Two Model Inputs for Assignment 2?

Course Q&A Natural Language Processing NLP with Attention Models
1

In evaluation part of assignment #2, for next_symbol function, why the model needs two inputs?

# model expects a tuple containing two padded tensors (with batch)
output, _ = model((padded_with_batch, padded_with_batch))

The reason I raised this question is that in assignment 1, the model needs two different input (english, german)
NMTAttn((input_tokens, padded_with_batch))

So compared to assignment 1, is the second input in for model in assignment 2 redundant?

2

Hi YIHUI,

It’s needed for the model to work, as the model expects a tuple as input.

3

image
image1087×638 102 KB

So does it mean the second input in the tuple is useless and it’s just an placeholder for this model structure?

4

Yes, that’s how I understand it.