I have been looking forward to use lightweighted LLM's for a project to convert human language to SQL queries for a database. To do this, I am reading tutorials and official docs on HuggingFace smolagents and how to use DeepSeek R1 Distill. However, I have a really poor GPU that cannot just simply handle any LLM locally. Therefore, I have seeking for a way to use a Cloud Endpoint instead of using it locally on my machine. I have found that smolagents has a wrapper for Hugging Face Inference API, but I am not sure if it would allow me to use a model like R1 Distill on some regular basis. I would appreciate if someone could explain to me if Hugging Face Inference API is suited for my project or if there are any requirements to meet, such as pricing? I am aware that this could be a very generic question and I could be encouraged to just read more on documentation and do some playgrounding, but I'd like to hace a simplified review so that I can then dive better into more docs myself. Thank you very much for your understanding and time. Continue reading...
