PCGrad¶ class torchjd.aggregation.PCGrad[source]¶ Aggregator as defined in algorithm 1 of Gradient Surgery for Multi-Task Learning. __call__(matrix, /)[source]¶ Computes the aggregation from the input matrix and applies all registered hooks. Parameters: matrix (Tensor) – The Jacobian to aggregate. Return type: Tensor class torchjd.aggregation.PCGradWeighting[source]¶ Weighting giving the weights of PCGrad. __call__(gramian, /)[source]¶ Computes the vector of weights from the input gramian and applies all registered hooks. Parameters: gramian (Tensor) – The gramian from which the weights must be extracted. Return type: Tensor