We should probably focus on one of them for the course. I personally prefer Ollama+Langchain, if ther eis enough time to introduce them properly...
The advantage of HuggingFace is that integration with episode 3 is seamless, as we are still using the transformers library. However it feels very draft-like code.
The advantage of Ollama (through langchain) is that it has a way more intuitive API, and is also more generalizable to advanced use cases, so it makes it easier to explain "next steps" if participants are interested. The problem is yet another library to introduce
We should probably focus on one of them for the course. I personally prefer Ollama+Langchain, if ther eis enough time to introduce them properly...
The advantage of HuggingFace is that integration with episode 3 is seamless, as we are still using the
transformerslibrary. However it feels very draft-like code.The advantage of Ollama (through langchain) is that it has a way more intuitive API, and is also more generalizable to advanced use cases, so it makes it easier to explain "next steps" if participants are interested. The problem is yet another library to introduce