sebastiansarasti commited on
Commit
fa00ec7
·
verified ·
1 Parent(s): 724e02f

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +2 -63
README.md CHANGED
@@ -57,68 +57,7 @@ class ModelColorization(nn.Module, PyTorchModelHubMixin):
57
  x = self.decoder(x)
58
  return x
59
 
60
- Here's your model card in Markdown format:
61
-
62
- md
63
- Copy code
64
- ---
65
- tags:
66
- - autoencoder
67
- - image-colorization
68
- - pytorch
69
- - pytorch_model_hub_mixin
70
- ---
71
-
72
- # Model Colorization Autoencoder
73
-
74
- ## Model Description
75
-
76
- This autoencoder model is designed for image colorization. It takes grayscale images as input and outputs colorized versions of those images. The model architecture consists of an encoder-decoder structure, where the encoder compresses the input image into a latent representation, and the decoder reconstructs the image in color.
77
-
78
- ### Architecture
79
-
80
- - **Encoder**: The encoder comprises three convolutional layers followed by max pooling and ReLU activations, each paired with batch normalization. It ends with a flattening layer and a fully connected layer to produce a latent vector.
81
- - **Decoder**: The decoder mirrors the encoder, using linear and transposed convolutional layers with ReLU activations and batch normalization. The final layer outputs a color image using a sigmoid activation function.
82
-
83
- The architecture details are as follows:
84
- ```python
85
- class ModelColorization(nn.Module, PyTorchModelHubMixin):
86
- def __init__(self):
87
- super(ModelColorization, self).__init__()
88
- self.encoder = nn.Sequential(
89
- nn.Conv2d(1, 64, kernel_size=3, stride=1, padding=1),
90
- nn.MaxPool2d(kernel_size=2, stride=2),
91
- nn.ReLU(),
92
- nn.BatchNorm2d(64),
93
- nn.Conv2d(64, 32, kernel_size=3, stride=1, padding=1),
94
- nn.MaxPool2d(kernel_size=2, stride=2),
95
- nn.ReLU(),
96
- nn.BatchNorm2d(32),
97
- nn.Conv2d(32, 16, kernel_size=3, stride=1, padding=1),
98
- nn.MaxPool2d(kernel_size=2, stride=2),
99
- nn.ReLU(),
100
- nn.BatchNorm2d(16),
101
- nn.Flatten(),
102
- nn.Linear(16*45*45, 4000),
103
- )
104
- self.decoder = nn.Sequential(
105
- nn.Linear(4000, 16 * 45 * 45),
106
- nn.ReLU(),
107
- nn.Unflatten(1, (16, 45, 45)),
108
- nn.ConvTranspose2d(16, 32, kernel_size=3, stride=2, padding=1, output_padding=1),
109
- nn.ReLU(),
110
- nn.BatchNorm2d(32),
111
- nn.ConvTranspose2d(32, 64, kernel_size=3, stride=2, padding=1, output_padding=1),
112
- nn.ReLU(),
113
- nn.BatchNorm2d(64),
114
- nn.ConvTranspose2d(64, 3, kernel_size=3, stride=2, padding=1, output_padding=1),
115
- nn.Sigmoid()
116
- )
117
-
118
- def forward(self, x):
119
- x = self.encoder(x)
120
- x = self.decoder(x)
121
- return x
122
 
123
  ### Training Details
124
  The model was trained using PyTorch for 5 epochs. Here are the training and validation losses observed during the training:
@@ -140,4 +79,4 @@ pip install torch torchvision transformers
140
  from transformers import AutoModel
141
 
142
  model = AutoModel.from_pretrained("sebastiansarasti/AutoEncoderImageColorization")
143
- ```
 
57
  x = self.decoder(x)
58
  return x
59
 
60
+ ```
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
61
 
62
  ### Training Details
63
  The model was trained using PyTorch for 5 epochs. Here are the training and validation losses observed during the training:
 
79
  from transformers import AutoModel
80
 
81
  model = AutoModel.from_pretrained("sebastiansarasti/AutoEncoderImageColorization")
82
+ ```python