The documentation here offers a good starting point for demonstrating how to use the Mistral's Python client to for chat completion tasks and to generate embeddings.
It could be improved by adding a full-code, annotated tutorial for each task. This could be incorporated into the documentation and each tutorial could also have an accompanying Jupyter or Colab notebook. I'm envisioning something similar to this Argilla tutorial, Few shot text classification with active learning using small-text and SetFit.
For generating embeddings, the tutorial could detail how to generate embeddings for an entire open dataset and perhaps add how to use them in a downstream task, like text classification. For the chat completion, the tutorial could be a bit more explanatory (rather than practical) and could resemble something similar to this notebook from OpenAI.
I would be more than happy to pick this up myself. To get started, I would just need access to Mistral's Platform and API (I'm currently on the waitlist).
The documentation here offers a good starting point for demonstrating how to use the Mistral's Python client to for chat completion tasks and to generate embeddings.
It could be improved by adding a full-code, annotated tutorial for each task. This could be incorporated into the documentation and each tutorial could also have an accompanying Jupyter or Colab notebook. I'm envisioning something similar to this Argilla tutorial, Few shot text classification with active learning using small-text and SetFit.
For generating embeddings, the tutorial could detail how to generate embeddings for an entire open dataset and perhaps add how to use them in a downstream task, like text classification. For the chat completion, the tutorial could be a bit more explanatory (rather than practical) and could resemble something similar to this notebook from OpenAI.
I would be more than happy to pick this up myself. To get started, I would just need access to Mistral's Platform and API (I'm currently on the waitlist).