Classify the sequence task

Background

Sequence problems are categorised by how many elements go in and come out:

input length	output length	category
1	1	one-to-one
> 1	1	many-to-one (e.g. sentiment)
1	> 1	one-to-many (e.g. captioning)
> 1	> 1	many-to-many (e.g. translation)

Problem statement

Implement task_category(input_len, output_len) returning the category string.

Input

• input_len — int $\ge 1$.
• output_len — int $\ge 1$.

Output

Returns one of "one-to-one", "many-to-one", "one-to-many", "many-to-many".

Examples

Example 1 — sentiment classification

Input:  input_len = 5, output_len = 1
Output: "many-to-one"

Example 2 — translation

Input:  input_len = 5, output_len = 7
Output: "many-to-many"

Constraints

• "many" means length > 1; "one" means length == 1.
• Return the exact category string.

Notes

• The category tells you the architecture shape: where the network reads inputs and where it must emit outputs.