1 write to query
Microsoft.ML.TorchSharp (1)
Roberta\Modules\AttentionSelf.cs (1)
37query = torch.nn.Linear(hiddenSize, hiddenSize, true);
2 references to query
Microsoft.ML.TorchSharp (2)
Roberta\Modules\AttentionSelf.cs (2)
48var mixedQueryLayer = query.forward(hiddenStates); 92query.Dispose();