當前位置: 首頁>>代碼示例>>Python>>正文

Python make_layers.conv_with_kaiming_uniform方法代碼示例

本文整理匯總了Python中maskrcnn_benchmark.modeling.make_layers.conv_with_kaiming_uniform方法的典型用法代碼示例。如果您正苦於以下問題：Python make_layers.conv_with_kaiming_uniform方法的具體用法？Python make_layers.conv_with_kaiming_uniform怎麽用？Python make_layers.conv_with_kaiming_uniform使用的例子？那麽, 這裏精選的方法代碼示例或許可以為您提供幫助。您也可以進一步了解該方法所在類maskrcnn_benchmark.modeling.make_layers的用法示例。


在下文中一共展示了make_layers.conv_with_kaiming_uniform方法的13個代碼示例，這些例子默認根據受歡迎程度排序。您可以為喜歡或者感覺有用的代碼點讚，您的評價將有助於係統推薦出更棒的Python代碼示例。

示例1: build_res2net_fpn_backbone

# 需要導入模塊: from maskrcnn_benchmark.modeling import make_layers [as 別名]
# 或者: from maskrcnn_benchmark.modeling.make_layers import conv_with_kaiming_uniform [as 別名]
def build_res2net_fpn_backbone(cfg):
    body = res2net.Res2Net(cfg)
    in_channels_stage2 = cfg.MODEL.RESNETS.RES2_OUT_CHANNELS
    out_channels = cfg.MODEL.RESNETS.BACKBONE_OUT_CHANNELS
    fpn = fpn_module.FPN(
        in_channels_list=[
            in_channels_stage2,
            in_channels_stage2 * 2,
            in_channels_stage2 * 4,
            in_channels_stage2 * 8,
        ],
        out_channels=out_channels,
        conv_block=conv_with_kaiming_uniform(
            cfg.MODEL.FPN.USE_GN, cfg.MODEL.FPN.USE_RELU
        ),
        top_blocks=fpn_module.LastLevelMaxPool(),
    )
    model = nn.Sequential(OrderedDict([("body", body), ("fpn", fpn)]))
    model.out_channels = out_channels
    return model
開發者ID:Res2Net，項目名稱:Res2Net-maskrcnn，代碼行數:22，代碼來源:res2net_builder.py

示例2: build_resnet_fpn_backbone

# 需要導入模塊: from maskrcnn_benchmark.modeling import make_layers [as 別名]
# 或者: from maskrcnn_benchmark.modeling.make_layers import conv_with_kaiming_uniform [as 別名]
def build_resnet_fpn_backbone(cfg):
    body = resnet.ResNet(cfg)
    in_channels_stage2 = cfg.MODEL.RESNETS.RES2_OUT_CHANNELS
    out_channels = cfg.MODEL.RESNETS.BACKBONE_OUT_CHANNELS
    fpn = fpn_module.FPN(
        in_channels_list=[
            in_channels_stage2,
            in_channels_stage2 * 2,
            in_channels_stage2 * 4,
            in_channels_stage2 * 8,
        ],
        out_channels=out_channels,
        conv_block=conv_with_kaiming_uniform(
            cfg.MODEL.FPN.USE_GN, cfg.MODEL.FPN.USE_RELU
        ),
        top_blocks=fpn_module.LastLevelMaxPool(),
    )
    model = nn.Sequential(OrderedDict([("body", body), ("fpn", fpn)]))
    model.out_channels = out_channels
    return model
開發者ID:Res2Net，項目名稱:Res2Net-maskrcnn，代碼行數:22，代碼來源:backbone.py

示例3: build_resnet_fpn_p3p7_backbone

# 需要導入模塊: from maskrcnn_benchmark.modeling import make_layers [as 別名]
# 或者: from maskrcnn_benchmark.modeling.make_layers import conv_with_kaiming_uniform [as 別名]
def build_resnet_fpn_p3p7_backbone(cfg):
    body = resnet.ResNet(cfg)
    in_channels_stage2 = cfg.MODEL.RESNETS.RES2_OUT_CHANNELS
    out_channels = cfg.MODEL.RESNETS.BACKBONE_OUT_CHANNELS
    in_channels_p6p7 = in_channels_stage2 * 8 if cfg.MODEL.RETINANET.USE_C5 \
        else out_channels
    fpn = fpn_module.FPN(
        in_channels_list=[
            0,
            in_channels_stage2 * 2,
            in_channels_stage2 * 4,
            in_channels_stage2 * 8,
        ],
        out_channels=out_channels,
        conv_block=conv_with_kaiming_uniform(
            cfg.MODEL.FPN.USE_GN, cfg.MODEL.FPN.USE_RELU
        ),
        top_blocks=fpn_module.LastLevelP6P7(in_channels_p6p7, out_channels),
    )
    model = nn.Sequential(OrderedDict([("body", body), ("fpn", fpn)]))
    model.out_channels = out_channels
    return model
開發者ID:Res2Net，項目名稱:Res2Net-maskrcnn，代碼行數:24，代碼來源:backbone.py

示例4: build_detnasnet_fpn_backbone

# 需要導入模塊: from maskrcnn_benchmark.modeling import make_layers [as 別名]
# 或者: from maskrcnn_benchmark.modeling.make_layers import conv_with_kaiming_uniform [as 別名]
def build_detnasnet_fpn_backbone(cfg):
    body = detnasnet.ShuffleNetV2DetNAS(cfg)
    out_channels = cfg.MODEL.RESNETS.BACKBONE_OUT_CHANNELS
    if '300M' in cfg.MODEL.BACKBONE.CONV_BODY:
        in_channels_list = [64, 160, 320, 640,]
    elif '1.3G' in cfg.MODEL.BACKBONE.CONV_BODY:
        in_channels_list = [96, 240, 480, 960,]
    elif '3.8G' in cfg.MODEL.BACKBONE.CONV_BODY:
        in_channels_list = [172, 432, 864, 1728,]
    else:
        raise ValueError("Wrong backbone size.")

    fpn = fpn_module.FPN(
        in_channels_list= in_channels_list,
        out_channels=out_channels,
        conv_block=conv_with_kaiming_uniform(
            cfg.MODEL.FPN.USE_GN, cfg.MODEL.FPN.USE_RELU, cfg.MODEL.FPN.USE_SYNCBN
        ),
        top_blocks=fpn_module.LastLevelMaxPool(),
    )
    if 'search' in cfg.MODEL.BACKBONE.CONV_BODY:
        return body, fpn
    model = nn.Sequential(OrderedDict([("body", body), ("fpn", fpn)]))
    model.out_channels = out_channels
    return model
開發者ID:megvii-model，項目名稱:DetNAS，代碼行數:27，代碼來源:backbone.py

示例5: build_detnasnet_fpn_p3p7_backbone

# 需要導入模塊: from maskrcnn_benchmark.modeling import make_layers [as 別名]
# 或者: from maskrcnn_benchmark.modeling.make_layers import conv_with_kaiming_uniform [as 別名]
def build_detnasnet_fpn_p3p7_backbone(cfg):
    body = detnasnet.ShuffleNetV2DetNAS(cfg)
    out_channels = cfg.MODEL.RESNETS.BACKBONE_OUT_CHANNELS
    fpn = fpn_module.FPN(
        in_channels_list=[0, 160, 320, 640,],
        out_channels=out_channels,
        conv_block=conv_with_kaiming_uniform(
            cfg.MODEL.FPN.USE_GN, cfg.MODEL.FPN.USE_RELU, cfg.MODEL.FPN.USE_SYNCBN
        ),
        top_blocks=fpn_module.LastLevelP6P7(out_channels, out_channels, cfg.MODEL.RETINANET.P6P7_USE_SYNCBN),
    )
    if 'search' in cfg.MODEL.BACKBONE.CONV_BODY:
        return body, fpn
    model = nn.Sequential(OrderedDict([("body", body), ("fpn", fpn)]))
    model.out_channels = out_channels
    return model
開發者ID:megvii-model，項目名稱:DetNAS，代碼行數:18，代碼來源:backbone.py

示例6: build_resnet_fpn_backbone

# 需要導入模塊: from maskrcnn_benchmark.modeling import make_layers [as 別名]
# 或者: from maskrcnn_benchmark.modeling.make_layers import conv_with_kaiming_uniform [as 別名]
def build_resnet_fpn_backbone(cfg):
    body = resnet.ResNet(cfg)
    in_channels_stage2 = cfg.MODEL.RESNETS.RES2_OUT_CHANNELS
    out_channels = cfg.MODEL.BACKBONE.OUT_CHANNELS
    fpn = fpn_module.FPN(
        in_channels_list=[
            in_channels_stage2,
            in_channels_stage2 * 2,
            in_channels_stage2 * 4,
            in_channels_stage2 * 8,
        ],
        out_channels=out_channels,
        conv_block=conv_with_kaiming_uniform(
            cfg.MODEL.FPN.USE_GN, cfg.MODEL.FPN.USE_RELU
        ),
        top_blocks=fpn_module.LastLevelMaxPool(),
    )
    model = nn.Sequential(OrderedDict([("body", body), ("fpn", fpn)]))
    return model
開發者ID:clw5180，項目名稱:remote_sensing_object_detection_2019，代碼行數:21，代碼來源:backbone.py

示例7: build_resnet_fpn_p4p7_backbone

# 需要導入模塊: from maskrcnn_benchmark.modeling import make_layers [as 別名]
# 或者: from maskrcnn_benchmark.modeling.make_layers import conv_with_kaiming_uniform [as 別名]
def build_resnet_fpn_p4p7_backbone(cfg):
    body = resnet.ResNet(cfg)
    in_channels_stage2 = cfg.MODEL.RESNETS.RES2_OUT_CHANNELS
    out_channels = cfg.MODEL.RESNETS.BACKBONE_OUT_CHANNELS
    in_channels_p6p7 = in_channels_stage2 * 8 if cfg.MODEL.RETINANET.USE_C5 \
        else out_channels
    fpn = fpn_module.FPN(
        in_channels_list=[
            0,
            0,
            in_channels_stage2 * 4,
            in_channels_stage2 * 8,
        ],
        out_channels=out_channels,
        conv_block=conv_with_kaiming_uniform(
            cfg.MODEL.FPN.USE_GN, cfg.MODEL.FPN.USE_RELU
        ),
        top_blocks=fpn_module.LastLevelP6P7(in_channels_p6p7, out_channels),
    )
    model = nn.Sequential(OrderedDict([("body", body), ("fpn", fpn)]))
    model.out_channels = out_channels
    return model
開發者ID:ChenJoya，項目名稱:sampling-free，代碼行數:24，代碼來源:backbone.py

示例8: build_resnet_fpn_p3p7_backbone

# 需要導入模塊: from maskrcnn_benchmark.modeling import make_layers [as 別名]
# 或者: from maskrcnn_benchmark.modeling.make_layers import conv_with_kaiming_uniform [as 別名]
def build_resnet_fpn_p3p7_backbone(cfg):
    body = resnet.ResNet(cfg)
    in_channels_stage2 = cfg.MODEL.RESNETS.RES2_OUT_CHANNELS
    out_channels = cfg.MODEL.BACKBONE.OUT_CHANNELS
    in_channels_p6p7 = in_channels_stage2 * 8 if cfg.MODEL.RETINANET.USE_C5 \
        else out_channels
    fpn = fpn_module.FPN(
        in_channels_list=[
            0,
            in_channels_stage2 * 2,
            in_channels_stage2 * 4,
            in_channels_stage2 * 8,
        ],
        out_channels=out_channels,
        conv_block=conv_with_kaiming_uniform(
            cfg.MODEL.FPN.USE_GN, cfg.MODEL.FPN.USE_RELU
        ),
        top_blocks=fpn_module.LastLevelP6P7(in_channels_p6p7, out_channels),
    )
    model = nn.Sequential(OrderedDict([("body", body), ("fpn", fpn)]))
    return model
開發者ID:mlperf，項目名稱:training，代碼行數:23，代碼來源:backbone.py

示例9: build_mnv2_fpn_backbone

# 需要導入模塊: from maskrcnn_benchmark.modeling import make_layers [as 別名]
# 或者: from maskrcnn_benchmark.modeling.make_layers import conv_with_kaiming_uniform [as 別名]
def build_mnv2_fpn_backbone(cfg):
    body = mobilenet.MobileNetV2(cfg)
    in_channels_stage2 = cfg.MODEL.BACKBONE.ENCODER_OUT_CHANNELS
    out_channels = cfg.MODEL.BACKBONE.OUT_CHANNELS
    fpn = fpn_module.FPN(
        in_channels_list=[
            0,
            in_channels_stage2[1],
            in_channels_stage2[2],
            in_channels_stage2[3],
        ],
        out_channels=out_channels,
        conv_block=conv_with_kaiming_uniform(
            cfg.MODEL.FPN.USE_GN, cfg.MODEL.FPN.USE_RELU
        ),
        top_blocks=fpn_module.LastLevelP6P7(out_channels, out_channels),
    )
    if cfg.MODEL.BACKBONE.SPLIT:
        # separate backbone and fpn output
        return body, fpn
    else:
        model = nn.Sequential(OrderedDict([("body", body), ("fpn", fpn)]))
        if cfg.MODEL.PANOPTIC.DECODER != "none":
            return model, None
    return model
開發者ID:Lausannen，項目名稱:NAS-FCOS，代碼行數:27，代碼來源:backbone.py

示例10: build_resnet_fpn_backbone

# 需要導入模塊: from maskrcnn_benchmark.modeling import make_layers [as 別名]
# 或者: from maskrcnn_benchmark.modeling.make_layers import conv_with_kaiming_uniform [as 別名]
def build_resnet_fpn_backbone(cfg):
    body = resnet.ResNet(cfg)
    in_channels_stage2 = cfg.MODEL.RESNETS.RES2_OUT_CHANNELS
    out_channels = cfg.MODEL.RESNETS.BACKBONE_OUT_CHANNELS
    fpn = fpn_module.FPN(
        in_channels_list=[
            in_channels_stage2,
            in_channels_stage2 * 2,
            in_channels_stage2 * 4,
            in_channels_stage2 * 8,
        ],
        out_channels=out_channels,
        conv_block=conv_with_kaiming_uniform(
            cfg.MODEL.FPN.USE_GN, cfg.MODEL.FPN.USE_RELU
        ),
        top_blocks=fpn_module.LastLevelMaxPool(),
        upsample_rates=cfg.MODEL.FPN.UPSAMPLE_RATE,  # add by hui
        upsample_mode=cfg.MODEL.FPN.UPSAMPLE_MODE  # add by hui
    )
    model = nn.Sequential(OrderedDict([("body", body), ("fpn", fpn)]))
    model.out_channels = out_channels
    return model
開發者ID:ucas-vg，項目名稱:TinyBenchmark，代碼行數:24，代碼來源:backbone.py

示例11: build_resnet_fpn_p3p7_backbone

# 需要導入模塊: from maskrcnn_benchmark.modeling import make_layers [as 別名]
# 或者: from maskrcnn_benchmark.modeling.make_layers import conv_with_kaiming_uniform [as 別名]
def build_resnet_fpn_p3p7_backbone(cfg):
    body = resnet.ResNet(cfg)
    in_channels_stage2 = cfg.MODEL.RESNETS.RES2_OUT_CHANNELS
    out_channels = cfg.MODEL.RESNETS.BACKBONE_OUT_CHANNELS
    in_channels_p6p7 = in_channels_stage2 * 8 if cfg.MODEL.RETINANET.USE_C5 \
        else out_channels
    fpn = fpn_module.FPN(
        in_channels_list=[
            0,
            in_channels_stage2 * 2,
            in_channels_stage2 * 4,
            in_channels_stage2 * 8,
        ],
        out_channels=out_channels,
        conv_block=conv_with_kaiming_uniform(
            cfg.MODEL.FPN.USE_GN, cfg.MODEL.FPN.USE_RELU
        ),
        top_blocks=fpn_module.LastLevelP6P7(in_channels_p6p7, out_channels),
        upsample_rates=cfg.MODEL.FPN.UPSAMPLE_RATE,   # add by hui
        upsample_mode=cfg.MODEL.FPN.UPSAMPLE_MODE     # add by hui
    )
    model = nn.Sequential(OrderedDict([("body", body), ("fpn", fpn)]))
    model.out_channels = out_channels
    return model
開發者ID:ucas-vg，項目名稱:TinyBenchmark，代碼行數:26，代碼來源:backbone.py

示例12: build_res2net_fpn_p3p7_backbone

# 需要導入模塊: from maskrcnn_benchmark.modeling import make_layers [as 別名]
# 或者: from maskrcnn_benchmark.modeling.make_layers import conv_with_kaiming_uniform [as 別名]
def build_res2net_fpn_p3p7_backbone(cfg):
    body = res2net.Res2Net(cfg)
    in_channels_stage2 = cfg.MODEL.RESNETS.RES2_OUT_CHANNELS
    out_channels = cfg.MODEL.RESNETS.BACKBONE_OUT_CHANNELS
    in_channels_p6p7 = in_channels_stage2 * 8 if cfg.MODEL.RETINANET.USE_C5 \
        else out_channels
    fpn = fpn_module.FPN(
        in_channels_list=[
            0,
            in_channels_stage2 * 2,
            in_channels_stage2 * 4,
            in_channels_stage2 * 8,
        ],
        out_channels=out_channels,
        conv_block=conv_with_kaiming_uniform(
            cfg.MODEL.FPN.USE_GN, cfg.MODEL.FPN.USE_RELU
        ),
        top_blocks=fpn_module.LastLevelP6P7(in_channels_p6p7, out_channels),
    )
    model = nn.Sequential(OrderedDict([("body", body), ("fpn", fpn)]))
    model.out_channels = out_channels
    return model


# def build_backbone(cfg):
#     assert cfg.MODEL.BACKBONE.CONV_BODY in registry.BACKBONES, \
#         "cfg.MODEL.BACKBONE.CONV_BODY: {} are not registered in registry".format(
#             cfg.MODEL.BACKBONE.CONV_BODY
#         )
#     return registry.BACKBONES[cfg.MODEL.BACKBONE.CONV_BODY](cfg)
開發者ID:Res2Net，項目名稱:Res2Net-maskrcnn，代碼行數:32，代碼來源:res2net_builder.py

示例13: add_conv_body_fpn

# 需要導入模塊: from maskrcnn_benchmark.modeling import make_layers [as 別名]
# 或者: from maskrcnn_benchmark.modeling.make_layers import conv_with_kaiming_uniform [as 別名]
def add_conv_body_fpn(cfg, dim_in=3):
    builder, arch_def = create_builder(cfg)

    body = FBNetTrunk(builder, arch_def, dim_in)
    in_channels_stage2 = cfg.MODEL.BACKBONE.ENCODER_OUT_CHANNELS
    out_channels = cfg.MODEL.BACKBONE.OUT_CHANNELS
    fpn = fpn_module.FPN(
        in_channels_list=[
            0,
            in_channels_stage2[1],
            in_channels_stage2[2],
            in_channels_stage2[3]
        ],
        out_channels=out_channels,
        conv_block=conv_with_kaiming_uniform(
            cfg.MODEL.FPN.USE_GN, cfg.MODEL.FPN.USE_RELU
        ),
        top_blocks=fpn_module.LastLevelP6P7(out_channels, out_channels),
    )
    if cfg.MODEL.BACKBONE.SPLIT:
        # separate backbone and fpn output
        return body, fpn
    else:
        model = nn.Sequential(OrderedDict([("body", body), ("fpn", fpn)]))
        if cfg.MODEL.PANOPTIC.DECODER != "none":
            return model, None
    return model
開發者ID:Lausannen，項目名稱:NAS-FCOS，代碼行數:29，代碼來源:fbnet.py


注：本文中的maskrcnn_benchmark.modeling.make_layers.conv_with_kaiming_uniform方法示例由純淨天空整理自Github/MSDocs等開源代碼及文檔管理平台，相關代碼片段篩選自各路編程大神貢獻的開源項目，源碼版權歸原作者所有，傳播和使用請參考對應項目的License；未經允許，請勿轉載。