Hi Albert, awesome work with this package and love the tidymodels flavor! Support for Databricks Foundation Models would be great. This would include the following serving endpoints:
DBRX @ databricks-dbrx-instruct - This is Mosaic AI's chat LLM rebranded post-Databricks acquisition
Mixtral 8x7B @ databricks-mixtral-8x7b-instruct
Meta Lllama 3 @ databricks-meta-llama-3-70b-instruct
Here are some additional resources that might help you:
chattr repo: I requested a similar chattr feature in #97 and their team updated / merged into main from #99
Using curl I was successful hitting /serving-endpoints/{name}/invocations
Databricks docs: While I think the focus of tidychatmodels is specifically chat LLMs, there are a few embedding model endpoints on Databricks and here's more info on those models for reference: Supported Models for Pay-per-Token
Hi Albert, awesome work with this package and love the
tidymodelsflavor! Support for Databricks Foundation Models would be great. This would include the following serving endpoints:databricks-dbrx-instruct- This is Mosaic AI's chat LLM rebranded post-Databricks acquisitiondatabricks-mixtral-8x7b-instructdatabricks-meta-llama-3-70b-instructHere are some additional resources that might help you:
chattrfeature in #97 and their team updated / merged into main from #99curlI was successful hitting/serving-endpoints/{name}/invocationstidychatmodelsis specifically chat LLMs, there are a few embedding model endpoints on Databricks and here's more info on those models for reference: Supported Models for Pay-per-Token