Skip to content

hybrid_interdependence

Bases: interdependence

A hybrid interdependence class combining multiple interdependence functions.

This class enables the combination of multiple interdependence functions and applies a fusion function to aggregate the results. It supports configurations for both interdependence functions and the fusion function.

Notes

Formally, given the input data batch \(\mathbf{X} \in R^{b \times m}\), we can define a set of data and structural interdependence functions \(\xi_1, \xi_2, \cdots, \xi_k: R^{b \times m} \to R^{m \times m'}\) to measure the interdependence relationships among the attributes. These functions can be effectively fused together as follows:

\[
    \begin{equation}
    \begin{aligned}
    \xi(\mathbf{X}) &= \text{fusion} \left( \xi_1(\mathbf{X}), \xi_2(\mathbf{X}), \cdots, \xi_k(\mathbf{X}) \right)\\
    &= \text{fusion} \left( \mathbf{A}_1, \mathbf{A}_2, \cdots, \mathbf{A}_k \right)\\
    &= \mathbf{A} \in R^{m \times m'},
    \end{aligned}
    \end{equation}
\]

where \(\mathbf{A}_i = \xi_i(\mathbf{X})\) denotes the interdependence matrix obtained by function \(\xi_i, \forall i \in \{1, 2, \cdots, k\}\). Different fusion strategies can be used to define the \(\text{fusion}(\cdot)\) operator used above, which will be introduced in the following subsection specifically.

Attributes:

Name Type Description
interdependence_functions list

List of interdependence function objects.
fusion_function Callable

The fusion function used to combine the outputs of the interdependence functions.
require_parameters bool

Indicates whether any interdependence or fusion function requires parameters.
require_data bool

Indicates whether any interdependence or fusion function requires input data.

Methods:

Name Description
__init__

Initializes the hybrid interdependence function.
get_function_parameter_numbers

Retrieves the number of parameters required by each interdependence function and the fusion function.
calculate_l

Calculates the total number of parameters required for the hybrid interdependence.
calculate_b_prime

Computes the number of rows in the output tensor after applying interdependence.
calculate_m_prime

Computes the number of columns in the output tensor after applying interdependence.
calculate_A

Computes the hybrid interdependence matrix.
Source code in tinybig/interdependence/hybrid_interdependence.py 
 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
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
class hybrid_interdependence(interdependence):
    r"""
        A hybrid interdependence class combining multiple interdependence functions.

        This class enables the combination of multiple interdependence functions and applies
        a fusion function to aggregate the results. It supports configurations for both
        interdependence functions and the fusion function.

        Notes
        ----------
        Formally, given the input data batch $\mathbf{X} \in R^{b \times m}$, we can define a set of data and
        structural interdependence functions $\xi_1, \xi_2, \cdots, \xi_k: R^{b \times m} \to R^{m \times m'}$
        to measure the interdependence relationships among the attributes. These functions can be effectively fused together as follows:

        $$
            \begin{equation}
            \begin{aligned}
            \xi(\mathbf{X}) &= \text{fusion} \left( \xi_1(\mathbf{X}), \xi_2(\mathbf{X}), \cdots, \xi_k(\mathbf{X}) \right)\\
            &= \text{fusion} \left( \mathbf{A}_1, \mathbf{A}_2, \cdots, \mathbf{A}_k \right)\\
            &= \mathbf{A} \in R^{m \times m'},
            \end{aligned}
            \end{equation}
        $$

        where $\mathbf{A}_i = \xi_i(\mathbf{X})$ denotes the interdependence matrix obtained by function $\xi_i, \forall i \in \{1, 2, \cdots, k\}$.
        Different fusion strategies can be used to define the $\text{fusion}(\cdot)$ operator used above, which will be introduced in the following subsection specifically.

        Attributes
        ----------
        interdependence_functions : list
            List of interdependence function objects.
        fusion_function : Callable
            The fusion function used to combine the outputs of the interdependence functions.
        require_parameters : bool
            Indicates whether any interdependence or fusion function requires parameters.
        require_data : bool
            Indicates whether any interdependence or fusion function requires input data.

        Methods
        -------
        __init__(...)
            Initializes the hybrid interdependence function.
        get_function_parameter_numbers()
            Retrieves the number of parameters required by each interdependence function and the fusion function.
        calculate_l()
            Calculates the total number of parameters required for the hybrid interdependence.
        calculate_b_prime(b=None)
            Computes the number of rows in the output tensor after applying interdependence.
        calculate_m_prime(m=None)
            Computes the number of columns in the output tensor after applying interdependence.
        calculate_A(x=None, w=None, device='cpu', *args, **kwargs)
            Computes the hybrid interdependence matrix.
    """

    def __init__(
        self,
        b: int, m: int,
        interdependence_type: str = 'attribute',
        name: str = 'hybrid_interdependence',
        interdependence_functions: list = None,
        interdependence_function_configs: list = None,
        fusion_function: Callable = None,
        fusion_function_config: dict = None,
        device: str = 'cpu',
        *args, **kwargs
    ):
        """
            Initializes the hybrid interdependence function.

            Parameters
            ----------
            b : int
                Number of rows in the input tensor.
            m : int
                Number of columns in the input tensor.
            interdependence_type : str, optional
                Type of interdependence ('attribute', 'instance', etc.). Defaults to 'attribute'.
            name : str, optional
                Name of the interdependence function. Defaults to 'hybrid_interdependence'.
            interdependence_functions : list, optional
                List of pre-initialized interdependence function objects. Defaults to None.
            interdependence_function_configs : list, optional
                List of configuration dictionaries for initializing interdependence functions. Defaults to None.
            fusion_function : Callable, optional
                Pre-initialized fusion function object. Defaults to None.
            fusion_function_config : dict, optional
                Configuration dictionary for initializing the fusion function. Defaults to None.
            device : str, optional
                Device for computation (e.g., 'cpu' or 'cuda'). Defaults to 'cpu'.
            *args : tuple
                Additional positional arguments for the parent class.
            **kwargs : dict
                Additional keyword arguments for the parent class.

            Raises
            ------
            ValueError
                If neither `interdependence_functions` nor `interdependence_function_configs` are provided.
                If neither `fusion_function` nor `fusion_function_config` is provided.
        """

        if interdependence_functions is not None:
            self.interdependence_functions = interdependence_functions
            for func in self.interdependence_functions:
                func.b = b
                func.m = m
                func.interdependence_type = interdependence_type
        elif interdependence_function_configs is not None:
            self.interdependence_functions = []
            for function_config in interdependence_function_configs:
                assert 'function_class' in function_config
                function_class = function_config['function_class']
                function_parameters = function_config['function_parameters'] if 'function_parameters' in function_config else {}
                function_parameters['b'] = b
                function_parameters['m'] = m
                function_parameters['interdependence_type'] = interdependence_type
                self.interdependence_functions.append(config.get_obj_from_str(function_class)(**function_parameters))
        else:
            raise ValueError('No interdependence functions or configurations are specified...')

        require_parameters = any([func.require_parameters for func in self.interdependence_functions])
        require_data = any([func.require_data for func in self.interdependence_functions])
        super().__init__(b=b, m=m, name=name, interdependence_type=interdependence_type, require_parameters=require_parameters, require_data=require_data, device=device, *args, **kwargs)

        if fusion_function is not None:
            self.fusion_function = fusion_function
        elif fusion_function_config is not None:
            assert 'function_class' in fusion_function_config
            function_class = fusion_function_config['function_class']
            function_parameters = fusion_function_config['function_parameters'] if 'function_parameters' in fusion_function_config else {}
            self.fusion_function = config.get_obj_from_str(function_class)(**function_parameters)
        else:
            raise ValueError('No fusion function or configurations are specified...')

    def get_function_parameter_numbers(self):
        """
            Retrieves the number of parameters required by each interdependence function and the fusion function.

            Returns
            -------
            list of int
                A list containing the number of parameters for each interdependence function and the fusion function.
        """
        l_list = []
        for func in self.interdependence_functions:
            if func.require_parameters:
                l_list.append(func.calculate_l())
            else:
                l_list.append(0)
        if self.fusion_function.require_parameters:
            l_list.append(self.fusion_function.calculate_l())
        else:
            l_list.append(0)
        return l_list

    def calculate_l(self):
        """
            Calculates the total number of parameters required for the hybrid interdependence.

            Returns
            -------
            int
                Total number of parameters required.
        """
        l_list = self.get_function_parameter_numbers()
        return sum(l_list)

    def calculate_b_prime(self, b: int = None):
        """
            Computes the number of rows in the output tensor after applying interdependence.

            Parameters
            ----------
            b : int, optional
                Number of rows in the input tensor. Defaults to `self.b`.

            Returns
            -------
            int
                The number of rows in the output tensor.
        """
        b_prime_list = [func.calculate_b_prime(b=b) for func in self.interdependence_functions]
        return self.fusion_function.calculate_n(dims=b_prime_list)

    def calculate_m_prime(self, m: int = None):
        """
            Computes the number of columns in the output tensor after applying interdependence.

            Parameters
            ----------
            m : int, optional
                Number of columns in the input tensor. Defaults to `self.m`.

            Returns
            -------
            int
                The number of columns in the output tensor.
        """
        m_prime_list = [func.calculate_b_prime(m=m) for func in self.interdependence_functions]
        return self.fusion_function.calculate_n(dims=m_prime_list)

    def calculate_A(self, x: torch.Tensor = None, w: torch.nn.Parameter = None, device: str = 'cpu', *args, **kwargs):
        r"""
            Computes the hybrid interdependence matrix.

            Formally, given the input data batch $\mathbf{X} \in R^{b \times m}$, we can define a set of data and
            structural interdependence functions $\xi_1, \xi_2, \cdots, \xi_k: R^{b \times m} \to R^{m \times m'}$
            to measure the interdependence relationships among the attributes. These functions can be effectively fused together as follows:

            $$
                \begin{equation}
                \begin{aligned}
                \xi(\mathbf{X}) &= \text{fusion} \left( \xi_1(\mathbf{X}), \xi_2(\mathbf{X}), \cdots, \xi_k(\mathbf{X}) \right)\\
                &= \text{fusion} \left( \mathbf{A}_1, \mathbf{A}_2, \cdots, \mathbf{A}_k \right)\\
                &= \mathbf{A} \in R^{m \times m'},
                \end{aligned}
                \end{equation}
            $$

            where $\mathbf{A}_i = \xi_i(\mathbf{X})$ denotes the interdependence matrix obtained by the function $\xi_i, \forall i \in \{1, 2, \cdots, k\}$.
            Different fusion strategies can be used to define the $\text{fusion}(\cdot)$ operator used above, which will be introduced in the following subsection specifically.


            Parameters
            ----------
            x : torch.Tensor, optional
                Input tensor of shape `(batch_size, num_features)`. Defaults to None.
            w : torch.nn.Parameter, optional
                Parameter tensor. Required if interdependence or fusion functions need parameters. Defaults to None.
            device : str, optional
                Device for computation (e.g., 'cpu' or 'cuda'). Defaults to 'cpu'.
            *args : tuple
                Additional positional arguments for interdependence functions.
            **kwargs : dict
                Additional keyword arguments for interdependence functions.

            Returns
            -------
            torch.Tensor
                The computed hybrid interdependence matrix.

            Raises
            ------
            AssertionError
                If the parameter tensor `w` does not match the required size.
            ValueError
                If neither input data nor parameters are provided when required.
        """
        if not self.require_data and not self.require_parameters and self.A is not None:
            return self.A
        else:
            parameter_numbers = self.get_function_parameter_numbers()
            assert w.numel() == sum(parameter_numbers)

            x = self.pre_process(x=x, device=device)

            A_list = []
            sparse_tag = False
            if w is not None:
                w_segments = torch.split(w, parameter_numbers, dim=-1)
            else:
                w_segments = [None] * len(parameter_numbers)

            for func, w_segment in zip(self.interdependence_functions, w_segments):
                A = func.calculate_A(x=x, w=w_segment, device=device, *args, **kwargs)
                if A.is_sparse:
                    A = A.to_dense()
                    sparse_tag = True
                A_list.append(A)

            if self.fusion_function.require_parameters:
                A = self.fusion_function(x=A_list, w=w_segments[-1], device=device, *args, **kwargs)
            else:
                A = self.fusion_function(x=A_list, device=device, *args, **kwargs)

            A = self.post_process(x=A, device=device)

            if self.interdependence_type in ['column', 'right', 'attribute', 'attribute_interdependence']:
                assert A.shape == (self.m, self.calculate_m_prime())
            elif self.interdependence_type in ['row', 'left', 'instance', 'instance_interdependence']:
                assert A.shape == (self.b, self.calculate_b_prime())
            if sparse_tag:
                A = A.to_sparse_coo()

            if not self.require_data and not self.require_parameters and self.A is None:
                self.A = A

            return A

__init__(b, m, interdependence_type='attribute', name='hybrid_interdependence', interdependence_functions=None, interdependence_function_configs=None, fusion_function=None, fusion_function_config=None, device='cpu', *args, **kwargs)

Initializes the hybrid interdependence function.

Parameters:

Name Type Description Default
b int

Number of rows in the input tensor.

 required
m int

Number of columns in the input tensor.

 required
interdependence_type str

Type of interdependence ('attribute', 'instance', etc.). Defaults to 'attribute'.

 'attribute'
name str

Name of the interdependence function. Defaults to 'hybrid_interdependence'.

 'hybrid_interdependence'
interdependence_functions list

List of pre-initialized interdependence function objects. Defaults to None.

 None
interdependence_function_configs list

List of configuration dictionaries for initializing interdependence functions. Defaults to None.

 None
fusion_function Callable

Pre-initialized fusion function object. Defaults to None.

 None
fusion_function_config dict

Configuration dictionary for initializing the fusion function. Defaults to None.

 None
device str

Device for computation (e.g., 'cpu' or 'cuda'). Defaults to 'cpu'.

 'cpu'
*args tuple

Additional positional arguments for the parent class.

 ()
**kwargs dict

Additional keyword arguments for the parent class.

 {}

Raises:

Type Description
ValueError

If neither interdependence_functions nor interdependence_function_configs are provided. If neither fusion_function nor fusion_function_config is provided.
Source code in tinybig/interdependence/hybrid_interdependence.py 
 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
def __init__(
    self,
    b: int, m: int,
    interdependence_type: str = 'attribute',
    name: str = 'hybrid_interdependence',
    interdependence_functions: list = None,
    interdependence_function_configs: list = None,
    fusion_function: Callable = None,
    fusion_function_config: dict = None,
    device: str = 'cpu',
    *args, **kwargs
):
    """
        Initializes the hybrid interdependence function.

        Parameters
        ----------
        b : int
            Number of rows in the input tensor.
        m : int
            Number of columns in the input tensor.
        interdependence_type : str, optional
            Type of interdependence ('attribute', 'instance', etc.). Defaults to 'attribute'.
        name : str, optional
            Name of the interdependence function. Defaults to 'hybrid_interdependence'.
        interdependence_functions : list, optional
            List of pre-initialized interdependence function objects. Defaults to None.
        interdependence_function_configs : list, optional
            List of configuration dictionaries for initializing interdependence functions. Defaults to None.
        fusion_function : Callable, optional
            Pre-initialized fusion function object. Defaults to None.
        fusion_function_config : dict, optional
            Configuration dictionary for initializing the fusion function. Defaults to None.
        device : str, optional
            Device for computation (e.g., 'cpu' or 'cuda'). Defaults to 'cpu'.
        *args : tuple
            Additional positional arguments for the parent class.
        **kwargs : dict
            Additional keyword arguments for the parent class.

        Raises
        ------
        ValueError
            If neither `interdependence_functions` nor `interdependence_function_configs` are provided.
            If neither `fusion_function` nor `fusion_function_config` is provided.
    """

    if interdependence_functions is not None:
        self.interdependence_functions = interdependence_functions
        for func in self.interdependence_functions:
            func.b = b
            func.m = m
            func.interdependence_type = interdependence_type
    elif interdependence_function_configs is not None:
        self.interdependence_functions = []
        for function_config in interdependence_function_configs:
            assert 'function_class' in function_config
            function_class = function_config['function_class']
            function_parameters = function_config['function_parameters'] if 'function_parameters' in function_config else {}
            function_parameters['b'] = b
            function_parameters['m'] = m
            function_parameters['interdependence_type'] = interdependence_type
            self.interdependence_functions.append(config.get_obj_from_str(function_class)(**function_parameters))
    else:
        raise ValueError('No interdependence functions or configurations are specified...')

    require_parameters = any([func.require_parameters for func in self.interdependence_functions])
    require_data = any([func.require_data for func in self.interdependence_functions])
    super().__init__(b=b, m=m, name=name, interdependence_type=interdependence_type, require_parameters=require_parameters, require_data=require_data, device=device, *args, **kwargs)

    if fusion_function is not None:
        self.fusion_function = fusion_function
    elif fusion_function_config is not None:
        assert 'function_class' in fusion_function_config
        function_class = fusion_function_config['function_class']
        function_parameters = fusion_function_config['function_parameters'] if 'function_parameters' in fusion_function_config else {}
        self.fusion_function = config.get_obj_from_str(function_class)(**function_parameters)
    else:
        raise ValueError('No fusion function or configurations are specified...')

calculate_A(x=None, w=None, device='cpu', *args, **kwargs)

Computes the hybrid interdependence matrix.

Formally, given the input data batch \(\mathbf{X} \in R^{b \times m}\), we can define a set of data and structural interdependence functions \(\xi_1, \xi_2, \cdots, \xi_k: R^{b \times m} \to R^{m \times m'}\) to measure the interdependence relationships among the attributes. These functions can be effectively fused together as follows:

\[
    \begin{equation}
    \begin{aligned}
    \xi(\mathbf{X}) &= \text{fusion} \left( \xi_1(\mathbf{X}), \xi_2(\mathbf{X}), \cdots, \xi_k(\mathbf{X}) \right)\\
    &= \text{fusion} \left( \mathbf{A}_1, \mathbf{A}_2, \cdots, \mathbf{A}_k \right)\\
    &= \mathbf{A} \in R^{m \times m'},
    \end{aligned}
    \end{equation}
\]

where \(\mathbf{A}_i = \xi_i(\mathbf{X})\) denotes the interdependence matrix obtained by the function \(\xi_i, \forall i \in \{1, 2, \cdots, k\}\). Different fusion strategies can be used to define the \(\text{fusion}(\cdot)\) operator used above, which will be introduced in the following subsection specifically.

Parameters:

Name Type Description Default
x Tensor

Input tensor of shape (batch_size, num_features). Defaults to None.

 None
w Parameter

Parameter tensor. Required if interdependence or fusion functions need parameters. Defaults to None.

 None
device str

Device for computation (e.g., 'cpu' or 'cuda'). Defaults to 'cpu'.

 'cpu'
*args tuple

Additional positional arguments for interdependence functions.

 ()
**kwargs dict

Additional keyword arguments for interdependence functions.

 {}

Returns:

Type Description
Tensor

The computed hybrid interdependence matrix.

Raises:

Type Description
AssertionError

If the parameter tensor w does not match the required size.
ValueError

If neither input data nor parameters are provided when required.
Source code in tinybig/interdependence/hybrid_interdependence.py 
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
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
def calculate_A(self, x: torch.Tensor = None, w: torch.nn.Parameter = None, device: str = 'cpu', *args, **kwargs):
    r"""
        Computes the hybrid interdependence matrix.

        Formally, given the input data batch $\mathbf{X} \in R^{b \times m}$, we can define a set of data and
        structural interdependence functions $\xi_1, \xi_2, \cdots, \xi_k: R^{b \times m} \to R^{m \times m'}$
        to measure the interdependence relationships among the attributes. These functions can be effectively fused together as follows:

        $$
            \begin{equation}
            \begin{aligned}
            \xi(\mathbf{X}) &= \text{fusion} \left( \xi_1(\mathbf{X}), \xi_2(\mathbf{X}), \cdots, \xi_k(\mathbf{X}) \right)\\
            &= \text{fusion} \left( \mathbf{A}_1, \mathbf{A}_2, \cdots, \mathbf{A}_k \right)\\
            &= \mathbf{A} \in R^{m \times m'},
            \end{aligned}
            \end{equation}
        $$

        where $\mathbf{A}_i = \xi_i(\mathbf{X})$ denotes the interdependence matrix obtained by the function $\xi_i, \forall i \in \{1, 2, \cdots, k\}$.
        Different fusion strategies can be used to define the $\text{fusion}(\cdot)$ operator used above, which will be introduced in the following subsection specifically.


        Parameters
        ----------
        x : torch.Tensor, optional
            Input tensor of shape `(batch_size, num_features)`. Defaults to None.
        w : torch.nn.Parameter, optional
            Parameter tensor. Required if interdependence or fusion functions need parameters. Defaults to None.
        device : str, optional
            Device for computation (e.g., 'cpu' or 'cuda'). Defaults to 'cpu'.
        *args : tuple
            Additional positional arguments for interdependence functions.
        **kwargs : dict
            Additional keyword arguments for interdependence functions.

        Returns
        -------
        torch.Tensor
            The computed hybrid interdependence matrix.

        Raises
        ------
        AssertionError
            If the parameter tensor `w` does not match the required size.
        ValueError
            If neither input data nor parameters are provided when required.
    """
    if not self.require_data and not self.require_parameters and self.A is not None:
        return self.A
    else:
        parameter_numbers = self.get_function_parameter_numbers()
        assert w.numel() == sum(parameter_numbers)

        x = self.pre_process(x=x, device=device)

        A_list = []
        sparse_tag = False
        if w is not None:
            w_segments = torch.split(w, parameter_numbers, dim=-1)
        else:
            w_segments = [None] * len(parameter_numbers)

        for func, w_segment in zip(self.interdependence_functions, w_segments):
            A = func.calculate_A(x=x, w=w_segment, device=device, *args, **kwargs)
            if A.is_sparse:
                A = A.to_dense()
                sparse_tag = True
            A_list.append(A)

        if self.fusion_function.require_parameters:
            A = self.fusion_function(x=A_list, w=w_segments[-1], device=device, *args, **kwargs)
        else:
            A = self.fusion_function(x=A_list, device=device, *args, **kwargs)

        A = self.post_process(x=A, device=device)

        if self.interdependence_type in ['column', 'right', 'attribute', 'attribute_interdependence']:
            assert A.shape == (self.m, self.calculate_m_prime())
        elif self.interdependence_type in ['row', 'left', 'instance', 'instance_interdependence']:
            assert A.shape == (self.b, self.calculate_b_prime())
        if sparse_tag:
            A = A.to_sparse_coo()

        if not self.require_data and not self.require_parameters and self.A is None:
            self.A = A

        return A

calculate_b_prime(b=None)

Computes the number of rows in the output tensor after applying interdependence.

Parameters:

Name Type Description Default
b int

Number of rows in the input tensor. Defaults to self.b.

 None

Returns:

Type Description
int

The number of rows in the output tensor.
Source code in tinybig/interdependence/hybrid_interdependence.py 
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
def calculate_b_prime(self, b: int = None):
    """
        Computes the number of rows in the output tensor after applying interdependence.

        Parameters
        ----------
        b : int, optional
            Number of rows in the input tensor. Defaults to `self.b`.

        Returns
        -------
        int
            The number of rows in the output tensor.
    """
    b_prime_list = [func.calculate_b_prime(b=b) for func in self.interdependence_functions]
    return self.fusion_function.calculate_n(dims=b_prime_list)

calculate_l()

Calculates the total number of parameters required for the hybrid interdependence.

Returns:

Type Description
int

Total number of parameters required.
Source code in tinybig/interdependence/hybrid_interdependence.py 
177
178
179
180
181
182
183
184
185
186
187
def calculate_l(self):
    """
        Calculates the total number of parameters required for the hybrid interdependence.

        Returns
        -------
        int
            Total number of parameters required.
    """
    l_list = self.get_function_parameter_numbers()
    return sum(l_list)

calculate_m_prime(m=None)

Computes the number of columns in the output tensor after applying interdependence.

Parameters:

Name Type Description Default
m int

Number of columns in the input tensor. Defaults to self.m.

 None

Returns:

Type Description
int

The number of columns in the output tensor.
Source code in tinybig/interdependence/hybrid_interdependence.py 
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
def calculate_m_prime(self, m: int = None):
    """
        Computes the number of columns in the output tensor after applying interdependence.

        Parameters
        ----------
        m : int, optional
            Number of columns in the input tensor. Defaults to `self.m`.

        Returns
        -------
        int
            The number of columns in the output tensor.
    """
    m_prime_list = [func.calculate_b_prime(m=m) for func in self.interdependence_functions]
    return self.fusion_function.calculate_n(dims=m_prime_list)

get_function_parameter_numbers()

Retrieves the number of parameters required by each interdependence function and the fusion function.

Returns:

Type Description
list of int

A list containing the number of parameters for each interdependence function and the fusion function.
Source code in tinybig/interdependence/hybrid_interdependence.py 
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
def get_function_parameter_numbers(self):
    """
        Retrieves the number of parameters required by each interdependence function and the fusion function.

        Returns
        -------
        list of int
            A list containing the number of parameters for each interdependence function and the fusion function.
    """
    l_list = []
    for func in self.interdependence_functions:
        if func.require_parameters:
            l_list.append(func.calculate_l())
        else:
            l_list.append(0)
    if self.fusion_function.require_parameters:
        l_list.append(self.fusion_function.calculate_l())
    else:
        l_list.append(0)
    return l_list