Open ali-nazir opened 5 years ago
First, this is another repo. Bert-as-a-service is developed by someone else, so you should post this issue there.
Regardless of that, it seems that you are trying to make client requests on the server side.
You can use the same machine for both server and client, but of course, not the same process.
Open another terminal window and use bert client to make the requests.
I am a newbie and studying on BERT. I tried to run the example as mentioned on the page https://bert-as-service.readthedocs.io/en/latest/section/get-start.html but whenever I give the command the server connects but nothing appears, I wanted to run the example as mentioned in the page but nothing appears on my screen it only says that Ready and listening! but I could not convert my text into vectors, as it appears in first clip, https://raw.githubusercontent.com/hanxiao/bert-as-service/master/.github/demo.gif I can write something but nothing happens. I use this command for pre-trained multi cased
python3 /home/ali/.local/bin/bert-serving-start -model_dir ./multi_cased_L-12_H-768_A-12/ -num_worker=4
on my ubuntu 18.04.System: HP omen 17 Intel core i7, 8750H Nvidia GTX1050 4GB.
whenever I run the command this happens![Screenshot from 2019-03-21 21-19-05](https://user-images.githubusercontent.com/49047042/55203047-4686e500-51ca-11e9-818c-0333bf32f351.png)
at the end of the screenshot I used this code (example), as mentioned in the 2nd link(gif)
from bert_serving.client import BertClient bc = BertClient() bc.encode(['First do it', 'then do it right', 'then do it better'])
rephrasing the question: I use this command to connect to Bert server
ali@Omen:~/Desktop/Work$ python3 /home/ali/.local/bin/bert-serving-start -model_dir ./multi_cased_L-12_H-768_A-12/ -num_worker=4
and I get this as output (as mentioned in screenshot)
usage: /home/ali/.local/bin/bert-serving-start -model_dir ./multi_cased_L-12_H-768_A-12/ -num_worker=4 ARG VALUE
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 = ./multi_cased_L-12_H-768_A-12/ num_worker = 4 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:[i:i: 66]:freeze, optimize and export graph, could take a while... I:GRAPHOPT:[gra:opt: 52]:model config: ./multi_cased_L-12_H-768_A-12/bert_config.json I:GRAPHOPT:[gra:opt: 55]:checkpoint: ./multi_cased_L-12_H-768_A-12/bert_model.ckpt I:GRAPHOPT:[gra:opt: 59]:build graph...
WARNING: The TensorFlow contrib module will not be included in TensorFlow 2.0. For more information, please see:
I:GRAPHOPT:[gra:opt:128]:load parameters from checkpoint... I:GRAPHOPT:[gra:opt:132]:optimize... I:GRAPHOPT:[gra:opt:140]:freeze... I:GRAPHOPT:[gra:opt:145]:write graph to a tmp file: /tmp/tmp74_auga3 I:VENTILATOR:[i:i: 74]:optimized graph is stored at: /tmp/tmp74_auga3 I:VENTILATOR:[i:_ru:118]:bind all sockets I:VENTILATOR:[__i:_ru:122]:open 8 ventilator-worker sockets I:VENTILATOR:[i:_ru:125]:start the sink I:SINK:[i:_ru:284]:ready I:VENTILATOR:[__i:_ge:202]:get devices W:VENTILATOR:[i:_ge:217]:only 1 out of 1 GPU(s) is available/free, but "-num_worker=4" W:VENTILATOR:[i:_ge:219]:multiple workers will be allocated to one GPU, may not scale well and may raise out-of-memory I:VENTILATOR:[__i:_ge:235]:device map: worker 0 -> gpu 0 worker 1 -> gpu 0 worker 2 -> gpu 0 worker 3 -> gpu 0 I:WORKER-3:[i:_ru:492]:use device gpu: 0, load graph from /tmp/tmp74_auga3 I:WORKER-0:[i:_ru:492]:use device gpu: 0, load graph from /tmp/tmp74_auga3 I:WORKER-2:[i:_ru:492]:use device gpu: 0, load graph from /tmp/tmp74_auga3 I:WORKER-1:[i:_ru:492]:use device gpu: 0, load graph from /tmp/tmp74_auga3 I:WORKER-2:[i:gen:520]:ready and listening! I:WORKER-1:[i:gen:520]:ready and listening! I:WORKER-3:[i:gen:520]:ready and listening! I:WORKER-0:[__i:gen:520]:ready and listening!