Spaces:
Running
Running
| # Amazon Web Services (AWS) | |
| | Feature | Available | | |
| | --------------------------- | --------- | | |
| | [Tools](../tools) | No | | |
| | [Multimodal](../multimodal) | No | | |
| You may specify your Amazon SageMaker instance as an endpoint for Chat UI: | |
| ```ini | |
| MODELS=`[{ | |
| "name": "your-model", | |
| "displayName": "Your Model", | |
| "description": "Your description", | |
| "parameters": { | |
| "max_new_tokens": 4096 | |
| }, | |
| "endpoints": [ | |
| { | |
| "type" : "aws", | |
| "service" : "sagemaker" | |
| "url": "", | |
| "accessKey": "", | |
| "secretKey" : "", | |
| "sessionToken": "", | |
| "region": "", | |
| "weight": 1 | |
| } | |
| ] | |
| }]` | |
| ``` | |
| You can also set `"service": "lambda"` to use a lambda instance. | |
| You can get the `accessKey` and `secretKey` from your AWS user, under programmatic access. | |