to get started mining, first start your model endpoint.
if you are using runpod, please create the model instance first.
you can do this by using a container image 'ghcr.io/huggingface/text-generation-inference:1.1.0' and then running the following command in the 'docker command': --model-id YOUR_HF_MODEL
for your port, specify 8080
when it is done spinning up, navigate to 'pods' tab and click 'connect'. You will see a card that has your port. click this, and that url is what you will use in --sybil.api_url
if you are not using runpod, then install pm2 and run the following command:
cd miners/sybil/tgi
pm2 start model_endpoint.sh
then once this is running, start the miner. it will automatically connect at 0.0.0.0:8080
there is no more need to specify /generate in the api_url! this is done automatically.
Miners
to get started mining, first start your model endpoint.
if you are using runpod, please create the model instance first.
you can do this by using a container image 'ghcr.io/huggingface/text-generation-inference:1.1.0' and then running the following command in the 'docker command': --model-id YOUR_HF_MODEL
for your port, specify 8080
when it is done spinning up, navigate to 'pods' tab and click 'connect'. You will see a card that has your port. click this, and that url is what you will use in --sybil.api_url
if you are not using runpod, then install pm2 and run the following command:
there is no more need to specify /generate in the api_url! this is done automatically.