Model Trained Using AutoTrain

Problem type: Entity Extraction
Model ID: 1676759313
CO2 Emissions (in grams): 0.0110

Validation Metrics

Loss: 0.097
Accuracy: 0.973
Precision: 0.903
Recall: 0.921
F1: 0.912

Usage

You can use cURL to access this model:

$ curl -X POST -H "Authorization: Bearer YOUR_API_KEY" -H "Content-Type: application/json" -d '{"inputs": "I love AutoTrain"}' https://api-inference.huggingface.co/models/hemangjoshi37a/autotrain-stocks-ner-2000-sample-test-1676759313

Or Python API:

from transformers import AutoModelForTokenClassification, AutoTokenizer

model = AutoModelForTokenClassification.from_pretrained("hemangjoshi37a/autotrain-stocks-ner-2000-sample-test-1676759313", use_auth_token=True)

tokenizer = AutoTokenizer.from_pretrained("hemangjoshi37a/autotrain-stocks-ner-2000-sample-test-1676759313", use_auth_token=True)

inputs = tokenizer("I love AutoTrain", return_tensors="pt")

outputs = model(**inputs)

GitHub Link to this project : Telegram Trade Msg Backtest ML

Need custom model for your application? : Place a order on hjLabs.in : Custom Token Classification or Named Entity Recognition (NER) model as in Natural Language Processing (NLP) Machine Learning

What this repository contains? :

Label data using LabelStudio NER(Named Entity Recognition or Token Classification) tool. convert to
Convert LabelStudio CSV or JSON to HuggingFace-autoTrain dataset conversion script
Train NER model on Hugginface-autoTrain.
Use Hugginface-autoTrain model to predict labels on new data in LabelStudio using LabelStudio-ML-Backend.
Define python function to predict labels using Hugginface-autoTrain model.
Only label new data from newly predicted-labels-dataset that has falsified labels.
Backtest Truely labelled dataset against real historical data of the stock using zerodha kiteconnect and jugaad_trader.
Evaluate total gained percentage since inception summation-wise and compounded and plot.
Listen to telegram channel for new LIVE messages using telegram API for algotrading.
Serve the app as flask web API for web request and respond to it as labelled tokens.
Outperforming or underperforming results of the telegram channel tips against exchange index by percentage.