This assumes you have a running yagna and YAGNA_APPKEY variable is set.
# Install packages
pip3 install jupyterlab golem-kernel
# Add golem-kernel to the list of known kernels
python3 -m golem_kernel install
# Start jupyter in a browser. Golem kernel should be available.
jupyter-lab
Check the demo.ipynb notebook.
Shows the amount of tokens on the testnet and mainnet.
Get some funds, if you need them. You can only get testnet tokens this way.
%fund rinkeby
Define a budget for computations.
%budget rinkeby 1
Connect to a provider.
You can specify:
- Minimal amount of RAM, disk and cores
- Provider selection strategy (available strategies: "bestprice" - default, cheapest provider, "random" - random provider)
- Image hash - creating Golem images. This
defaults to an image built from the
providerdirectory. Your image must include everything specified inprovider/Dockerfile(exceptnumpy).
Sample usage:
%connect mem>4 disk>100 cores>2 strategy=random image_hash=5389c01c128f94f14653bc0b56822c22b4b3987737ef8f3c0ac61946
End work (on the current provider, you can connect to another one later) and pay.