def D_arch()

in BigGAN_PyTorch/BigGAN.py [0:0]

• 43 lines of code
• 17 McCabe index (conditional complexity)

def D_arch(ch=64, attention="64", ksize="333333", dilation="111111"):
    arch = {}
    arch[256] = {
        "in_channels": [3] + [ch * item for item in [1, 2, 4, 8, 8, 16]],
        "out_channels": [item * ch for item in [1, 2, 4, 8, 8, 16, 16]],
        "downsample": [True] * 6 + [False],
        "resolution": [128, 64, 32, 16, 8, 4, 4],
        "attention": {
            2 ** i: 2 ** i in [int(item) for item in attention.split("_")]
            for i in range(2, 8)
        },
    }
    arch[128] = {
        "in_channels": [3] + [ch * item for item in [1, 2, 4, 8, 16]],
        "out_channels": [item * ch for item in [1, 2, 4, 8, 16, 16]],
        "downsample": [True] * 5 + [False],
        "resolution": [64, 32, 16, 8, 4, 4],
        "attention": {
            2 ** i: 2 ** i in [int(item) for item in attention.split("_")]
            for i in range(2, 8)
        },
    }
    arch[64] = {
        "in_channels": [3] + [ch * item for item in [1, 2, 4, 8]],
        "out_channels": [item * ch for item in [1, 2, 4, 8, 16]],
        "downsample": [True] * 4 + [False],
        "resolution": [32, 16, 8, 4, 4],
        "attention": {
            2 ** i: 2 ** i in [int(item) for item in attention.split("_")]
            for i in range(2, 7)
        },
    }
    arch[32] = {
        "in_channels": [3] + [item * ch for item in [4, 4, 4]],
        "out_channels": [item * ch for item in [4, 4, 4, 4]],
        "downsample": [True, True, False, False],
        "resolution": [16, 16, 16, 16],
        "attention": {
            2 ** i: 2 ** i in [int(item) for item in attention.split("_")]
            for i in range(2, 6)
        },
    }
    return arch