Initial commit. Trained model, Flask web app, Docker container

.gitattributes

@@ -33,3 +33,7 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+models/model filter=lfs diff=lfs merge=lfs -text
+training/checkpoints/model_epoch_1 filter=lfs diff=lfs merge=lfs -text
+training/checkpoints/model_epoch_2 filter=lfs diff=lfs merge=lfs -text
+training/checkpoints/model_epoch_3 filter=lfs diff=lfs merge=lfs -text

Dockerfile

@@ -0,0 +1,17 @@
+FROM arm64v8/python:3.8-slim-buster
+
+WORKDIR /app
+
+COPY requirements.txt /app/
+
+RUN apt-get update && apt-get install -y libstdc++6 \
+    && pip install --trusted-host pypi.python.org -r requirements.txt \
+    && apt-get clean
+
+COPY app /app/app/
+
+COPY models /app/models
+
+EXPOSE 9696
+
+CMD ["python", "app/app.py"]

app/app.py

@@ -0,0 +1,39 @@
+from flask import Flask, jsonify, request
+from utils import predict_single, predict_batch
+
+
+app = Flask(__name__)
+
+
+
+@app.route('/')
+@app.route('/home')
+def status():
+    return jsonify({'status': 'ok'})
+
+
+@app.route('/predict', methods=['POST'])
+def predict():
+    data = request.get_json()
+
+    if 'text' not in data:
+        return jsonify({'error': 'Missing "text" parameter'}), 400
+    
+    tweets = data['text']
+
+    if len(tweets) == 1:
+        response = predict_single(tweets[0])
+    elif len(tweets) > 1:
+        response = predict_batch(tweets)
+    else:
+        return jsonify({'error': 'Zero text strings posted'}), 400
+
+    return jsonify({
+        'inputs': tweets,
+        'predictions': response
+        })
+
+
+
+if __name__ == '__main__':
+    app.run(debug=True, host='0.0.0.0', port=9696)

app/routes.py

@@ -0,0 +1,32 @@
+from flask import Flask, jsonify, request
+from utils import predict_single, predict_batch
+
+app = Flask(__name__)
+
+
+@app.route('/')
+@app.route('/home')
+def status():
+    return jsonify({'status': 'ok'})
+
+
+@app.route('/predict', methods=['POST'])
+def predict():
+    data = request.get_json()
+
+    if 'text' not in data:
+        return jsonify({'error': 'Missing "text" parameter'}), 400
+    
+    tweets = data['text']
+
+    if len(tweets) == 1:
+        response = predict_single(tweets[0])
+    elif len(tweets) > 1:
+        response = predict_batch(tweets)
+    else:
+        return jsonify({'error': 'Zero text strings posted'}), 400
+
+    return jsonify({
+        'inputs': tweets,
+        'predictions': response
+        })

app/utils.py

@@ -0,0 +1,32 @@
+import torch
+import torch.nn.functional as F
+import pickle
+import re
+
+
+model = torch.load("models/model", map_location='cpu')
+tokenizer = torch.load("models/tokenizer")
+
+with open("models/label_dict", 'rb') as file:
+    label_dict = pickle.load(file)
+
+def preprocess_string(tweet: str) -> str:
+    tweet = tweet.lower().strip()
+    tweet = re.sub(r'[^\w\s]', '', tweet)
+    return tweet
+
+def predict_single(tweet: str) -> str:
+    clean_tweet = preprocess_string(tweet)
+    input = tokenizer(clean_tweet, return_tensors='pt', truncation=True)
+    output = model(**input)
+    pred = torch.max(F.softmax(output.logits, dim=-1), dim=-1)[1]
+    pred = pred.data.item()
+    return label_dict[pred]
+
+def predict_batch(tweets):
+    clean_tweets = [preprocess_string(tweet) for tweet in tweets]
+    inputs = tokenizer(clean_tweets, return_tensors='pt', padding=True, truncation=True)
+    outputs = model(**inputs)
+    preds = torch.max(F.softmax(outputs.logits, dim=-1), dim=-1)[1]
+    preds = preds.tolist()
+    return [label_dict[pred] for pred in preds]

models/model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f2f2873ded67e0c26f97306d9ba4d017a6f733692ea4614b16ca043add388976
+size 267884825

requirements.txt

@@ -0,0 +1,4 @@
+--find-links https://download.pytorch.org/whl/torch_stable.html
+torch==1.9.0
+transformers==4.26.1
+flask==2.3.3

serve_predictions.ipynb

@@ -0,0 +1,71 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import requests\n",
+    "import json"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'inputs': [\"That's just so romantic, I guess you're really falling for her!\",\n",
+       "  'Never have I felt so disgusted in my life.',\n",
+       "  'The sun was shining, we had a great lunch and then took a hike - it was the greatest day of my life.',\n",
+       "  'It came out of nowhere, noboy was expecting that.'],\n",
+       " 'predictions': ['love', 'anger', 'joy', 'fear']}"
+      ]
+     },
+     "execution_count": 2,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "response = requests.post(\n",
+    "    url='http://127.0.0.1:9696/predict',\n",
+    "    json={'text': \n",
+    "          [\"That's just so romantic, I guess you're really falling for her!\",\n",
+    "           \"Never have I felt so disgusted in my life.\",\n",
+    "           \"The sun was shining, we had a great lunch and then took a hike - it was the greatest day of my life.\",\n",
+    "           \"It came out of nowhere, noboy was expecting that.\"\n",
+    "           ]}\n",
+    ")\n",
+    "\n",
+    "response = json.loads(response.text)\n",
+    "response"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "hugging",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.9.7"
+  },
+  "orig_nbformat": 4
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

training/checkpoints/model_epoch_1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:68c581923c7e0ce1a9708dab5506d500baaf36225727b40da5a7f89b8f58b528
+size 267892337

training/checkpoints/model_epoch_2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4679e143d9c933638ff9c8c0762ed7a54f20f8bae5fe4c257599d4c872a9a0e2
+size 267892337

training/checkpoints/model_epoch_3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a6dfd452565645d50ba5247202448f9148dc51b1a57583e10af6303988249f51
+size 267892337