# FiBiNET

**Repository Path**: magicat128/fi-bi-net

## Basic Information

- **Project Name**: FiBiNET
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: MIT
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2021-05-25
- **Last Updated**: 2021-05-25

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# FiBiNET

This is a TenforFlow implementation of FiBiNET on Ascend 910 for CTR prediction task, as described in paper:

Tongwen Huang, Zhiqi Zhang, Junlin Zhang. [FiBiNET: Combining Feature Importance and Bilinear feature Interaction for Click-Through Rate Prediction](https://arxiv.org/pdf/1905.09433.pdf). arXiv preprint arXiv:1905.09433, 2019.

This implementation achieves around 0.8245 AUC and 0.4224 Logloss on the MovieLens-1M dataset. The GPU version gets 0.8231 AUC and 0.4175 Logloss.

## Requirements

- Tensorflow 1.15
- Python 3.7
- Ascend 910 (Image path: swr.cn-north-4.myhuaweicloud.com/ascend-share/3.3.0.alpha002_tensorflow-ascend910-cp37-euleros2.8-aarch64-training:1.15.0-2.0.12_0412)

## Usage

```
python3 train.py \
	--data_url PATH_TO_DATA \
	--train_url PATH_TO_OUTPUT \
	--epoch 3 \
	--embedding_size 32 \
	--batch_size 1024 \
	--learning_rate 0.01 \
	--field_size 7
```