LSTMP

Description

Computes a one-layer LSTMP.

Input

• x: input tensor of type float16.
• wx: input tensor of type float16.
• bias: input tensor of type float16.
• wr: input tensor of type float16.
• project: input tensor of type float16.
• real_mask (optional): input tensor of type float16.
• init_h (optional): input tensor of type float16.
• init_c (optional): input tensor of type float16.

Output

1. y: output tensor of type float16.
2. output_h: output tensor of type float16.
3. output_c: output tensor of type float16.

Constraints

This operator is customized for specific user scenarios and is not recommended in irrelevant scenarios.

ONNX Opset Support

Opset v8/v9/v10/v11/v12/v13/v14/v15/v16