Skip to content

OpenAPI spec has wrong examples #754

@Kablys

Description

@Kablys

System Info

Info
docker run -p 8080:80 ghcr.io/huggingface/text-embeddings-inference:cpu-1.8
curl 127.0.0.1:8080/info | jq { "model_id": "BAAI/bge-large-en-v1.5", "model_sha": null, "model_dtype": "float32", "model_type": { "embedding": { "pooling": "cls" } }, "max_concurrent_requests": 512, "max_input_length": 512, "max_batch_tokens": 16384, "max_batch_requests": 8, "max_client_batch_size": 32, "auto_truncate": false, "tokenization_workers": 16, "version": "1.8.3", "sha": "78502d8e61223d2c73d4bb7aeaea46787e90d596", "docker_label": "sha-78502d8" }

Information

  • Docker
  • The CLI directly

Tasks

  • An officially supported command
  • My own modifications

Reproduction

  1. Open http://localhost:8080/docs/#/Text%20Embeddings%20Inference/embed
  2. Press "Try it out" and get an example:
{
  "dimensions": null,
  "inputs": "string",
  "normalize": true,
  "prompt_name": "null",
  "truncate": false,
  "truncation_direction": "right"
}
  1. "Execute"
  2. Get error code 422:
Failed to deserialize the JSON body into the target type: truncation_direction: unknown variant `right`, expected `Left` or `Right` at line 7 column 33

Expected behavior

Example should have correct values for prompt_name and truncation_direction:

{
  "dimensions": null,
  "inputs": "string",
  "normalize": true,
  "prompt_name": null,
  "truncate": false,
  "truncation_direction": "Right"
}

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions