Fixes descriptions in the Inference APIs #5566
Conversation
|
Following you can find the validation changes against the target branch for the APIs.
You can validate these APIs yourself by using the |
| * Applies only to the `sparse_embedding` and `text_embedding` task types. | ||
| * Not applicable to the `rerank`, `completion`, or `chat_completion` task types. |
There was a problem hiding this comment.
These generic messages might be confusing, because for example Cohere only supports the following task types:
- completion
- rerank
- text_embedding
So consider deleting sparse_embedding and chat_completion here. I'd inspect all of these services to avoid confusing users by mention applicabilities that aren't available
There was a problem hiding this comment.
You’re absolutely right. Thinking it through, that also means chunking_settings isn’t applicable for certain inference endpoints, even though now it appears for all of them. Could this be a bug, @davidkyle?
For example, the Create an Anthropic inference endpoint API supports only one task type (completion) so chunking doesn’t apply there.
There was a problem hiding this comment.
Yes chunking_settings only applies to the text_embedding and sparse_embedding task types
leemthompo
dismissed
their stale review
Sorry was on auto-pilot and I thought this was a backport 🙈
I noticed a small nit that might confuse users
Co-authored-by: Liam Thompson <leemthompo@gmail.com>
| completion, | ||
| rerank, | ||
| space_embedding, | ||
| sparse_embedding, |
| * | ||
| * NOTE: When creating an inference endpoint, the associated machine learning model is automatically deployed if it is not | ||
| * already running. After creating the endpoint, wait for the model deployment to complete before using it. You can verify | ||
| * the deployment status by using the Get trained model statistics API. In the response, look for "state": "fully_allocated" | ||
| * and ensure the "allocation_count" matches the "target_allocation_count". Avoid creating multiple endpoints for the same | ||
| * model unless required, as each endpoint consumes significant resources. | ||
| * |
There was a problem hiding this comment.
This text is in PutElasticsearchRequest.ts and PutElserRequest.ts and specific to those inference services we don't need it here too
| * | |
| * NOTE: When creating an inference endpoint, the associated machine learning model is automatically deployed if it is not | |
| * already running. After creating the endpoint, wait for the model deployment to complete before using it. You can verify | |
| * the deployment status by using the Get trained model statistics API. In the response, look for "state": "fully_allocated" | |
| * and ensure the "allocation_count" matches the "target_allocation_count". Avoid creating multiple endpoints for the same | |
| * model unless required, as each endpoint consumes significant resources. | |
| * |
| * Applies only to the `sparse_embedding` and `text_embedding` task types. | ||
| * Not applicable to the `rerank`, `completion`, or `chat_completion` task types. |
There was a problem hiding this comment.
Yes chunking_settings only applies to the text_embedding and sparse_embedding task types
This PR fixes description errors discovered during the https://github.com/elastic/docs-content-internal/issues/280 of the 8.x and 9.x API documentation.
Closes #5508, #5509, #5510, #5511, #5512