«Ошибка»: «Прогнозирование не удалось: вектор IDF не установлен» с командой почтальона и curl ⇐ Python

[Anonymous]

Код: Выделить всё

from sklearn.pipeline import make_pipeline
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.linear_model import LogisticRegression
import joblib

#Make a pipeline with TFIDF vectorizier and logistic regresssion model

pipeline = make_pipeline(
(TfidfVectorizer()),
LogisticRegression())

#Train the model with pipeline set above

pipeline.fit(train_df['cleaned_sentence'], train_df['category'])

# Save fully fitted pipeline (includes vectorizer + model)
joblib.dump(pipeline, "logistic_model.pkl")
print("Saved: logistic_model.pkl")
< /code>
После сохранения модели создано изображение Docker и развернутые требования, Main.py и Dockerfile для GCP < /p>
# build docker and push to gcp
gcloud builds submit --tag gcr.io/finalprojectwk5/sentiment-classifier

# gcp deploy from google cloud shell sdk
gcloud run deploy sentiment-classifier --image=gcr.io/finalprojectwk5/sentiment-
classifier --platform=managed --region=us-central1 --allow-unauthenticated --
memory=1024Mi

# Curl command with Postman
curl -X POST -H "Content-Type: application/json" -d "{\"text\": \"prefix great service
confusion read crustal clear crass filthy nonsensical see coming\"}" https://sentiment-
classifier-763594834115.us-central1.run.app/predict

Я получаю неудачный прогноз Eror>

Подробнее здесь: https://stackoverflow.com/questions/795 ... d-curl-com