Allora Worker x Reputer as modify with compatible CPU mode, but you can setup with your handle!!!

1. Cloning code from this repository

git clone https://github.com/arcxteam/allora-cpu.git
cd allora-cpu

2. Install Docker and necessary libraries (optional)

chmod +x init.sh
./init.sh

3. Setup Preparation

• Run the worker - directory
• Setting wallet & phrase or manual config.json
• Setting docker-compose.yaml & docker-compose-reputer.yaml =>> Coingecko API=CG-xxxxx

cd allora-cpu/allora-node

chmod +x ./init.config.sh
./init.config.sh "wallet_name" "mnemonic" "coingecko_api_key"

• example: chmod +x ./init.config.sh ./init.config.sh "bangsat-kau" "kalo pake model sendiri pasti dapat point xxx xxx" "CG-xxxxxx"

4. Upgrade & Build Your Worker

docker compose pull

docker compose up --build -d 

• Check logs as to running

docker compose logs -f 

---------------------------------------- Congrats Your Setup is Completed -------------------------------------

1. Now if you want to have your own unique model

• Run command below!!
• Buidl with the train_models.py file
• To edit for the best your 'own model' and run
• Save the model (ctrl X + Y) save ENTER

nano train_models.py 

2. How to train the model?

• Run command below!!

chmod +x ./start-train-model.sh
./start-train-model.sh

3. How to check train the model?

• Run docker command!!
• Copied of CONTAINER ID
• docker logs -f xxxxxxxxxx
• Enter and check the logs

sudo docker ps

docker logs -f 

4. Training the model in ranges 30-60 Minutes!! plz keep until success

• Check your IMAGES run as 'allora-train-model:1.0.0'
• Check logs for training horizon symbol
• 10 Minutues & 20 Minutes & 1440 Minutes
• The containerID with train will be exited auto after complete

Thanks to hiephtdev & 0xtnpxsgt