Docs/examples nv ai foundation models #396
Conversation
Signed-off-by: Patricia Pampanelli <[email protected]>
patriciapampanelli
added
the
documentation
There was a problem hiding this comment.
The README for the bot is good 👍 .
However, to avoid duplicating the ABC bot config, you should restructure it as a user guide in the /docs/user_guides/llm/nvidia_ai_endpoints folder.
The user guide should be written in a notebook and then coverted to .md (https://github.com/NVIDIA/NeMo-Guardrails/blob/develop/CONTRIBUTING.md#jupyter-notebook-documentation).
The user guide should copy the abc configuration and only change the engine/model part and show that it works (e.g., similar to how this guide uses the step 6 from the getting started guide and makes changes: https://github.com/NVIDIA/NeMo-Guardrails/tree/develop/docs/user_guides/jailbreak_detection_heuristics).
Last but not least, if enough time, we should also evaluate the llama2_70b and playground_mixtral_8x7b models. See #387 for more context.
Remove nvidia ai endpoints reference Signed-off-by: Patricia Pampanelli <[email protected]>
| from nemoguardrails.llm.providers import register_llm_provider | ||
| from langchain_nvidia_ai_endpoints import ChatNVIDIA | ||
|
|
||
| register_llm_provider("nvidia_ai_endpoints", ChatNVIDIA) |
| "role": "context", | ||
| "content": { | ||
| "relevant_chunks": """ | ||
| Employees are eligible for the following time off: |
There was a problem hiding this comment.
The ABC bot has the knowledge base configured, so providing explicit relevant chunks should not be needed. It should work without.
|
Thanks @patriciapampanelli ! Looks good. I've made a few tweaks to make this similar to the Vertex AI notebook. We'll do the evaluation in a future PR. |
No description provided.