https://github.com/LiveHelperChat/intent-deeppavlov
Simple intent recognition using DeepPavlov.ai
https://github.com/LiveHelperChat/intent-deeppavlov
Last synced: about 2 months ago
JSON representation
Simple intent recognition using DeepPavlov.ai
- Host: GitHub
- URL: https://github.com/LiveHelperChat/intent-deeppavlov
- Owner: LiveHelperChat
- Created: 2020-12-16T11:04:02.000Z (over 4 years ago)
- Default Branch: main
- Last Pushed: 2020-12-16T12:40:24.000Z (over 4 years ago)
- Last Synced: 2024-04-13T21:55:06.694Z (about 1 year ago)
- Language: Dockerfile
- Size: 248 KB
- Stars: 1
- Watchers: 4
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
Simple intent detection service based on DeepPavlov.ai.
**This should be used only if you have large amount of training data**
Training is happening on image build event. Configuration should be similar to https://doc.livehelperchat.com/docs/bot/deeppavlov-faq/ except response location should be `0:0` instead of `0:0:0`
* You should edit `Dockerfiles/deep/train/train.json` with your data before building an image.
* See `Dockerfiles/deep/Dockerfile` for getting `dstc2_fastText_model.bin` file
```shell
docker-compose -f docker-compose.yml build
docker-compose -f docker-compose.yml up
```
Run as service once it's build
```shell
docker-compose -f docker-compose.yml up -d
```
Testing
```
curl -X POST "http://localhost:5000/model" -H "accept: application/json" -H "Content-Type: application/json" -d "{\"x\":[\"I want to book place for two\"]}"
```
Response if you did not change train data
```json
[["BookRestaurant",[0.2782818078994751,0.22770161926746368,0.30094414949417114,0.6002139449119568,0.2533802390098572,0.25423356890678406,0.21291431784629822],3]]
```