|
38 | 38 | "The `Model Run` object represents the experiment itself. It is a specific instance of a `Model` with preconfigured hyperparameters (training data). You can upload inferences across each `Model Run`, filter by IoU score, and compare your model's predictions against the annotations from your training data.\n", |
39 | 39 | "\n", |
40 | 40 | "## Steps\n", |
41 | | - "\n", |
42 | | - "1. Have a set of ground truth labels in a project\n", |
43 | | - "2. Install the latest SDK release (At this time that is 3.0.0rc0)\n", |
44 | | - "3. Create a `Model`\n", |
45 | | - "4. Create a `Model Run`\n", |
46 | | - "5. Compute predictions\n", |
47 | | - "6. Compute model performance metrics\n", |
48 | | - "7. Upload labels, predictions, and metrics\n", |
49 | | - "8. Navigate to the `Models` tab on Labelbox\n", |
| 41 | + "1. Make sure you are signed up for the beta. If not navigate here https://labelbox.com/product/model-diagnostics\n", |
| 42 | + "2. Have a set of ground truth labels in a project\n", |
| 43 | + "3. Install the latest SDK release (At this time that is 3.0.0rc1)\n", |
| 44 | + "4. Create a `Model`\n", |
| 45 | + "5. Create a `Model Run`\n", |
| 46 | + "6. Compute predictions\n", |
| 47 | + "7. Compute model performance metrics\n", |
| 48 | + "8. Upload labels, predictions, and metrics\n", |
| 49 | + "9. Navigate to the `Models` tab on Labelbox\n", |
50 | 50 | "\n", |
51 | 51 | "## Best practices\n", |
52 | | - "Currently there is a limit of 2000 images per model run. We suggest uploading lower performing examples from your test set." |
| 52 | + "Currently there is a limit of 2000 images per model run. We suggest uploading lower performing examples from your test set.\n" |
53 | 53 | ] |
54 | 54 | }, |
55 | 55 | { |
|
0 commit comments