---
title: Task hackathon
tag: hackathon irifellow
---

Task
===
## Architecture
- User - Web (frontend) - Server (App engine) - ML engine
## Phase 1
### Export model and deploy to mlengine
- [source](https://medium.com/searce/creating-deep-learning-models-training-and-deploying-it-on-google-cloud-ml-engine-using-9a4ed6a84076)

#ENVIRONMENT VARIABLES
$ export MODEL_NAME kaggle_housing_price_prediction
$ export MODEL_PATH=gs://cloud-ml-job-
  bucket/export/exporter/1545905371
#CREATE MODEL
$ gcloud ml-engine models create $MODEL_NAME
$ gcloud ml-engine versions create "version_1" --model $MODEL_NAME -
  -origin $MODEL_PATH

### Web app:
- [ ] Find template (Bao)
- [ ] upload image
- [ ] Send image to mlengine (Kien)
- [ ] Have database to store price and label
- [ ] Calculate price and show item on screen
- [ ] Preprocess return result (Filter out none make sense detected item) [coco label](https://raw.githubusercontent.com/amikelive/coco-labels/master/coco-labels-paper.txt)

import numpy as np
from googleapiclient import discovery
from flask import Flask, jsonify, render_template, request
import os
os.environ['GOOGLE_APPLICATION_CREDENTIALS'] = './cloudml-b436315c38f2.json'
PROJECT = "gentle-voltage-235901"
MODEL   = "mnist"
INPUT_NODE = "flatten_input"
VERSION = "v1"

def predict_json(project, model, instances, version=None):
    """Send json data to a deployed model for prediction.

    Args:
        project (str): project where the Cloud ML Engine Model is deployed.
        model (str): model name.
        instances ([Mapping[str: Any]]): Keys should be the names of Tensors
            your deployed model expects as inputs. Values should be datatypes
            convertible to Tensors, or (potentially nested) lists of datatypes
            convertible to tensors.
        version: str, version of the model to target.
    Returns:
        Mapping[str: any]: dictionary of prediction results defined by the
            model.
    """
    # Create the ML Engine service object.
    # To authenticate set the environment variable
    # GOOGLE_APPLICATION_CREDENTIALS=<path_to_service_account_file>
    # CREDENTIALS = app_engine.Credentials()
    # service = discovery.build('ml', 'v1', credentials=CREDENTIALS)
    service = discovery.build('ml', 'v1')
    name = 'projects/{}/models/{}'.format(project, model)

    if version is not None:
        name += '/versions/{}'.format(version)
    response = service.projects().predict(
        name=name,
        body={'instances': instances}
    ).execute()

    if 'error' in response:
        raise RuntimeError(response['error'])

    return response['predictions']

## Phase 2
### Model:
#### Train on D2S dataset
- [ ] https://www.mvtec.com/company/research/datasets/mvtec-d2s/
- [ ] Train fast r-cnn with backbone [vgg, resnet, inceptionet v2 ...]
- [ ] Collect data of vietnamese product
- [ ] Crawl Price of produce

### Web app:
- [] Capture from webcam
##
## Bonus:
- [ ] Know trade mark of each product (exp: diffrent price base on brand: bottle of water - aquafina)