| """ |
| VGG16 in MLX with endpoints for relu1_2, relu2_2, relu3_3, relu4_2, relu4_3, |
| relu5_2, relu5_3. Loads weights from a torchvision-exported npz |
| (see export_vgg16_npz.py). |
| """ |
|
|
| import mlx.core as mx |
| import mlx.nn as nn |
| import numpy as np |
|
|
|
|
| def _conv(in_ch, out_ch, kernel_size=3, padding=1): |
| return nn.Conv2d(in_ch, out_ch, kernel_size=kernel_size, padding=padding, bias=True) |
|
|
|
|
| class VGG16(nn.Module): |
| def __init__(self): |
| super().__init__() |
| self.layers = [ |
| _conv(3, 64), |
| nn.ReLU(), |
| _conv(64, 64), |
| nn.ReLU(), |
| nn.MaxPool2d(kernel_size=2, stride=2), |
| _conv(64, 128), |
| nn.ReLU(), |
| _conv(128, 128), |
| nn.ReLU(), |
| nn.MaxPool2d(kernel_size=2, stride=2), |
| _conv(128, 256), |
| nn.ReLU(), |
| _conv(256, 256), |
| nn.ReLU(), |
| _conv(256, 256), |
| nn.ReLU(), |
| nn.MaxPool2d(kernel_size=2, stride=2), |
| _conv(256, 512), |
| nn.ReLU(), |
| _conv(512, 512), |
| nn.ReLU(), |
| _conv(512, 512), |
| nn.ReLU(), |
| nn.MaxPool2d(kernel_size=2, stride=2), |
| _conv(512, 512), |
| nn.ReLU(), |
| _conv(512, 512), |
| nn.ReLU(), |
| _conv(512, 512), |
| nn.ReLU(), |
| nn.MaxPool2d(kernel_size=2, stride=2), |
| ] |
|
|
| |
| self.endpoint_indices = { |
| "relu1_2": 3, |
| "relu2_2": 8, |
| "relu3_3": 15, |
| "relu4_1": 18, |
| "relu4_2": 20, |
| "relu4_3": 22, |
| "relu5_1": 25, |
| "relu5_2": 27, |
| "relu5_3": 29, |
| } |
|
|
| def forward_with_endpoints(self, x): |
| endpoints = {} |
| for idx, layer in enumerate(self.layers): |
| x = layer(x) |
| for name, i in self.endpoint_indices.items(): |
| if idx == i: |
| endpoints[name] = x |
| return x, endpoints |
|
|
| def __call__(self, x): |
| _, endpoints = self.forward_with_endpoints(x) |
| return endpoints |
|
|
| def load_npz(self, path: str): |
| data = np.load(path) |
|
|
| def load_weight(key, transpose=False): |
| if key in data: |
| w = data[key] |
| elif f"{key}_int8" in data: |
| w_int8 = data[f"{key}_int8"] |
| scale = data[f"{key}_scale"] |
| w = w_int8.astype(scale.dtype) * scale |
| else: |
| raise ValueError(f"Missing key {key} in npz") |
| |
| if transpose and w.ndim == 4: |
| w = np.transpose(w, (0, 2, 3, 1)) |
| return mx.array(w) |
|
|
| conv_indices = [0, 2, 5, 7, 10, 12, 14, 17, 19, 21, 24, 26, 28] |
| for idx in conv_indices: |
| conv = self.layers[idx] |
| weight_key = f"features.{idx}.weight" |
| bias_key = f"features.{idx}.bias" |
| |
| conv.weight = load_weight(weight_key, transpose=True) |
| conv.bias = load_weight(bias_key) |
|
|