Query Parameters

• type: (string)

  Example:

  sample

• sortby: (string)

  A comma separated string of field names on which to sort. Add " DESC" after a field name to change to a descending sort.

  Examples:

  two field sort:
  

  owner,name

  descending sort:
  

  owner DESC, name, created_at

Body

Media type: application/json

Type: object

• name: required (string)

  The name of the dataset.

Example:

{
  name: "MyDataset",
}

Media type: example

Type:

{
  name: "MyDataset",
}

Body

Media type: application/json

Type: object

• action: required (delete)

  Indicates the action to be performed on this dataset. Possible actions are

  • delete -- deletes the indicated set of datasets
• dataset_id_list: required (list)

  The dataset_id_list is a list of dataset UUID strings. The list should not contain duplicate ids and must have at least one item in the list.

Example:

{
  action: "delete",
  dataset_id_list: [ "datasetId1", "datasetID2" ]
}

Body

Media type: multipart/form-data

Type: object

Examples:

Content-Disposition: form-data;
[
  name="files"; filename="safety.zip" Content-Type: multipart/form-data
]

<form enctype="multipart/form-data" action="{baseUri}/datasets/import" method="POST">
  Choose zip file to upload:
  <input name="files" type="file" /><br />
  <input type="submit" value="Upload File" />
</form>

{
  "curl -i -X POST -H \"Content-Type": "multipart/form-data\" -F \"files=@safety.zip\" {baseUri}/datasets/import"
}

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Query Parameters

• event_id: (string)

  This query parameter is only relevant when an object store (such as swift) is being used. In this environment, if event_id is specified, an SSE will be generated when the object store completes its delete operation. The SSE will set the event_id to the value supplied with this query parameter.

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Body

Media type: application/json

Type: object

• action: required (string)

  The action field identifies the action to be performed. Possible values are

  • rename -- renames the indicated dataset
  • calculate -- returns the number of files in the the dataset
  • mark_sample -- marks the dataset as a sample
  • unmark_sample -- removes the "sample" mark from the dataset
  • clone -- creates a new dataset by cloning the indicated dataset
  • preprocess -- perform transformation on the dataset. Currently, the only supported transformation is data_augmentation. Note that preprocess is a background operation. Progress can be monitored via the /event endpoint.
  • autolabel -- performs an "autolabel" operation on the images in the dataset.
• data: (string)

  This field only applies to the rename action and is required for that action. It identifies the new name for the dataset

• event_id: (string)

  This field only applies to the clone action and is optional for that action. If specified, SSE events will be generated. The event_id in the SSE will be set to the string "clone_dataset".

• name: (string)

  This field only applies to the clone action and is required for that action. It identifies the name for the dataset to be created.

• detector: (string)

  The detector property only applies to the preprocess action and is required for that action. It identifies the type of preprocessing to perform. Currently the only supported "detector" is data_augmenation.

• parameters: (object)

  The parameters property only applies to the preprocess action and is required for that action when performing data_augmentation preprocessing transformations. It identifies the set of augmentations to perform and the parameters for each of those augmentations that require parameters. The types of augmentations are

  • color -- manipulates color characteristics of the image
  • crop -- performs random cropping of the image
  • noise -- adds random noise to the new images
  • rotation -- randomly rotates the image
  • sharpness -- randomly increases the sharpness of the image
  • flip_horizontal
  • flip_vertical
  • focus_blur -- randomly blurs the focus
  • motion_blur -- randomly blurs motion
• dataset_name: (string)

  This property is only relevant to the preprocess action and is required when performing data_augmentation. It identifies the name of the new dataset that will be created to hold the transformed images. Note that the original images will also be present in the new dataset.

• id_list: (string)

  This property is only relevant to the preprocess action. It identifies the set of files to be transformed. If not present, all files in the dataset will be transformed.

Examples:

{
  "action": "rename",
  "data": "New_dataset"
}

{
  "action": "calculate"
}

{
  "action": "clone",
  "event_id": "yes",
  "name": "Dataset2"
}

{
   "action": "preprocess",
   "detector": "data_augmentation"
   "parameters": {
      "color": {
        "brightness": 50,
        "contrast": 50,
        "hue": 20,
        "saturation": 80,
        "reps": 8
      },
      "crop": {
        "max": 30,
        "reps": 1
      },
      "noise": {
        "max": 30,
        "reps": 4
      },
      "rotation": {
        "max": 150,
        "reps": 5
      },
      "sharpness": {
        "max": 30,
        "reps": 2
      },
      "flip_horizontal": false,
      "flip_vertical": true,
      "focus_blur": {
        "max": 30,
        "reps": 3
      },
      "motion_blur": {
        "max": 50,
        "reps": 3
      }
   },
   "dataset_name": "augmented dataset name",
   "id_list": [ "file-id1", "file-id2", "file-id3" ]
}

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Query Parameters

• sortby: (string)

  A comma separated string of field names on which to sort. Add " DESC" after a field name to change to a descending sort.

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Body

Media type: application/json

Type: object

• name: required (string)

  Identifies the name of the tag (action) to create

Example:

{
  name: "backhand"
}

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• tag_id: required (string)

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Query Parameters

• sortby: (string)

  A comma separated string of field names on which to sort. Add " DESC" after a field name to change to a descending sort.

  Example:

  ?sortyby=name DESC

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Body

Media type: application/json

Type: object

• name: required (string)

  The name of the new category

Example:

{
  name: "myTestCategory"
  }

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• category_id: required (string)

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• category_id: required (string)

Body

Media type: application/json

Type: object

• action: required (string)

  Identifies the action to perform on the category. Supported actions are:

  • rename -- change the name of the indicated category
  • calculate -- determine the number of files belonging to the category
• data: (string)

  This property is only relevant to the rename operation. It is required for rename. It identifies the new name for the category.

Examples:

{
  action: "rename",
  data: "New Category Name"
}

{
  action: "calculate"
}

Media type: properties

Type: object

Media type: examples

Type: object

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Query Parameters

• category_id: (string)

  Filter the files returned such that only files in the given category are returned.

  Example:

  ?category_id=8669b787-15b0-4731-be6c-a76dca76d466

• type: (string)

  Can be used to get only unlabeled files

  Example:

  ?type=unlabeled

• parent_id: (string)

  optional parameter to limit files to those that have the given file ID as a parent. This query parameter can be used to get all frames capture/extracted from the indicated video file.

  Example:

  ?parent_id=9059b787-15b0-4731-be6c-a76dca76d501

• sortby: (string)

  A comma separated string of field names on which to sort. Add " DESC" after a field name to change to a descending sort.

  Examples:

  two field sort:
  

  ?sortby=original_file_name

  descending sort:
  

  ?sortby=file_type DESC, original_file_name

• query: (string)

  A string specifying "generic" query terms. Each query term is a tuple consisting of three items (field name, comparison operator, and value to compare). Each item must be separated by spaces. Multiple terms are allowed but must be separated by commas. Multiple terms are combined in an AND manner. There is currently no support for OR combination.

  The field name can be any field in the File object. User metadata can be included by specifying user_metadata.my-metadata-field-name. The following comparison operators are supported.

  • == (equal)
  • != (not equal)
  • > (greater than)
  • >= (greater than or equal)
  • < (less than)
  • <= (less than or equal)

  String values must be enclosed in either single or double quotes.

  If query is provided, other filter parameters are ignored.

  Examples:

  single term query:
  

  ?query=user_metadata.quantity == 100

  multi-term query:
  

  ?query=user_metadata.count == 6, user_metadata.type == 'lug nuts'

• limit: (integer)

  optional parameter to limit the number of items to return

  Example:

  ?limit=20 // limit the number of elements to return

• skip: (integer)

  optional parameter to skip the number of items before including items in the return set

  Example:

  ?skip=40 // skip 40 elements

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Body

Media type: multipart/form-data

Type: object

• files: required (string)

  File to be uploaded into the dataset. It is possible to upload multiple files in a single post operation.

• category_id: (string)

  The UUID of the category with which the file is to be associated. This property is only relevant if the file is to be used in classification training

• category_name: (string)

  The name of the category with which the file is to be associated. This parameter is only releveant if the file is to be used in classification training.

• labels: (object)

  Optional labels to be saved with the uploaded file. The labels input is the same as the save-labels endpoint. Note that if labels are supplied, only 1 file can be uploaded.

  • labels: required (array of object)

    Items: items

    • name: required (string)

      Label name (aka tag name)

    • confidence: (number)

      If this label was generated by an inference, the confidence score from that inference

    • generate_type: (one of manual, inferred)

      How the label was generated. Default value is manual, but could be inferred if the label was generated by an inference operation.

    • bndbox: required (object)

      Bounding box for the label. Either the bndbox or the segment_polygons property is required. If both are provided, bndbox is ignored. required: false

      • xmin: required (number - minimum: 0)
      • ymin: required (number - minimum: 0)
      • xmax: required (number - minimum: 0)
      • ymax: required (number - minimum: 0)
    • segment_polygons: (array)

      segment_polygons is a 3 dimensional array of points that represent the boundary of the object. The first dimension holds boundaries. It is an array to all for discontingous object identification which can happen when an object is occluded and essentially cut into pieces.

      The 2nd dimension contains arrays of points with each point array (the 3rd dimension) holding an 'x' and 'y' coordinate in that order.

• user_metadata: (object)

  Optional user metadata key/value pairs to be associated with the uploaded file. If multiple files are in the upload list, the metadata is applied to each file. However, metadata is not applied if the file is a zip file.

  The input metadata is in the same format as described by the post to datasets/{DSID}/files/{FILE_ID}/user-metadata endpoint.

Examples:

// Post params
Content-Disposition: form-data;
  [
    name="files"; filename="flower5.jpg" Content-Type: image/jpeg => image binary data
    name="category_id" => eg. "ef01b8b0-8d7b-4078-903d-6f1009b52998"
    name="category_name" => eg. "Tree"
  ]

<form enctype="multipart/form-data" action="{baseUri}/datasets/{id}/files" method="POST">
  Choose jpg/png file to upload:
  <input name="files" type="file" /><br />
  <input name="category_id" type="text" value="ef01b8b0-8d7b-4078-903d-6f1009b52998" /><br />
  <input name="category_name" type="text" value="Tree" /><br />
  <input type="submit" value="Upload File" />
</form>

curl -i -X POST -H "Content-Type: multipart/form-data" -F "files=@mytemp.jpg" -F "category_id=ef01b8b0-8d7b-4078-903d-6f1009b52998 -F "category_name=Tree" baseURI/datasets/DS-UUID/files

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Body

Media type: application/json

Type: object

• action: required (delete)

  Indicates the action to be performed on this file. Possible actions are

  • delete -- deletes the indicated set of files
• file_id_list: required (string)

  The file_id_list is a list of file UUID strings. The list should not contain duplicate ids and must have at least one item in the list.

Example:

{
  action: "delete",
  file_id_list: [ "FileId1", "FileID2", "FileId3" ]
}

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• file_id: required (string)

  Dataset File UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• file_id: required (string)

  Dataset File UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Body

Media type: application/json

Type: object

• action: (one of autolabel, capture, change_category ] required: true description: | Indicates the action to be performed on this file Possible actions are... - `autolabel` -- indicates that automatic labeling should be applied to the indicated video file. The user must provide the ID for the model with which to label. It is the users responsibility to train the model before attempting to auto-label. - `capture` -- indicates that a frame should be capture from the indicated video file. The `capture` action is only valid on video files. - `change_category` -- changes the category to which a file is assigned. method: enum: [ , anual", auto)

  related to capture action to indicate the type of capture to perform. This property is required for the capture action.

• model_id: (string)

  related to the autolabel action. This property identifies the model to be used when performing the auto-label action. This property is required for the autolabel action.

• target_category_id: (string)

  This property only applies to the change_category action and is required if that action is requested. The UUID of the category to change to is the parameter for this property.

• time_interval: (integer)

  relevant to the autolabel action and the capture action when the method is "auto". This property identifies the interval (in whole milliseconds) at which frames should be captured.
  This property is required for the autolabel action.

• time_offset: (integer)

  relevant only to the capture action to identify the time offset (in whole milliseconds) from the beginning of the video at which to capture a single frame.
  This property is required for the capture action.

Examples:

{
  action: "autolabel",,
  time_interval: "5000",
  model_id: "UUID_of_deployed_model"
}

{
  action: "capture",
  time_offset : 1000,
  method: "manual",
}

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• file_id: required (string)

  Dataset File UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• file_id: required (string)

  Dataset File UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Body

Media type: application/json

Type: object

• name: required (string)

  Identifies the name of the action tag to which the label belongs.

• tag_id: required (string)

  ID of the tag identified by the name property.

• start_time: required (integer)

  The time offset into the file at which the action starts. The offset is in whole milliseconds

• end_time: required (integer)

  The time offset into the file at which the action ends. The offset is in whole milliseconds

Example:

{
  "name": "backhand",
  "tag_id": "40a35da0-2f93-48b9-8584-093bf2451974",
  "start_time": 9010,
  "end_time": 9527
  }

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• file_id: required (string)

  Dataset File UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• label_id: required (string)

  Action Label UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• file_id: required (string)

  Dataset File UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• label_id: required (string)

  Action Label UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Body

Media type: application/json

Type: object

• name: (string)

  Identifies the name of the action tag to which the label should be changed.

• start_time: (integer)

  The time offset into the file at which the action starts. The offset is in whole milliseconds

• end_time: (integer)

  The time offset into the file at which the action ends. The offset is in whole milliseconds

Examples:

{
  "name": "backhand"
}

{
  "start_time": 11010,
  "end_time": 11527
}

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• file_id: required (string)

  Dataset File UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• file_id: required (string)

  Dataset File UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Body

Media type: application/json

Type: object

• name: required (string)

  Identifies the name of the tag (class) to which the label belongs

• bndbox: required (object)

  Identifies the upper left and lower right corners of the label's bounding box.

  • xmin: required (integer)

    Upper left corner 'x' coordinate

  • ymin: required (integer)

    Upper left corner 'y' coordinate

  • xmax: required (integer)

    Lower right corner 'x' coordinate

  • ymax: required (integer)

    Lower right corner 'y' coordinate

• segment_polygons: (list)

  segment_polygons is a 3 dimensional array of points that represent the boundary of the object. The first dimension holds boundaries. It is an array to all for discontingous object identification which can happen when an object is occluded and essentially cut into pieces.

  The 2nd dimension contains arrays of points with each point array (the 3rd dimension) holding an 'x' and 'y' coordinate in that order.

Example:

[
  {
    "name": "badge",
    "bndbox": {
      "xmin": 999,
      "ymin": 2009,
      "xmax": 1452,
      "ymax": 2522
    },
    "segment_polygons": [
      [
        [ 1225, 2009 ],
        [ 1452, 2265 ],
        [ 1225, 2522 ],
        [ 999, 2265 ]
      ]
    ]
  },
  {
    "name": "cap",
    "bndbox": {
      "xmin": 3112,
      "ymin": 1226,
      "xmax": 3542,
      "ymax": 1766
    }
  }
]

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• file_id: required (string)

  Dataset File UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Body

Media type: application/json

Type: object

Example:

{
  “location”: “final axle assembly”,
  “shift": “first”
}

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• file_id: required (string)

  Dataset File UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• file_id: required (string)

  Dataset File UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Body

Media type: application/json

Type: array of string

Example:

[
  "key-name",
  "another key to delete"
]

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• file_id: required (string)

  Dataset File UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Body

Media type: application/json

Type: object

• tag_id: required (string)

  Identifies the name of the tag (class) to which the label belongs. This tag (class) must already exist.

• generate_type: (one of manual, auto, suggested, inferred)

  Optional information about how the label was created. If not specified, the default is manual.

• confidence: (number)

  Optional confidence that the label is correct. Range is 0.0 to 1.0.

  Example:

  0.85

• model_id: (string)

  UUID of the model that generated the label. Irrelevant for manual labels.

• bndbox: required (object)

  Identifies the upper left and lower right corners of the label's bounding box.

  • xmin: required (integer)

    Upper left corner 'x' coordinate

  • ymin: required (integer)

    Upper left corner 'y' coordinate

  • xmax: required (integer)

    Lower right corner 'x' coordinate

  • ymax: required (integer)

    Lower right corner 'y' coordinate

• segment_polygons: (list)

  segment_polygons is a 3 dimensional array of points that represent the boundary of the object. The first dimension holds boundaries. It is an array to all for discontingous object identification which can happen when an object is occluded and essentially cut into pieces.

  The 2nd dimension contains arrays of points with each point array (the 3rd dimension) holding an 'x' and 'y' coordinate in that order.

  NOTE: if segment_polygons is present, the bndbox property is ignored and the bounding box information is calculated from the contents of this property.

Examples:

{
  "tag_id": UUID-OF_TAG,
  "bndbox": {
    "xmin": 999,
    "ymin": 2009,
    "xmax": 1452,
    "ymax": 2522
  }
}

{
  "tag_id": UUID-OF_TAG,
  "segment_polygons": [
    [
      [ 1225, 2009 ],
      [ 1452, 2265 ],
      [ 1225, 2522 ],
      [ 999, 2265 ]
    ]
  ]
}

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• file_id: required (string)

  Dataset File UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Query Parameters

• ids: (string)

  A comma separated list of tags UUIDs whose labels are to be retrieved

  Example:

  ?ids=45123c471-15b0-4731-be6c-a76dca76d466,62ab4714-15b0-4731-be6c-a76dca76d420

• tag_ids: (string)

  A comma separated list of object UUIDs to retrieve

  Example:

  ?tag_ids=8669b787-15b0-4731-be6c-a76dca76d466,9769b787-15b0-4731-be6c-a76dca76d420

• model_id: (string)

  UUID of the model that was used to generate the labels to be retrieved.

  Example:

  ?model_id=tag_ids=a146cdb0-15b0-4731-be6c-a76dca76d424

• generate_type: (string)

  The type of method used to create the labels. Possible values are
  

  • manual -- entered by hand, or manually approved auto-label
  • auto -- generated by an auto-label process

  Example:

  ?generate_type=auto

• min_conf: (string)

  The minimum confidence level of labels to be returned. This filter only applies to labels that were generated by an auto-label operation. Value can range from 0.1 to 1.0.

  Example:

  ?min-conf=0.75

• max_conf: (string)

  The maximum confidence level of labels to be returned. This filter only applies to labels that were generated by an auto-label operation. Value can range from 0.1 to 1.0.

  Example:

  ?max-conf=0.49

• limit: (integer)

  optional parameter to limit the number of items to return

  Example:

  ?limit=20

• skip: (integer)

  optional parameter to skip the number of items before including items in the return set

  Example:

  ?skip=40

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• file_id: required (string)

  Dataset File UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Query Parameters

• ids: (string)

  A comma separated list of tags UUIDs whose labels are to be deleted

  Example:

  ?ids=45123c471-15b0-4731-be6c-a76dca76d466,62ab4714-15b0-4731-be6c-a76dca76d420

• tag_ids: (string)

  A comma separated list of object UUIDs to delete

  Example:

  ?tag_ids=8669b787-15b0-4731-be6c-a76dca76d466,9769b787-15b0-4731-be6c-a76dca76d420

• model_id: (string)

  UUID of the model that was used to generate the labels to be deleted.

  Example:

  ?model_id=tag_ids=a146cdb0-15b0-4731-be6c-a76dca76d424

• generate_type: (string)

  The type of method used to create the labels. Possible values are
  

  • manual -- entered by hand, or manually approved auto-label
    
  • auto -- generated by an auto-label process

  Example:

  ?generate_type=auto

• min_conf: (string)

  The minimum confidence level of labels to be deleted. This filter only applies to labels that were generated by an auto-label operation. Value can range from 0.1 to 1.0.

  Example:

  ?min-conf=0.75

• max_conf: (string)

  The maximum confidence level of labels to be deleted. This filter only applies to labels that were generated by an auto-label operation. Value can range from 0.1 to 1.0.

  Example:

  ?max-conf=0.49

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• file_id: required (string)

  Dataset File UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Body

Media type: application/json

Type: object

• target_dataset_id: required (string)

  UUID of the destination dataset

• files: required (array of string)

  Array of file UUIDs to copy

Example:

{
  "target_dataset_id": "123",
  "files": [
    "e48df358-1b73-411a-9600-25e1d21f33cb"
  ]
}

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• file_id: required (string)

  Dataset File UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Body

Media type: application/json

Type: object

• target_dataset_id: required (string)

  UUID of the destination dataset

• files: required (array of string)

  Array of file UUIDs to copy

Example:

{
  "target_dataset_id": "123",
  "files": [
    "e48df358-1b73-411a-9600-25e1d21f33cb"
  ]
}

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Query Parameters

• ids: (string)

  A comma separated list of tags UUIDs whose labels are to be retrieved

  Example:

  ?ids=45123c471-15b0-4731-be6c-a76dca76d466,62ab4714-15b0-4731-be6c-a76dca76d420

• tag_ids: (string)

  A comma separated list of object UUIDs to retrieve

  Example:

  ?tag_ids=8669b787-15b0-4731-be6c-a76dca76d466,9769b787-15b0-4731-be6c-a76dca76d420

• file_ids: (string)

  A comma separated list of file UUIDs whose object labels are to be retrieved

  Example:

  ?tag_ids=12345678-15b0-4731-be6c-a76dca76d466,abcdef05-15b0-4731-be6c-a76dca76d420

• model_id: (string)

  UUID of the model that was used to generate the labels to be retrieved.

  Example:

  ?model_id=tag_ids=a146cdb0-15b0-4731-be6c-a76dca76d424

• generate_type: (string)

  The type of method used to create the labels. Possible values are
  

  • manual -- entered by hand, or manually approved auto-label
  • auto -- generated by an auto-label process

  Example:

  ?generate_type=auto

• min_conf: (string)

  The minimum confidence level of labels to be returned. This filter only applies to labels that were generated by an auto-label operation. Value can range from 0.1 to 1.0.

  Example:

  ?min-conf=0.75

• max_conf: (string)

  The maximum confidence level of labels to be returned. This filter only applies to labels that were generated by an auto-label operation. Value can range from 0.1 to 1.0.

  Example:

  ?max-conf=0.49

• limit: (integer)

  optional parameter to limit the number of items to return

  Example:

  ?limit=20

• skip: (integer)

  optional parameter to skip the number of items before including items in the return set

  Example:

  ?skip=40

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Query Parameters

• ids: (string)

  A comma separated list of tags UUIDs whose labels are to be deleted

  Example:

  ?ids=45123c471-15b0-4731-be6c-a76dca76d466,62ab4714-15b0-4731-be6c-a76dca76d420

• tag_ids: (string)

  A comma separated list of object UUIDs to delete

  Example:

  ?tag_ids=8669b787-15b0-4731-be6c-a76dca76d466,9769b787-15b0-4731-be6c-a76dca76d420

• file_ids: (string)

  A comma separated list of file UUIDs whose object labels are to be deleted

  Example:

  ?tag_ids=12345678-15b0-4731-be6c-a76dca76d466,abcdef05-15b0-4731-be6c-a76dca76d420

• model_id: (string)

  UUID of the model that was used to generate the labels to be deleted.

  Example:

  ?model_id=tag_ids=a146cdb0-15b0-4731-be6c-a76dca76d424

• generate_type: (string)

  The type of method used to create the labels. Possible values are
  

  • manual -- entered by hand, or manually approved auto-label
    
  • auto -- generated by an auto-label process

  Example:

  ?generate_type=auto

• min_conf: (string)

  The minimum confidence level of labels to be deleted. This filter only applies to labels that were generated by an auto-label operation. Value can range from 0.1 to 1.0.

  Example:

  ?min-conf=0.75

• max_conf: (string)

  The maximum confidence level of labels to be deleted. This filter only applies to labels that were generated by an auto-label operation. Value can range from 0.1 to 1.0.

  Example:

  ?max-conf=0.49

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• label_id: required (string)

  Object Label UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• label_id: required (string)

  Object Label UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• label_id: required (string)

  Object Label UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Body

Media type: application/json

Type: object

• tag_id: (string)

  Identifies the UUID of the tag (class) to which the label should belong.

• generate_type: (one of manual, auto, suggested)

  Changes the generate_type of the label.

• confidence: (number)

  Changes the confidence property of the label. Confidance values range from 0.0 to 1.0.

  Example:

  0.85

• model_id: (string)

  Changes the mode_id of the label.

• bndbox: (object)

  Changes the bounding box for the label.

  NOTE that the bndbox will be ignored if the object label contains a segment_polygons property. In this case, the bounding box values are calculated from the segment_polygons value.

  • xmin: required (integer)

    Upper left corner 'x' coordinate

  • ymin: required (integer)

    Upper left corner 'y' coordinate

  • xmax: required (integer)

    Lower right corner 'x' coordinate

  • ymax: required (integer)

    Lower right corner 'y' coordinate

• segment_polygons: (list)

  Changes the polygon information for the label. All points in the polygon must be provided when changing the polygon information.

  segment_polygons is a 3 dimensional array of points that represent the boundary of the object. The first dimension holds boundaries. It is an array to all for discontingous object identification which can happen when an object is occluded and essentially cut into pieces.

  The 2nd dimension contains arrays of points with each point array (the 3rd dimension) holding an 'x' and 'y' coordinate in that order.

  NOTE: if segment_polygons is present, the bndbox property is ignored and the bounding box information is calculated from the contents of this property.

Examples:

{
  "tag_id": UUID-OF_TAG
}

{
  "segment_polygons": [
    [
      [ 1225, 2009 ],
      [ 1452, 2265 ],
      [ 1225, 2522 ],
      [ 999, 2265 ]
    ]
  ]
}

{
  "generate_type": "manual"
  "segment_polygons": [
    [
      [ 1225, 2009 ],
      [ 1452, 2265 ],
      [ 1225, 2522 ],
      [ 999, 2265 ]
    ]
  ]
}

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Query Parameters

• sortby: (string)

  A comma separated string of field names on which to sort. Add " DESC" after a field name to change to a descending sort.

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Body

Media type: application/json

Type: object

• name: required (string)

  Identifies the name of the tag (class) to create

Example:

{
  name: "elephant"
}

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• tag_id: required (string)

  Object Tag UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• tag_id: required (string)

  Object Tag UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Body

Media type: application/json

Type: object

• action: required (string)

  Identifies the action to peform. Possible actions are

  • rename - Change the name of the tag.
• data: (string)

  The data property only applies to the rename action and is required for that action. It identifies the new name for the tag.

Example:

{
  action: "rename",
  data: "new_name"
}

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Body

Media type: application/json

Type: object

• name: required (string - minLength: 1 - maxLength: 64)

  Name of the new user key.

• description: (string - minLength: 1 - maxLength: 128)

  Optional description of the key (purpose, value format, etc).

Example:

{
  "name": "new key",
  "description": "description of the purpose of this key"
}

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• key_name: required (string)

  The name of the target key.

  Example:

  key-Name

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• key_name: required (string)

  The name of the target key.

  Example:

  key-Name

URI Parameters

• id: required (string)

  Dataset UUID.

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• key_name: required (string)

  The name of the target key.

  Example:

  key-Name

Body

Media type: application/json

Type: object

• name: (string - minLength: 1 - maxLength: 64)

  New name of the user key.

• description: (string - minLength: 1 - maxLength: 128)

  New description of the user key.

Examples:

{
  "name": "changed key name"
}

{
  "name": "changed key name",
  "description": "changed description"
}

Query Parameters

• pg_id: (String)

  Filter the returned list so that only training tasks belonging to the indicated project group are shown.

  Example:

  ?pg_id=6e9f142d-6970-45c4-afac-5ac120631cab

• status: required (string)

  Filter the returned list so that only training tasks with the indicated status are shown.

  Example:

  ?status=failed

• sortby: (String)

  A comma separated string of field names on which to sort. Add " DESC" after a field name to change to a descending sort.

  Examples:

  two field sort:
  

  ?sortby=name

  descending sort:
  

  ?sortby=status DESC, name

Body

Media type: application/json

Type: object

• action: required (string)

  Indicates the training task to be created. Supported actions are

  • pre-analysis -- analyze dataset to ensure minimum requirements for Image Classification training are met.
  • create -- start a training task for classification training.
  • pre-analysis-cod -- analyze dataset to ensure minimum requirements for Object Detection training are met.
  • create-cod-task -- start a training task for object detection training.
• dataset_id: required (string)

  The UUID for the dataset to be trained.

• name: (string)

  The name field gives a name to the training task. It is required for actions that create a training task, but it is not used for pre-analysis tasks.

• nn_arch: (string)

  Optional parameter that applies only to COD training. This parameter identifies the type of COD training to perform. Possible values are...

  • tiny_yolo_v2 -- indicates training should be for tiny YOLO v2.
  • frcnn -- indicates FRCNN training should be performed
  • frcnn_mrcnn -- indicates segmentation training should be performed

  The default value is frcnn.

• pretrained_model: (string)

  Optional parameter to identify the trained model to use as the base neural network from which to start the transfer learning training.

• embing_id: (string)

  Optional parameter that applies to COD training only.

• userdnn_id: (string)

  Optional parameter to identify the user loaded custom DNN to use for training.

• bitwidth: (string)

  Optional parameter that applies to COD YOLO training where the resulting model is to be used in an FPGA.

• strategy: (object)

  Optional object use to modify default training hyperparameters. The content of the strategy document varies by training type and neural network architecture.

  • scenario: (string)

    Optional parameter that applies to CIC training only. Possible values are

    • "flowers"
    • "chineseFood"
    • "general"
  • max_iter: (string)

    Number of iterations to train. Applies to all training types.

    Values range from 100 to 1,000,000. Default value is as follows

    • for CIC training -- 1,500
    • for COD FRCNN training -- 4,000
    • for COD Tiny YOLO v2 -- 40,000
    • for COD Segmentation -- 4,000
  • test_iter: (string)

    Optional parameter that applies to CIC training only.

    Values range from 1 to 1,000. Default value is 100.

  • test_interval: (string)

    Optional parameter that applies to CIC training only. This parameter indicates the iteration interval at which testing will be done. Values range from 1 to 1,000. Default value is 20.

  • learning_rate: (string)

    Optional parameter that applies to all training types.

    Possible values range from 0.0 to 0.01.

  • weight_decay: (string)

    Optional parameter that applies to all training types.

    Possible values range from 0.0 to 0.5. Default value is 0.0005

  • ratio: (string)

    Optional parameter that applies to all types of COD training. It identifies the percentage of images to use for training, and inversely the percentage of images to reserve for testing.

    Possible values range from 0.5 to 1.0. Default value is 0.8.

Examples:

{
  "action": "pre-analysis", // pre analysis for cic task
  "dataset_id": "6d69b787-15b0-4731-be6c-a76dca76d466",
  "strategy": { // optional, if not set, will use default value
    "max_iter": 1500, // optional, [100-1000000], default to 1500
    "test_iter": 100, // optional, [1-1000], default to 100
    "test_interval": 20, // optional, [1-1000], default to 20
    "learning_rate": 0.001, // optional, (0-0.01], default to 0.001
    "weight_decay": 0.0005 // optional, (0-0.5), default to 0.0005
  }
}

{
  "action": "create", // create cic task
  "dataset_id": "6d69b787-15b0-4731-be6c-a76dca76d466",
  "strategy": { // optional, if not set, will use default value
    "max_iter": 1500, // optional, [100-1000000], default to 1500
    "test_iter": 100, // optional, [1-1000], default to 100
    "test_interval": 20, // optional, [1-1000], default to 20
    "learning_rate": 0.001, // optional, (0-0.01], default to 0.001
    "weight_decay": 0.0005 // optional, (0-0.5), default to 0.0005
  }
  "name": "cictest"
}
// reponse for action "pre-analysis"
{
  "action": "pre-analysis-cod", // pre analysis for cod task
  "dataset_id": "6d69b787-15b0-4731-be6c-a76dca76d466",
  "strategy": { // optional, if not set, will use default value
    "max_iter": 4000, // optional, [100-1000000], default to 4000
    "momentum": 0.9, // optional, [0-1.0], default to 0.9
    "ratio": 0.8, // optional, [0.5-1.0], default to 0.8
    "learning_rate": 0.001, // optional, (0-0.01], default to 0.001
    "weight_decay": 0.0005 // optional, (0-0.5), default to 0.0005
  }
}
{
  "action": "create-cod-task", // create cod task
  "dataset_id": "6d69b787-15b0-4731-be6c-a76dca76d466",
  "strategy": { // optional, if not set, will use default value
    "max_iter": 4000, // optional, [100-1000000], default to 4000
    "momentum": 0.9, // optional, [0-1.0], default to 0.9
    "ratio": 0.8, // optional, [0.5-1.0], default to 0.8
    "learning_rate": 0.001, // optional, (0-0.01], default to 0.001
    "weight_decay": 0.0005 // optional, (0-0.5), default to 0.0005
  }
  "name": "codtest"
}

URI Parameters

• id: required (string)

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

URI Parameters

• id: required (string)

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

URI Parameters

• id: required (string)

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Body

Media type: application/json

Type: object

Example:

{
  action: "rename",
  data: "xxx" // the new name of task
}

URI Parameters

• id: required (string)

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

Query Parameters

• usage: (String)

  Filter the returned list so that only trained models trained for the indicated usage are shown. Possible usage values are "cic", "cod", or "act"

  Example:

  ?usage:cod

• pg_id: (String)

  Filter the returned list so that only trained models belonging to the indicated project group are shown.

  Example:

  ?pg_id=6e9f142d-6970-45c4-afac-5ac120631cab

• production_status: required (string)

  Filter the returned list so that only trained models with the indicated production_status are shown.

  Example:

  ?production_status=rejected

• sortby: (String)

  A comma separated string of field names on which to sort. Add " DESC" after a field name to change to a descending sort.

  Examples:

  two field sort:
  

  ?sortby=name

  descending sort:
  

  ?sortby=status DESC, name

Body

Media type: multipart/form-data

Type: object

Example:

// Post params
Content-Disposition: form-data;
  [
    name="files"; filename="mymodel.tar.gz" Content-Type: .tar.gz => binary data
  ]

// HTML usage
<form enctype="multipart/form-data" action="{baseUri}/trained-models" method="POST">
    Choose a file to upload:
    <input name="files" type="file" /><br />
    <input type="submit" value="Upload File" />
</form>

// curl usage
curl -i -X POST -H "Content-Type: multipart/form-data" -F "files=@mymodel.tar.gz" {baseUri}/trained-models

URI Parameters

• id: required (string)

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

URI Parameters

• id: required (string)

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

URI Parameters

• id: required (string)

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

URI Parameters

• id: required (string)

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• assettype: required (string)

  The type of asset on which to operate.
  Currently the supported asset types are coreml and tensorrt.

URI Parameters

• id: required (string)

  Example:

  a48c3779-eb14-4697-a490-eeb0125196f4

• assettype: required (string)

  The type of asset on which to operate.
  Currently the supported asset types are coreml and tensorrt.

Body

Media type: application/json

Type: object

Example:

{
  action: "delete",
  model_list: [ "modelID1", "modelID2" ] //IDs of trained models to be deleted
}

Query Parameters

• sortby: (String)

  A comma separated string of field names on which to sort. Add " DESC" after a field name to change to a descending sort.

  Examples:

  two field sort:
  

  ?sortby=name

  descending sort:
  

  ?sortby=usage DESC, name

Body

Media type: application/json

Type: object

• trained_model_id: required (string)

  The UUID of the trained model to deploy.

• name: required (string)

  The name of the deployed model.

• usage: required (string)

  The usage by which the model was trained. Possible values are cic, cod, act.

• accel_type: (string)

  Optional field to identify the type of acceleration method to use. Possible values are GPU, GPU_TENSORRT, CPU, FPGA_XFDNN_8_BIT, FPGA_XFDNN_16_BIT

• userdnn_id: (string)

  The custom model UUID. This value is to be set when deploying a custom model.

• dnnscript_id: (string)

  The DNN script UUID to use on inference.

• save_inference: (string)

  The dataset UUID where all inference results performed on this webapi will be saved.

Example:

{
  "trained_model_id:" "a48c3779-eb14-4697-a490-eeb0125196f4",
  "name": "Deployed Model Name",
  "usage": "cic",
  "accel_type": "GPU"
}

Media type: properties

Type: object

Media type: example

Type:

{
  "trained_model_id:" "a48c3779-eb14-4697-a490-eeb0125196f4",
  "name": "Deployed Model Name",
  "usage": "cic",
  "accel_type": "GPU"
}

URI Parameters

• id: required (string)

  Deployed Model UUID (also known as a web API)

  Example:

  fbb78a52-ae41-4000-bbe5-ad0d116c4e9d

Body

Media type: application/json

Type: object

Example:

{
  "action": "delete",
  "webapi_list": [ "WebAPIID1", "WebAPIID2" ] // IDs of deployed models to be deleted
}

URI Parameters

• id: required (string)

  Unique ID of the deployed model

  Example:

  4c3259c3-e300-41a7-b031-d6fbf508cab7

Query Parameters

• imageUrl: required (string)

  URL for an image. May be on the internet or a local file

  Example:

  http://IP/path/fff287bc-412b-4b5d-9f8a-a3af27ca74d7.jpg

• containHeatMap: (string)

  Flag parameter to request inclusion of the heat map in classification results. This parameter is only applicable on classification models and is ignored on on all model types.

  Possible values are "true" to include the heat map, or "false" to exclude the heat map.
  The default value is "false" if the parameter is not provided.

• containRle: (string)

  Flag parameter to request inclusion of the compressed RLE string in object detection segmentaion results. This parameter is only applicable on object detection models that can provide instance segmentation results. It is ignored on on all model types. Possible values are "true" to include the compressed RLE string, or "false" to exclude it.>br> The default value is "false" if the parameter is not provided.

• containPolygon: (string)

  Flag parameter to request inclusion of the polygon boundary information in object detection segmentaion results. This parameter is only applicable on object detection models that can provide instance segmentation results. It is ignored on on all model types.

  Possible values are "true" to include the polygon information, or "false" to exclude it.
  The default value is "true" if the parameter is not provided.

• confthre: (float)

  Parameter to specify the confidence threshold to use when providing results. This parameter applies to both classification and object detection models. It is a floating point value that specifies the threshold at which possible matches are to be excluded from the result set. A value of 0.5 will exclude possible matches with a confidence score below 50%.

  Possible values range from 0.0 to 0.9.
  The default value is 0.0 if this parameter is not specified.

• clsnum: (integer)

  This parameter is used to identify how many different possible matches are included in the results. For classification, this means that if there are multiple possible matches with a confidence score above the confthre value, up to the indicated value matches are included in the result. Values are integer values where 0 is the default (which means to include all possible matches in the results).

• waitForResults: (string)

  This parameter is used for video inferencing. By default, video inference results are reported asynchronously -- meaning that the call to do the inference will return while processing continues on the video. Inference results are available via the /inferences endpoint.

  If this parameter is passed with a value of "true" (case is ignored), the endpoint will not return until the video has completed processing. The JSON results will be returned at that time. Note that the JSON results will also be available via the /inferences endpoint.

• genCaption: (string)

  This parameter is used for video inferencing. It controls whether or not a "captioned" video should be generated and stored with the inferencing results. The default varies by the detection being done. For Action Detection, the default is to generate a captioned video. For Object Detection, the default is to NOT generate a captioned video.

  The captioned video is accessible via the /inferences endpoint.

URI Parameters

• id: required (string)

  Unique ID of the deployed model

  Example:

  4c3259c3-e300-41a7-b031-d6fbf508cab7

Body

Media type: multipart/form-data

Type: object

• files: required (string)

  octet file stream of the file to be inferred.

• containHeatMap: (string)

  Flag parameter to request inclusion of the heat map in classification results. This parameter is only applicable on classification models and is ignored on on all model types.

  Possible values are "true" to include the heat map, or "false" to exclude the heat map.>br> The default value is "false" if the parameter is not provided.

• containRle: (string)

  Flag parameter to request inclusion of the compressed RLE string in object detection segmentaion results. This parameter is only applicable on object detection models that can provide instance segmentation results. It is ignored on on all model types. Possible values are "true" to include the compressed RLE string, or "false" to exclude it.
  The default value is "false" if the parameter is not provided.

• containPolygon: (string)

  Flag parameter to request inclusion of the polygon boundary information in object detection segmentaion results. This parameter is only applicable on object detection models that can provide instance segmentation results. It is ignored on on all model types.

  Possible values are "true" to include the polygon information, or "false" to exclude it.
  The default value is "true" if the parameter is not provided.

• confthre: (float)

  Parameter to specify the confidence threshold to use when providing results. This parameter applies to both classification and object detection models. It is a floating point value that specifies the threshold at which possible matches are to be excluded from the result set. A value of 0.5 will exclude possible matches with a confidence score below 50%.

  Possible values range for 0.0 to 0.9.
  The default value is 0.0 if this parameter is not specified.

• clsnum: (integer)

  This parameter is used to identify how many different possible matches are included in the results. For classification, this means that if there are multiple possible matches with a confidence score above the confthre value, up to the indicated value matches are included in the result. Values are integer values where 0 is the default (which means to include all possible matches in the results).

• waitForResults: (string)

  This parameter is used for video inferencing. By default, video inference results are reported asynchronously -- meaning that the call to do the inference will return while processing continues on the video. Inference results are available via the /inferences endpoint.

  If this parameter is passed with a value of "true" (case is ignored), the endpoint will not return until the video has completed processing. The JSON results will be returned at that time. Note that the JSON results will also be available via the /inferences endpoint.

• genCaption: (string)

  This parameter is used for video inferencing. It controls whether or not a "captioned" video should be generated and stored with the inferencing results. The default varies by the detection being done. For Action Detection, the default is to generate a captioned video. For Object Detection, the default is to NOT generate a captioned video.

  The captioned video is accessible via the /inferences endpoint.

Media type: example

Type:

// Post params
Content-Disposition: form-data;
**** THIS EXAMPLE NEEDS TO BE UPDATED OR REMOVED  ****
  [
    name="files"; filename="mytemp.jpg" Content-Type: image/jpeg => image binary data
  ]

// HTML usage
**** THIS EXAMPLE NEEDS TO BE UPDATED OR REMOVED  ****
<form enctype="multipart/form-data" action="{baseUri}/dlapis/4c3259c3-e300-41a7-b031-d6fbf508cab7" method="POST">
    Choose a file to upload:
    <input name="files" type="file" /><br />
    <input type="submit" value="Upload File" />
</form>

// curl usage
curl -i -F files=@mytemp.jpg {baseUri}/dlapis/4c3259c3-e300-41a7-b031-d6fbf508cab7

Query Parameters

• newer_than: (string)

  optional parameter to filter the inference results to only get results with updates more recent than the given date/time stamp of the form "YYYY/MM/DDTHH:MM:SS". Note that all times are UTC times.

  Example:

  ?newer_than="2019/06/01T09:11:48"

• model_id: (string)

  Optional parameter to limit results to those inference performed by the given model.

  Example:

  ?model_id="model-id-string"

• status: (string)

  optional parameter to filter the background tasks to only get tasks with the given status. Possible status values are starting, working, completed, and failed.

• sortby: (String)

  A comma separated string of field names on which to sort. Add " DESC" after a field name to change to a descending sort.

  Examples:

  two field sort:
  

  ?sortby=model_id

  descending sort:
  

  ?sortby=model_id, status DESC, total_frames