_channels,
85 def bn_with_initialize(out_channels): argument
[all …]
|
| /third_party/mindspore/tests/st/pynative/ |
| D | test_pynative_resnet50_gpu.py | 55 def conv3x3(in_channels, out_channels, stride=1, padding=0): argument
57 return nn.Conv2d(in_channels, out_channels,
62 def conv1x1(in_channels, out_channels, stride=1, padding=0): argument
64 return nn.Conv2d(in_channels, out_channels,
69 def conv7x7(in_channels, out_channels, stride=1, padding=0): argument
71 return nn.Conv2d(in_channels, out_channels,
76 def bn_with_initialize(out_channels): argument
77 shape = (out_channels)
81 bn = nn.BatchNorm2d(out_channels, momentum=0.99, eps=0.00001, gamma_init='Uniform',
86 def bn_with_initialize_last(out_channels): argument
[all …]
|
| D | test_pynative_resnet50_ascend.py | 93 def conv3x3(in_channels, out_channels, stride=1, padding=0): argument
95 return nn.Conv2d(in_channels, out_channels,
100 def conv1x1(in_channels, out_channels, stride=1, padding=0): argument
102 return nn.Conv2d(in_channels, out_channels,
107 def conv7x7(in_channels, out_channels, stride=1, padding=0): argument
109 return nn.Conv2d(in_channels, out_channels,
114 def bn_with_initialize(out_channels): argument
115 shape = (out_channels)
119 bn = nn.BatchNorm2d(out_channels, momentum=0.99, eps=0.00001, gamma_init='Uniform',
124 def bn_with_initialize_last(out_channels): argument
[all …]
|
| /third_party/mindspore/tests/st/model_zoo_tests/yolov3_darknet53/src/ |
| D | darknet.py | 21 out_channels, argument
31 out_channels,
37 nn.BatchNorm2d(out_channels, momentum=0.1),
59 out_channels): argument
62 out_chls = out_channels//2
64 self.conv2 = conv_block(out_chls, out_channels, kernel_size=3, stride=1)
101 out_channels, argument
105 self.outchannel = out_channels[-1]
108 if not len(layer_nums) == len(in_channels) == len(out_channels) == 5:
115 out_channels[0],
[all …]
|
| /third_party/mindspore/tests/ut/python/model/ |
| D | res18_example.py | 27 def conv3x3(in_channels, out_channels, stride=1, padding=1, pad_mode='pad'): argument
29 return nn.Conv2d(in_channels, out_channels,
33 def conv1x1(in_channels, out_channels, stride=1, padding=0, pad_mode='pad'): argument
35 return nn.Conv2d(in_channels, out_channels,
47 out_channels, argument
52 out_chls = out_channels // self.expansion
59 self.conv3 = conv1x1(out_chls, out_channels, stride=1, padding=0)
60 self.bn3 = nn.BatchNorm2d(out_channels)
65 self.conv_down_sample = conv1x1(in_channels, out_channels,
67 self.bn_down_sample = nn.BatchNorm2d(out_channels)
[all …]
|
| /third_party/mindspore/tests/perf_test/ |
| D | resnet_example.py | 24 def conv3x3(in_channels, out_channels, stride=1, padding=1, pad_mode='pad'): argument
26 return nn.Conv2d(in_channels, out_channels,
30 def conv1x1(in_channels, out_channels, stride=1, padding=0, pad_mode='pad'): argument
32 return nn.Conv2d(in_channels, out_channels,
44 out_channels, argument
49 out_chls = out_channels // self.expansion
56 self.conv3 = conv1x1(out_chls, out_channels, stride=1, padding=0)
57 self.bn3 = nn.BatchNorm2d(out_channels)
62 self.conv_down_sample = conv1x1(in_channels, out_channels,
64 self.bn_down_sample = nn.BatchNorm2d(out_channels)
[all …]
|
| /third_party/mindspore/tests/ut/python/communication/ |
| D | test_data_parallel_resnet.py | 29 def conv3x3(in_channels, out_channels, stride=1, padding=1, pad_mode='pad'): argument
31 return nn.Conv2d(in_channels, out_channels,
35 def conv1x1(in_channels, out_channels, stride=1, padding=0, pad_mode='pad'): argument
37 return nn.Conv2d(in_channels, out_channels,
49 out_channels, argument
54 out_chls = out_channels // self.expansion
61 self.conv3 = conv1x1(out_chls, out_channels, stride=1, padding=0)
62 self.bn3 = nn.BatchNorm2d(out_channels)
67 self.conv_down_sample = conv1x1(in_channels, out_channels,
69 self.bn_down_sample = nn.BatchNorm2d(out_channels)
[all …]
|
| /third_party/mindspore/tests/ut/python/parallel/ |
| D | test_operator_model_parallel.py | 104 def conv3x3(in_channels, out_channels, stride=1): argument
106 weight_shape = (out_channels, in_channels, 3, 3)
108 conv = Conv2d(in_channels, out_channels,
115 def conv1x1(in_channels, out_channels, stride=1): argument
117 weight_shape = (out_channels, in_channels, 1, 1)
119 conv = Conv2d(in_channels, out_channels,
126 def conv7x7(in_channels, out_channels, stride=1): argument
128 weight_shape = (out_channels, in_channels, 7, 7)
130 conv = Conv2d(in_channels, out_channels,
147 def bn_with_initialize(out_channels): argument
[all …]
|
| D | test_auto_parallel_resnet_sharding_propagation2.py | 46 def _conv3x3(in_channels, out_channels, stride=1, padding=0, pad_mode='same'): argument
49 return nn.Conv2d(in_channels, out_channels,
53 def _conv1x1(in_channels, out_channels, stride=1, padding=0, pad_mode='same'): argument
56 return nn.Conv2d(in_channels, out_channels,
60 def _conv7x7(in_channels, out_channels, stride=1, padding=0, pad_mode='same'): argument
63 return nn.Conv2d(in_channels, out_channels,
77 out_channels, argument
82 out_chls = out_channels // self.expansion
89 self.conv3 = _conv1x1(out_chls, out_channels, stride=1)
90 self.bn3 = _fused_bn(out_channels, momentum=momentum)
[all …]
|
| D | test_auto_parallel_resnet_sharding_propagation.py | 47 def _conv3x3(in_channels, out_channels, stride=1, padding=0, pad_mode='same'): argument
50 return nn.Conv2d(in_channels, out_channels,
54 def _conv1x1(in_channels, out_channels, stride=1, padding=0, pad_mode='same'): argument
57 return nn.Conv2d(in_channels, out_channels,
61 def _conv7x7(in_channels, out_channels, stride=1, padding=0, pad_mode='same'): argument
64 return nn.Conv2d(in_channels, out_channels,
78 out_channels, argument
83 out_chls = out_channels // self.expansion
90 self.conv3 = _conv1x1(out_chls, out_channels, stride=1)
91 self.bn3 = _fused_bn(out_channels, momentum=momentum)
[all …]
|
| /third_party/mindspore/mindspore/nn/layer/ |
| D | conv.py | 38 out_channels, argument
53 … self.out_channels = Validator.check_positive_int(out_channels, 'out_channels', self.cls_name)
91 if out_channels % group != 0:
95 shape = [in_channels, out_channels // group, *kernel_size]
97 … shape = [out_channels, *kernel_size, in_channels // group] if self.format == "NHWC" else \
98 [out_channels, in_channels // group, *kernel_size]
102 self.bias = Parameter(initializer(self.bias_init, [out_channels]), name='bias')
225 out_channels, argument
243 out_channels,
254 self.conv2d = P.Conv2D(out_channel=self.out_channels,
[all …]
|
| /third_party/mindspore/tests/st/auto_parallel/ |
| D | resnet50_expand_loss.py | 45 def _conv3x3(in_channels, out_channels, stride=1, padding=0, pad_mode='same'): argument
47 return nn.Conv2d(in_channels, out_channels,
51 def _conv1x1(in_channels, out_channels, stride=1, padding=0, pad_mode='same'): argument
53 return nn.Conv2d(in_channels, out_channels,
57 def _conv7x7(in_channels, out_channels, stride=1, padding=0, pad_mode='same'): argument
59 return nn.Conv2d(in_channels, out_channels,
72 out_channels, argument
77 self.conv1 = _conv3x3(in_channels, out_channels, stride=stride)
78 self.bn1 = _fused_bn(out_channels, momentum=momentum)
79 self.conv2 = _conv3x3(out_channels, out_channels)
[all …]
|
| /third_party/gstreamer/gstplugins_base/gst-libs/gst/audio/ |
| D | audio-channel-mixer.c | 63 gint out_channels; member
109 gint in_channels, GstAudioChannelPosition * in_position, gint out_channels, in gst_audio_channel_mixer_fill_identical() argument
117 for (co = 0; co < out_channels; co++) { in gst_audio_channel_mixer_fill_identical()
139 GstAudioChannelPosition * in_position, gint out_channels, in gst_audio_channel_mixer_fill_compatible() argument
181 for (n = 0; n < out_channels; n++) { in gst_audio_channel_mixer_fill_compatible()
369 GstAudioChannelPosition * in_position, gint out_channels, in gst_audio_channel_mixer_fill_others() argument
399 gst_audio_channel_mixer_detect_pos (out_channels, out_position, in gst_audio_channel_mixer_fill_others()
570 gint out_channels) in gst_audio_channel_mixer_fill_normalize() argument
575 for (j = 0; j < out_channels; j++) { in gst_audio_channel_mixer_fill_normalize()
590 for (j = 0; j < out_channels; j++) { in gst_audio_channel_mixer_fill_normalize()
[all …]
|
| /third_party/mindspore/mindspore/lite/examples/export_models/models/ |
| D | NetworkInNetwork.py | 31 … nn.Conv2d(in_channels=num_channel, out_channels=192, kernel_size=5, stride=1, has_bias=False),
33 nn.Conv2d(in_channels=192, out_channels=160, kernel_size=1, stride=1, has_bias=True),
35 nn.Conv2d(in_channels=160, out_channels=96, kernel_size=1, stride=1, has_bias=True),
42 nn.Conv2d(in_channels=96, out_channels=192, kernel_size=5, stride=1, has_bias=False),
44 nn.Conv2d(in_channels=192, out_channels=192, kernel_size=1, stride=1, has_bias=True),
46 nn.Conv2d(in_channels=192, out_channels=192, kernel_size=1, stride=1, has_bias=True),
53 nn.Conv2d(in_channels=192, out_channels=192, kernel_size=3, stride=1, has_bias=False),
55 nn.Conv2d(in_channels=192, out_channels=192, kernel_size=1, stride=1, has_bias=True),
57 … nn.Conv2d(in_channels=192, out_channels=num_classes, kernel_size=1, stride=1, has_bias=True),
69 n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels
|
| /third_party/ffmpeg/libavresample/ |
| D | audio_mix.c | 40 int out_channels; member
63 int out_channels, int ptr_align, int samples_align, in ff_audio_mix_set_func() argument
68 (out_channels == am->out_matrix_channels || out_channels == 0)) { in ff_audio_mix_set_func()
81 if (out_channels) in ff_audio_mix_set_func()
83 in_channels, out_channels); in ff_audio_mix_set_func()
87 } else if (out_channels) { in ff_audio_mix_set_func()
89 out_channels); in ff_audio_mix_set_func()
335 am->out_channels); in mix_function_init()
364 am->out_channels = avr->out_channels; in ff_audio_mix_alloc()
373 double *matrix_dbl = av_mallocz(avr->out_channels * avr->in_channels * in ff_audio_mix_alloc()
[all …]
|
| /third_party/mindspore/tests/st/networks/models/deeplabv3/src/backbone/ |
| D | resnet_deeplab.py | 124 out_channels, argument
154 out_channels=out_channels,
158 in_channels = out_channels
300 out_channels, argument
306 mid_channels = out_channels // expansion
334 out_channels,
337 self.bn3 = nn.BatchNorm2d(out_channels, use_batch_statistics=use_batch_statistics)
338 if in_channels != out_channels:
340 out_channels,
343 bn = nn.BatchNorm2d(out_channels, use_batch_statistics=use_batch_statistics)
[all …]
|
| /third_party/mindspore/tests/st/fl/cross_silo_faster_rcnn/src/FasterRcnn/ |
| D | resnet.py | 29 def _conv(in_channels, out_channels, kernel_size=3, stride=1, padding=0, pad_mode='pad'): argument
31 shape = (out_channels, in_channels, kernel_size, kernel_size)
33 return nn.Conv2d(in_channels, out_channels,
74 out_channels, argument
78 if not len(layer_nums) == len(in_channels) == len(out_channels) == 4:
95 out_channel=out_channels[0],
102 out_channel=out_channels[1],
109 out_channel=out_channels[2],
116 out_channel=out_channels[3],
190 out_channels, argument
[all …]
|
| D | resnet50v1.py | 29 def _conv(in_channels, out_channels, kernel_size=3, stride=1, padding=0, pad_mode='pad'): argument
31 shape = (out_channels, in_channels, kernel_size, kernel_size)
33 return nn.Conv2d(in_channels, out_channels,
74 out_channels, argument
78 if not len(layer_nums) == len(in_channels) == len(out_channels) == 4:
95 out_channel=out_channels[0],
102 out_channel=out_channels[1],
109 out_channel=out_channels[2],
116 out_channel=out_channels[3],
190 out_channels, argument
[all …]
|
| /third_party/mindspore/tests/st/gnn/ |
| D | aggregator.py | 71 out_channels, argument
77 self.out_channels = Validator.check_positive_int(out_channels)
81 if weight_init.ndim != 2 or weight_init.shape[0] != out_channels or \
85 … self.weight = Parameter(initializer(weight_init, [out_channels, in_channels]), name="weight")
89 if bias_init.ndim != 1 or bias_init.shape[0] != out_channels:
92 self.bias = Parameter(initializer(bias_init, [out_channels]), name="bias")
103 output = F.reshape(output, (tensor_shape[0], tensor_shape[1], self.out_channels))
107 s = 'in_channels={}, out_channels={}'.format(self.in_channels, self.out_channels)
270 out_channels=self.out_channel,
275 out_channels=1)
[all …]
|
| /third_party/mindspore/tests/ut/python/transform/ |
| D | test_transform.py | 32 def conv3x3(in_channels, out_channels, stride=1, padding=1): argument
34 weight = Tensor(np.ones([out_channels, in_channels, 3, 3]).astype(np.float32))
35 return nn.Conv2d(in_channels, out_channels,
40 def conv1x1(in_channels, out_channels, stride=1, padding=0): argument
42 weight = Tensor(np.ones([out_channels, in_channels, 1, 1]).astype(np.float32))
43 return nn.Conv2d(in_channels, out_channels,
56 out_channels, argument
61 out_chls = out_channels // self.expansion
68 self.conv3 = conv1x1(out_chls, out_channels, stride=1, padding=0)
69 self.bn3 = nn.BatchNorm2d(out_channels)
[all …]
|