how can i make the bb3 model like flask api or websocket api

[lonnieLuPacvue]

i can run the bb3 model through below command
parlai safe_interactive --model-file zoo:bb3/bb3_3B/model --init-opt gen/r2c2_bb3

But i want make them like flask api or websocket,is there any solutions?
i read the terminal websocket and rewrite the config.yml like

tasks:
default:
onboard_world: MessengerBotChatOnboardWorld
task_world: MessengerBotChatTaskWorld
timeout: 1800
agents_required: 1
task_name: chatbot
world_module: parlai.chat_service.tasks.chatbot.worlds
overworld: MessengerOverworld
max_workers: 30
opt:
debug: True
models:
bb3_3B:
model: transformer/generator
model_file: zoo:bb3/bb3_3B/model
interactive_mode: True
# no_cuda: False
override:
#init_opt: arch/r2c2_base_3B
search_server: 172.28.107.241:8080
memory_decision: always
search_decision: always
additional_args:
page_id: 1

But i can chat with model.But the answer is all wrong.
Help~

[mojtaba-komeili]

Check out our interactive_web script.

[lonnieLuPacvue]

@mojtaba-komeili thanks, it's worked for me.But I find that when I start two client and send the request at the same time.The model deals it one by one.So,how I can do something to make it to be concurrent.

[mojtaba-komeili]

Our code is not optimized for performance in production much. This simple script is just for running small demos and debugging the model. For a simple solution, I recommend running multiple copies (maybe each inside a container). Or you could look into making it async.

[lonnieLuPacvue]

yes, I understand.But i want to now,Does this model can deal something with parallel

[mojtaba-komeili]

We do not provide that here. But we have knowledge of projects that were able to successfully do that.

[lonnieLuPacvue]

ok，thanks