[ { "function_name": "LogLinearNoise.__init__", "file_path": "noise_schedule.py", "function_calls": [ { "function_call_name": "super().__init__", "functionality": "Initializes the parent class `Noise`.", "input": "", "output": "An initialized `Noise` object instance.", "location_or_namespace": "noise_schedule.py" }, { "function_call_name": "self.total_noise", "functionality": "Computes the total noise at a specific timestep.", "input": "t (torch.Tensor)", "output": "Total noise value (torch.Tensor).", "location_or_namespace": "noise_schedule.py" }, { "function_call_name": "torch.tensor", "functionality": "Constructs a torch tensor.", "input": "data (object), dtype (torch.dtype, optional)", "output": "A torch.Tensor.", "location_or_namespace": "torch" } ] }, { "function_name": "LogLinearNoise.k", "file_path": "noise_schedule.py", "function_calls": [ { "function_call_name": "torch.tensor", "functionality": "Constructs a torch tensor.", "input": "data (object), dtype (torch.dtype, optional)", "output": "A torch.Tensor.", "location_or_namespace": "torch" } ] }, { "function_name": "LogLinearNoise.rate_noise", "file_path": "noise_schedule.py", "function_calls": [] }, { "function_name": "LogLinearNoise.total_noise", "file_path": "noise_schedule.py", "function_calls": [ { "function_call_name": "torch.log1p", "functionality": "Returns a new tensor with the natural logarithm of (1 + input).", "input": "input (torch.Tensor)", "output": "A torch.Tensor.", "location_or_namespace": "torch" } ] }, { "function_name": "PowerMeanNoise.__init__", "file_path": "noise_schedule.py", "function_calls": [ { "function_call_name": "super().__init__", "functionality": "Initializes the parent class `Noise`.", "input": "", "output": "An initialized `Noise` object instance.", "location_or_namespace": "noise_schedule.py" } ] }, { "function_name": "PowerMeanNoise.rho", "file_path": "noise_schedule.py", "function_calls": [ { "function_call_name": "torch.tensor", "functionality": "Constructs a torch tensor.", "input": "data (object), dtype (torch.dtype, optional)", "output": "A torch.Tensor.", "location_or_namespace": "torch" } ] }, { "function_name": "PowerMeanNoise.total_noise", "file_path": "noise_schedule.py", "function_calls": [ { "function_call_name": "self.rho", "functionality": "Returns the fixed rho value for the noise schedule.", "input": "", "output": "The rho value as a torch.Tensor.", "location_or_namespace": "noise_schedule.py" }, { "function_call_name": "torch.Tensor.pow", "functionality": "Takes the power of each element in the input tensor with an exponent and returns a tensor with the result.", "input": "exponent (float or torch.Tensor)", "output": "A torch.Tensor.", "location_or_namespace": "torch" } ] }, { "function_name": "PowerMeanNoise.inverse_to_t", "file_path": "noise_schedule.py", "function_calls": [ { "function_call_name": "self.rho", "functionality": "Returns the fixed rho value for the noise schedule.", "input": "", "output": "The rho value as a torch.Tensor.", "location_or_namespace": "noise_schedule.py" }, { "function_call_name": "torch.Tensor.pow", "functionality": "Takes the power of each element in the input tensor with an exponent and returns a tensor with the result.", "input": "exponent (float or torch.Tensor)", "output": "A torch.Tensor.", "location_or_namespace": "torch" } ] }, { "function_name": "PowerMeanNoise_PerColumn.__init__", "file_path": "noise_schedule.py", "function_calls": [ { "function_call_name": "super().__init__", "functionality": "Initializes the parent class `nn.Module`.", "input": "", "output": "An initialized `nn.Module` object instance.", "location_or_namespace": "torch.nn" }, { "function_call_name": "torch.nn.Parameter", "functionality": "A kind of Tensor that is to be considered a module parameter.", "input": "data (torch.Tensor)", "output": "A torch.nn.Parameter object.", "location_or_namespace": "torch.nn" }, { "function_call_name": "torch.tensor", "functionality": "Constructs a torch tensor.", "input": "data (object), dtype (torch.dtype, optional)", "output": "A torch.Tensor.", "location_or_namespace": "torch" } ] }, { "function_name": "PowerMeanNoise_PerColumn.rho", "file_path": "noise_schedule.py", "function_calls": [ { "function_call_name": "torch.nn.functional.softplus", "functionality": "Applies the softplus function element-wise: Softplus(x) = 1/beta * log(1 + exp(beta * x)).", "input": "input (torch.Tensor)", "output": "A torch.Tensor.", "location_or_namespace": "torch.nn.functional" } ] }, { "function_name": "PowerMeanNoise_PerColumn.total_noise", "file_path": "noise_schedule.py", "function_calls": [ { "function_call_name": "self.rho", "functionality": "Returns the softplus-transformed learnable rho for all numerical columns.", "input": "", "output": "A torch.Tensor of rho values.", "location_or_namespace": "noise_schedule.py" }, { "function_call_name": "torch.Tensor.pow", "functionality": "Takes the power of each element in the input tensor with an exponent and returns a tensor with the result.", "input": "exponent (float or torch.Tensor)", "output": "A torch.Tensor.", "location_or_namespace": "torch" } ] }, { "function_name": "PowerMeanNoise_PerColumn.rate_noise", "file_path": "noise_schedule.py", "function_calls": [] }, { "function_name": "PowerMeanNoise_PerColumn.inverse_to_t", "file_path": "noise_schedule.py", "function_calls": [ { "function_call_name": "self.rho", "functionality": "Returns the softplus-transformed learnable rho for all numerical columns.", "input": "", "output": "A torch.Tensor of rho values.", "location_or_namespace": "noise_schedule.py" }, { "function_call_name": "torch.Tensor.pow", "functionality": "Takes the power of each element in the input tensor with an exponent and returns a tensor with the result.", "input": "exponent (float or torch.Tensor)", "output": "A torch.Tensor.", "location_or_namespace": "torch" } ] }, { "function_name": "LogLinearNoise_PerColumn.__init__", "file_path": "noise_schedule.py", "function_calls": [ { "function_call_name": "super().__init__", "functionality": "Initializes the parent class `nn.Module`.", "input": "", "output": "An initialized `nn.Module` object instance.", "location_or_namespace": "torch.nn" }, { "function_call_name": "torch.nn.Parameter", "functionality": "A kind of Tensor that is to be considered a module parameter.", "input": "data (torch.Tensor)", "output": "A torch.nn.Parameter object.", "location_or_namespace": "torch.nn" }, { "function_call_name": "torch.tensor", "functionality": "Constructs a torch tensor.", "input": "data (object), dtype (torch.dtype, optional)", "output": "A torch.Tensor.", "location_or_namespace": "torch" } ] }, { "function_name": "LogLinearNoise_PerColumn.k", "file_path": "noise_schedule.py", "function_calls": [ { "function_call_name": "torch.nn.functional.softplus", "functionality": "Applies the softplus function element-wise: Softplus(x) = 1/beta * log(1 + exp(beta * x)).", "input": "input (torch.Tensor)", "output": "A torch.Tensor.", "location_or_namespace": "torch.nn.functional" } ] }, { "function_name": "LogLinearNoise_PerColumn.rate_noise", "file_path": "noise_schedule.py", "function_calls": [ { "function_call_name": "self.k", "functionality": "Returns the softplus-transformed learnable k for all categorical columns.", "input": "", "output": "A torch.Tensor of k values.", "location_or_namespace": "noise_schedule.py" }, { "function_call_name": "torch.Tensor.pow", "functionality": "Takes the power of each element in the input tensor with an exponent and returns a tensor with the result.", "input": "exponent (float or torch.Tensor)", "output": "A torch.Tensor.", "location_or_namespace": "torch" } ] }, { "function_name": "LogLinearNoise_PerColumn.total_noise", "file_path": "noise_schedule.py", "function_calls": [ { "function_call_name": "self.k", "functionality": "Returns the softplus-transformed learnable k for all categorical columns.", "input": "", "output": "A torch.Tensor of k values.", "location_or_namespace": "noise_schedule.py" }, { "function_call_name": "torch.Tensor.pow", "functionality": "Takes the power of each element in the input tensor with an exponent and returns a tensor with the result.", "input": "exponent (float or torch.Tensor)", "output": "A torch.Tensor.", "location_or_namespace": "torch" }, { "function_call_name": "torch.log1p", "functionality": "Returns a new tensor with the natural logarithm of (1 + input).", "input": "input (torch.Tensor)", "output": "A torch.Tensor.", "location_or_namespace": "torch" } ] } ]