How to create Hosted Inference API

I finished training my model](nloc2578/QAG · Hugging Face) (without push_to_hub()) on gg colab and all my result and files (config.json, pytorch_model.bin, tokenizer.json, …) are saved in Drive. Now I want to create Hosted Inference API (HIA) for Spaces later usage, and I see that I need to upload config.json, pytorch_model.bin to the Hub. I did it but I still dont have the HIA.

Please help!!!