LSTM version:
BERT version:
The version of the deployed model is a two-layer bidirectional LSTM.
Send a POST request to https://gender-regconition.herokuapp.com/gender with only one parameter named names which contains names separated by a comma. The response is the probability of the name being male or female.
Example: https://gender-regconition.herokuapp.com/gender?names=quốc khánh,nguyễn thanh nhàn,hương,lan hương
Response:

{
    "result": [
        [
            "quốc khánh",
            "nam",
            99.99995422363281
        ],
        [
            "nguyễn thanh nhàn",
            "nữ",
            99.87543487548828
        ],
        [
            "hương",
            "nam",
            61.728614807128906
        ],
        [
            "lan hương",
            "nữ",
            97.12590789794922
        ]
    ]
}

Character level base, four layers LSTM.
Here a some generated names by the model with seed b: bùi chí công
bùi kiên giang
bùi lọ khuê
bùi sơn lâm
bùi bảo lễ
bùi ngọc khanh
bùi thanh hoa
bùi văn nhật
bùi thành nhung
bùi hoài giang

The model has not deployed yet.