https://arxiv.org/abs/1809.03672
This repo introduce how to run Deep Interest Evoluation Network (DIEN) with OpenVINO backend.
Please refer to orignal repo for details: https://github.com/alibaba/ai-matrix/tree/master/macro_benchmark/DIEN_TF2
You can get the data from amazon website and process it using the script
sh prepare_data.sh
Because getting and processing the data is time consuming,so we had processed it and upload it for you. You can unzip it to use directly.
sh prepare_dataset.sh
When you see the files below, you can do the next work.
- cat_voc.pkl
- mid_voc.pkl
- uid_voc.pkl
- local_train_splitByUser
- local_test_splitByUser
- reviews-info
- item-info
pip install openvino openvino-dev[tensorflow]
mo --input_meta_graph dnn_best_model_trained/ckpt_noshuffDIEN3.meta \
--input "Inputs/mid_his_batch_ph[-1,-1],Inputs/cat_his_batch_ph[-1,-1],Inputs/uid_batch_ph[-1],Inputs/mid_batch_ph[-1],Inputs/cat_batch_ph[-1],Inputs/mask[-1,-1],Inputs/seq_len_ph[-1]" \
--output dien/fcn/Softmax --model_name DIEN -o openvino/FP32 --compress_to_fp16=False
./infer.sh tensorflow
./infer.sh openvino f32
./infer.sh openvino bf16
Please note, Xeon native supports BF16 infer precision since 4th Generation Intel® Xeon® Scalable Processors. Run BF16 on legacy Xeon platform may lead to performance degradation.