You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: md-docs/user_guide/data_schema.md
+9-3Lines changed: 9 additions & 3 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -56,15 +56,21 @@ Some tasks can have different data entities for the same Role, the Column object
56
56
| --|--|--|--|
57
57
| RAG User Input | INPUT | String | In RAG Tasks it is the user query submitted to the system. |
58
58
| RAG Retrieved Context | INPUT | String | In RAG Tasks it is the retrieved contexts (separated with the Task attribute *context separator*) that the retrieval system has selected to answer the query.|
59
-
| Model probability | PREDICTION | Depends on Task Type:<br><ul><li>RAG: Array 1</li><li>Classification Binary: Float</li><li>Classification Multiclass: Array 1</li><li>Classification Multilabel: Array 1</li></ul> | It is automatically created by ML cube Platform when the created Model has the flag additional probabilistic output set as True. The name has fixed template: <MODEL_NAME\>_probability\@<MODEL_VERSION\>.|
60
-
| Object detection prediction label| PREDICTION | Array 1 | It is automatically created when Task Type is Object detection. It is an array with length equal to the number of predicted bounding boxes where each element contains the class label assigned to the bounding box. The name has a fixed template: <MODEL_NAME\>_predicted_labels\@<MODEL_VERSION\>.|
61
-
| Object detection target label| TARGET | Array 1 | It is mandatory when Task Type is Object detection. It is an array with length equal to the number of actual bounding boxes where each element contains the class label assigned to the bounding box. |
59
+
| Model probability | PREDICTION | Depends on Task Type:<br><ul><li>RAG: Array 1</li><li>Classification Binary: Float</li><li>Classification Multiclass: Array 1</li><li>Classification Multilabel: Array 1</li><li>Semantic Segmentation: Array 3</li></ul> | It is automatically created by ML cube Platform when the created Model has the flag additional probabilistic output set as True. The name has fixed template: <MODEL_NAME\>_probability\@<MODEL_VERSION\>.|
60
+
| Object prediction label| PREDICTION | Array 1 | It is automatically created when Task Type is Object detection or Semantic Segmentation. It is an array with length equal to the number of predicted bounding boxes where each element contains the class label assigned to the bounding box. The name has a fixed template: <MODEL_NAME\>_predicted_labels\@<MODEL_VERSION\>.|
61
+
| Object target label| TARGET | Array 1 | It is mandatory when Task Type is Object detection or Semantic Segmentation. It is an array with length equal to the number of actual bounding boxes where each element contains the class label assigned to the bounding box. |
62
62
63
63
## Data schema constraints
64
64
65
65
Each combination of Task Type and Data Structure leads to different Data Schema requirements that must be satisfied when it is created for the Task.
66
66
For instance, image binary classification tasks requires only one input column object with image data type and target column object must be categorical with only two possible values.
67
67
68
+
!!! note
69
+
Object Detection and Semantic Segmentation have specific constraints about the _dims_ attribute of the TARGET and PREDICTION columns:
70
+
71
+
- Object Detection [-1, 4]: the first is for identified objects, the second is for bounding box specification: x_min, x_max, y_min, y_max
72
+
- Semantic Segmentation [-1, -1, 2]: the first is for identified objects, the second is for polygon vertices, the third is for vertices coordinates x, y
73
+
68
74
Here the list of constraints about quantities for each Role:
| USER INPUT RETRIEVED CONTEXT |||||||:material-check: |
93
+
| USER INPUT MODEL OUTPUT |||||||:material-check: |
94
+
| MODEL OUTPUT RETRIEVED CONTEXT |||||||:material-check: |
95
95
96
96
#### Monitoring Metrics
97
97
@@ -111,8 +111,13 @@ This table is subject to changes, as new metrics will be added in the future.
111
111
| MODEL PERPLEXITY | A measure of the uncertainty of an LLM when predicting the next words | PREDICTION | When the task type is RAG | A floating point value. |
112
112
| IMAGE BRIGHTNESS | The brightness of the image | INPUT | When the data structure is image | A floating point value. |
113
113
| IMAGE CONTRAST | The contrast of the image | INPUT | When the data structure is image | A floating point value. |
114
-
| BBOXES AREA | The average area of the predicted bounding boxes | PREDICTION | When the task type is Object Detection | A floating point value. |
115
-
| BBOXES QUANTITY | The average number of predicted bounding boxes per image | PREDICTION | When the task type is Object Detection | An integer value. |
114
+
| AVERAGE AREA PER OBJECT TYPE | Average area of identified objects of the same type. If a sample does not have labels of that type then this metric is missing. | PREDICTION | When Task Type is Object Detection or Semantic Segmentation | A floating point |
115
+
| QUANTITY PER OBJECT TYPE | Number of identified objects for each type in the image. | PREDICTION | When Task Type is Object Detection or Semantic Segmentation | A array with one dimensions. Each element corresponds to an object label. |
116
+
| TOTAL OBJECTS | Total number of identified objects independently from the object type. | PREDICTION | When Task Type is Object Detection or Semantic Segmentation | A floating point |
117
+
| OBJECT TYPES COUNT | Number of different object types identified in the image. It differs from the other because it only counts the number of different labels and not the number of objects per labels or objects in total. | PREDICTION | When Task Type is Object Detection or Semantic Segmentation | A floating point |
118
+
| MODEL ENTROPY | Uncertainty of the model predicting objects in the images | PREDICTION PROBABILITY | When Task Type is Object Detection or Semantic Segmentation | A floating point |
0 commit comments