bert-as-service 用 BERT 作為句子編碼器, 並通過 ZeroMQ 服務託管, 只需兩行程式碼就可以將句子對映成固定長度的向量表示; ## 準備 windows10 + python3.5 + tensorflow1.2.1 ## 安裝流程 1. 安裝 tensorflow, [參考](https://tensorflow.google.cn/install/pip#windows_1) 2. 安裝 bert-as-service bert-as-service, 依賴於 **python≥3.5** AND **tensorflow≥1.10**; ``` pip install bert-serving-server pip instlal bert-serving-client ``` 3. 下載中文 bert 預訓練的模型 | [BERT-Base, Uncased](https://storage.googleapis.com/bert_models/2018_10_18/uncased_L-12_H-768_A-12.zip) | 12-layer, 768-hidden, 12-heads, 110M parameters | | ------------------------------------------------------------ | ------------------------------------------------------------ | | [BERT-Large, Uncased](https://storage.googleapis.com/bert_models/2018_10_18/uncased_L-24_H-1024_A-16.zip) | 24-layer, 1024-hidden, 16-heads, 340M parameters | | [BERT-Base, Cased](https://storage.googleapis.com/bert_models/2018_10_18/cased_L-12_H-768_A-12.zip) | 12-layer, 768-hidden, 12-heads , 110M parameters | | [BERT-Large, Cased](https://storage.googleapis.com/bert_models/2018_10_18/cased_L-24_H-1024_A-16.zip) | 24-layer, 1024-hidden, 16-heads, 340M parameters | | [BERT-Base, Multilingual Cased (New)](https://storage.googleapis.com/bert_models/2018_11_23/multi_cased_L-12_H-768_A-12.zip) | 104 languages, 12-layer, 768-hidden, 12-heads, 110M parameters | | [BERT-Base, Multilingual Cased (Old)](https://storage.googleapis.com/bert_models/2018_11_03/multilingual_L-12_H-768_A-12.zip) | 102 languages, 12-layer, 768-hidden, 12-heads, 110M parameters | | [BERT-Base, Chinese](https://storage.googleapis.com/bert_models/2018_11_03/chinese_L-12_H-768_A-12.zip) | Chinese Simplified and Traditional, 12-layer, 768-hidden, 12-heads, 110M parameters | 4. 啟動 bert-as-serving 服務 ```shell bert-serving-start -model_dir /tmp/english_L-12_H-768_A-12/ -num_worker=2 //模型路徑自改 usage: xxxx\Anaconda3\envs\py35\Scripts\bert-serving-start -model_dir D:\env\bert\chinese_L-12_H-768_A-12 -num_worker=2 ARG VALUE __________________________________________________ ckpt_name = bert_model.ckpt config_name = bert_config.json cors = * cpu = False device_map = [] do_lower_case = True fixed_embed_length = False fp16 = False gpu_memory_fraction = 0.5 graph_tmp_dir = None http_max_connect = 10 http_port = None mask_cls_sep = False max_batch_size = 256 max_seq_len = 25 model_dir = D:\env\bert\chinese_L-12_H-768_A-12 no_position_embeddings = False no_special_token = False num_worker = 2 pooling_layer = [-2] pooling_strategy = REDUCE_MEAN port = 5555 port_out = 5556 prefetch_size = 10 priority_batch_size = 16 show_tokens_to_client = False tuned_model_dir = None verbose = False xla = False I:VENTILATOR:freeze, optimize and export graph, could take a while... I:GRAPHOPT:model config: D:\env\bert\chinese_L-12_H-768_A-12\bert_config.json I:GRAPHOPT:checkpoint: D:\env\bert\chinese_L-12_H-768_A-12\bert_model.ckpt I:GRAPHOPT:build graph... I:GRAPHOPT:load parameters from checkpoint... I:GRAPHOPT:optimize... I:GRAPHOPT:freeze... I:GRAPHOPT:write graph to a tmp file: C:\Users\Memento\AppData\Local\Temp\tmpo07002um I:VENTILATOR:bind all sockets I:VENTILATOR:open 8 ventilator-worker sockets I:VENTILATOR:start the sink I:SINK:ready I:VENTILATOR:get devices W:VENTILATOR:no GPU available, fall back to CPU I:VENTILATOR:device map: worker 0 -> cpu worker 1 -> cpu I:WORKER-0:use device cpu, load graph from C:\Users\Memento\AppData\Local\Temp\tmpo07002um I:WORKER-1:use device cpu, load graph from C:\Users\Memento\AppData\Local\Temp\tmpo07002um I:WORKER-0:ready and listening! I:WORKER-1:ready and listening! I:VENTILATOR:all set, ready to serve request! ``` 5. 用 python 模擬呼叫 bert-as-service 服務 ````python bc = BertClient(ip="localhost", check_version=False, check_length=False) vec = bc.encode(['你好', '你好呀', '我很好']) print(vec) ```` 輸出結果: ``` [[ 0.2894022 -0.13572647 0.07591158 ... -0.14091237 0.54630077 -0.30118054] [ 0.4535432 -0.03180456 0.3459639 ... -0.3121457 0.42606848 -0.50814617] [ 0.6313594 -0.22302179 0.16799903 ... -0.1614125 0.23098437 -0.5840646 ]] ``` ## [亮點](https://github.com/hanxiao/bert-as-service#highlights) -