BINN
This is the API reference for the BINN-package. For usage examples, see Examples. Note that the API is still stabilizing and will undergo changes.
BINN
Bases: LightningModule
Implements a Biologically Informed Neural Network (BINN). The BINN is implemented using the Lightning-framework. If you are unfamiliar with PyTorch, we suggest visiting their website: https://pytorch.org/
Parameters:
Name | Type | Description | Default |
---|---|---|---|
pathways |
Network
|
A Network object that defines the network topology. |
required |
activation |
str
|
Activation function to use. Defaults to "tanh". |
'tanh'
|
weight |
Tensor
|
Weights for loss function. Defaults to torch.Tensor([1, 1]). |
tensor([1, 1])
|
learning_rate |
float
|
Learning rate for optimizer. Defaults to 1e-4. |
0.0001
|
n_layers |
int
|
Number of layers in the network. Defaults to 4. |
4
|
scheduler |
str
|
Learning rate scheduler to use. Defaults to "plateau". |
'plateau'
|
optimizer |
str
|
Optimizer to use. Defaults to "adam". |
'adam'
|
validate |
bool
|
Whether to use validation data during training. Defaults to False. |
False
|
n_outputs |
int
|
Number of output nodes. Defaults to 2. |
2
|
dropout |
float
|
Dropout probability. Defaults to 0. |
0
|
residual |
bool
|
Whether to use residual connections. Defaults to False. |
False
|
Attributes:
Name | Type | Description |
---|---|---|
residual |
bool
|
Whether to use residual connections. |
pathways |
Network
|
A Network object that defines the network topology. |
n_layers |
int
|
Number of layers in the network. |
layer_names |
List[str]
|
List of layer names. |
features |
Index
|
A pandas Index object containing the input features. |
layers |
Module
|
The layers of the BINN. |
loss |
Module
|
The loss function used during training. |
learning_rate |
float
|
Learning rate for optimizer. |
scheduler |
str
|
Learning rate scheduler used. |
optimizer |
str
|
Optimizer used. |
validate |
bool
|
Whether to use validation data during training. |
Source code in binn/binn.py
11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 |
|
configure_optimizers()
Configures the optimizer and learning rate scheduler for training the BINN.
Returns:
Type | Description |
---|---|
A list of optimizers and a list of learning rate schedulers. |
Source code in binn/binn.py
187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 |
|
forward(x)
Performs a forward pass through the BINN.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
x |
Tensor
|
The input tensor to the BINN. |
required |
Returns:
Type | Description |
---|---|
tensor
|
torch.Tensor: The output tensor of the BINN. |
Source code in binn/binn.py
116 117 118 119 120 121 122 123 124 125 126 127 128 129 |
|
get_connectivity_matrices()
Returns the connectivity matrices underlying the BINN.
Returns:
Type | Description |
---|---|
list
|
The connectivity matrices as a list of Pandas DataFrames. |
Source code in binn/binn.py
228 229 230 231 232 233 234 235 |
|
init_weights()
Initializes the trainable parameters of the BINN.
Source code in binn/binn.py
243 244 245 246 247 |
|
reset_params()
Resets the trainable parameters of the BINN.
Source code in binn/binn.py
237 238 239 240 241 |
|
test_step(batch, _)
Implements a single testing step for the BINN.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
batch |
A tuple containing the input and output data for the current batch. |
required | |
_ |
The batch index, which is not used. |
required |
Source code in binn/binn.py
171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 |
|
training_step(batch, _)
Performs a single training step for the BINN.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
batch |
The batch of data to use for the training step. |
required | |
_ |
Not used. |
required |
Returns:
Type | Description |
---|---|
torch.Tensor: The loss tensor for the training step. |
Source code in binn/binn.py
131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 |
|
validation_step(batch, _)
Implements a single validation step for the BINN.
Parameters:
Name | Type | Description | Default |
---|---|---|---|
batch |
A tuple containing the input and output data for the current batch. |
required | |
_ |
The batch index, which is not used. |
required |
Source code in binn/binn.py
153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 |
|