đĻAdd Ollama endpoint
Last updated
Last updated
Ollama version: 0.1.19 or later
Running at least one model under Ollama
Go to MindMac -> Settings... -> Account or press â + , to open Account Setting.
Select Ollama as provider
Enter a name
If you don't change any Ollama configuration, just use http://localhost:11434/api/chat as URL
Enter a random text in API Key text area
Click Save to finish
Currently Ollama does not provide exact context length for each model, so if you want to control max tokens (or context length) parameter, you might need to enter that value manually for every Ollama model. Follow below steps to do so.
Go to MindMac -> Settings... -> Account or press â + , to open Account Setting.
Click on the brain icon to show Ollama model list.
Click on the edit button.
Enter max tokens value for that model and click Save to finish.
To get max tokens value for each model, you can find in configuration file (config.json for example) of that model on Hugging Face. For instance, with the model Mixtral-8x7B-Instruct-v0.1 , you can go to https://huggingface.co/mistralai/Mixtral-8x7B-Instruct-v0.1/blob/main/config.json and grab the value max_position_embeddings as its context length. Kindly be aware that there is no standard specifying which key determines the context length of a model in the model configuration file.
