Hello everyone.
Im trying to load others trained nlu models during conversations, i have found the API page that provide the “model” endpoint with PUT method.
But i found your error tracking quite hard to understand.
Application side i perform curl command:
curl -X PUT http://localhost:5005/model -H 'Content-type: application/json' -d '{"model_file":"/models/prestataire.tar.gz","model_server":{"url":"string","params":{ },"headers":{ },"basic_auth":{ },"token":"string","token_name":"string","wait_time_between_pulls":0},"remote_storage":"None"}'
response is:
{"version":"1.3.8","status":"failure","message":"An unexpected error occurred. Error: string","reason":"LoadingError","details":{},"help":null,"code":500}
Ok error 500. no probleme.
But Rasa side i have this:
rasa_1 | 2019-10-16 14:55:23 ERROR rasa.core.agent - Could not load model due to string.
This is this line i fully not understand. can you tell me what is wrong.
Model folder:
$ pwd
<....>/dockerrasa/models
$ ls
default.tar.gz prestataire.tar.gz
The cause of this error is “remove_storage” field is an enum and None isn’t in. but i don’t want to use remote storage or model server. So i perform this line:
curl -X PUT http://localhost:5005/model -H 'Content-type: application/json' -d '{"model_file":"/models/prestataire.tar.gz"}'
Response is:
{"version":"1.3.8","status":"failure","message":"Agent with name '\/models\/prestataire.tar.gz' could not be loaded.","reason":"BadRequest","details":{"parameter":"model","in":"query"},"help":null,"code":400}
Edit: i use docker compose
Both models are trainned using docker command:
docker run -v $(pwd):/app rasa/rasa:latest-full train --domain domain.yml --data data --out models --config config.yml --fixed-model-name default
docker run -v $(pwd):/app rasa/rasa:latest-full train --domain domain.yml --data data_prestataire --out models --config config.yml --fixed-model-name prestataire
$ ls data_prestataire
nlu.md