> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.conductorquantum.com/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.conductorquantum.com/_mcp/server.

# Overview

> ML models for automated analysis of quantum device data, powering Control's calibration workflow.

Machine learning models for automated analysis of quantum device data. Models power the analysis step of Control's calibration loop, replacing hours of manual interpretation with structured, automated results. Use them during device tune-up and characterization to quickly extract key device parameters from your measurement data.

Regardless of the specific model, the SDK always wraps the response in a `ModelResultInfo` object:

| Field             | Type       | Description                                    |
| ----------------- | ---------- | ---------------------------------------------- |
| `id`              | `str`      | Unique identifier for the execution            |
| `model`           | `str`      | Name of the executed model                     |
| `created_at`      | `datetime` | Timestamp (UTC) when the execution was created |
| `input_file_name` | `str`      | Name of the uploaded input file                |
| `input_file_size` | `int`      | Size of the input file in bytes                |
| `output`          | `Dict`     | Model‑specific output (boolean or dictionary)  |

## Model Versioning

Some of our models use a versioning scheme (v0, v1, etc.) where higher numbers generally indicate improved
accuracy and performance. When selecting a model version, consider:

* **Newer versions** typically offer better accuracy and robustness
* **Specialized versions** for specific data types or conditions
* **Performance variations** based on your measurement characteristics

We recommend testing multiple versions to find the optimal one for your quantum device data.

### Charge Stability Diagram Analysis

| Model                                                 | Description                                                                    | Input Shape | Output Keys               | Endpoints                                    |
| ----------------------------------------------------- | ------------------------------------------------------------------------------ | ----------- | ------------------------- | -------------------------------------------- |
| `charge-stability-diagram-segmenter-v0`               | Segments a charge stability diagram into different regions                     | (128, 128)  | `segmentation`            | [Run](/api-reference/control-api/models/run) |
| `charge-stability-diagram-binary-classifier-v0-16x16` | Classifies a charge stability diagram as no dot, single, or double dot         | (16, 16)    | `classification`          | [Run](/api-reference/control-api/models/run) |
| `charge-stability-diagram-binary-classifier-v1-16x16` | Classifies a charge stability diagram as no dot, single, or double dot         | (16, 16)    | `classification`          | [Run](/api-reference/control-api/models/run) |
| `charge-stability-diagram-binary-classifier-v2-16x16` | Classifies a charge stability diagram as no dot, single, or double dot         | (16, 16)    | `classification`          | [Run](/api-reference/control-api/models/run) |
| `charge-stability-diagram-binary-classifier-v3-16x16` | Classifies a charge stability diagram as no dot, single, or double dot         | (16, 16)    | `classification`, `score` | [Run](/api-reference/control-api/models/run) |
| `charge-stability-diagram-binary-classifier-v0-48x48` | Classifies a charge stability diagram as no dot, single, or double dot         | (48, 48)    | `classification`, `score` | [Run](/api-reference/control-api/models/run) |
| `charge-stability-diagram-binary-classifier-v1-48x48` | Classifies a charge stability diagram as no dot, single, or double dot         | (48, 48)    | `classification`, `score` | [Run](/api-reference/control-api/models/run) |
| `charge-stability-diagram-transition-detector-v0`     | Detects the locations of charge transition lines in a charge stability diagram | (n, m)      | `transition_lines`        | [Run](/api-reference/control-api/models/run) |
| `charge-stability-diagram-transition-detector-v1`     | Detects the locations of charge transition lines in a charge stability diagram | (n, m)      | `transition_lines`        | [Run](/api-reference/control-api/models/run) |
| `charge-stability-diagram-transition-detector-v2`     | Detects the locations of charge transition lines in a charge stability diagram | (n, m)      | `transition_lines`        | [Run](/api-reference/control-api/models/run) |
| `charge-stability-diagram-transition-detector-v3`     | Detects the locations of charge transition lines in a charge stability diagram | (n, m)      | `transition_lines`        | [Run](/api-reference/control-api/models/run) |

You can see how to use the charge stability diagram models on the [Charge Stability Diagram Models](/control/models/charge-stability-diagram) page.

### Coulomb Blockade Analysis

| Model                                    | Description                                                                   | Input Shape | Output Keys               | Endpoints                                    |
| ---------------------------------------- | ----------------------------------------------------------------------------- | ----------- | ------------------------- | -------------------------------------------- |
| `coulomb-blockade-classifier-v0`         | Determines if a current measurement exhibits Coulomb blockade characteristics | (n, )       | `classification`          | [Run](/api-reference/control-api/models/run) |
| `coulomb-blockade-classifier-v1`         | Determines if a current measurement exhibits Coulomb blockade characteristics | (n, )       | `classification`          | [Run](/api-reference/control-api/models/run) |
| `coulomb-blockade-classifier-v2`         | Determines if a current measurement exhibits Coulomb blockade characteristics | (128, )     | `classification`, `score` | [Run](/api-reference/control-api/models/run) |
| `coulomb-blockade-classifier-v3`         | Determines if a current measurement exhibits Coulomb blockade characteristics | (128, )     | `classification`, `score` | [Run](/api-reference/control-api/models/run) |
| `coulomb-blockade-peak-detector-v0`      | Detects the locations of Coulomb blockade peaks in current measurement        | (n, )       | `peak_indices`            | [Run](/api-reference/control-api/models/run) |
| `coulomb-blockade-peak-detector-v1`      | Detects the locations of Coulomb blockade peaks in current measurement        | (n, )       | `peak_indices`            | [Run](/api-reference/control-api/models/run) |
| `coulomb-blockade-peak-detector-v1-mini` | Detects the locations of Coulomb blockade peaks in current measurement        | (n, )       | `peak_indices`            | [Run](/api-reference/control-api/models/run) |
| `coulomb-blockade-peak-detector-v2`      | Detects the locations of Coulomb blockade peaks in current measurement        | (n, )       | `peak_indices`            | [Run](/api-reference/control-api/models/run) |

You can see how to use the Coulomb Blockade models on the [Coulomb Blockade Models](/control/models/coulomb-blockade) page.

Note: `coulomb-blockade-peak-detector-v1-mini` is a quicker version of `coulomb-blockade-peak-detector-v1` that gives a good trade-off between accuracy and speed.

### Coulomb Diamond Analysis

| Model                          | Description                                                                       | Input Shape | Output Keys        | Endpoints                                    |
| ------------------------------ | --------------------------------------------------------------------------------- | ----------- | ------------------ | -------------------------------------------- |
| `coulomb-diamond-segmenter-v0` | Segments a Coulomb diamond conductance grid to Coulomb diamond regions            | (n, m)      | `segmentation`     | [Run](/api-reference/control-api/models/run) |
| `coulomb-diamond-detector-v0`  | Detects the vertices of Coulomb diamond structures from a binary mask             | (n, m)      | `diamond_vertices` | [Run](/api-reference/control-api/models/run) |
| `coulomb-diamond-detector-v1`  | Detects the vertices of Coulomb diamond structures directly from conductance data | (n, n)      | `diamond_vertices` | [Run](/api-reference/control-api/models/run) |
| `coulomb-diamond-detector-v2`  | Detects the vertices of Coulomb diamond structures directly from conductance data | (n, n)      | `diamond_vertices` | [Run](/api-reference/control-api/models/run) |

You can see how to use the Coulomb Diamond models on the [Coulomb Diamond Models](/control/models/coulomb-diamond) page.

### Pinch-off Analysis

| Model                              | Description                                                                                                       | Input Shape | Output Keys                                             | Endpoints                                    |
| ---------------------------------- | ----------------------------------------------------------------------------------------------------------------- | ----------- | ------------------------------------------------------- | -------------------------------------------- |
| `pinch-off-classifier-v0`          | Determines if a current measurement exhibits pinch-off characteristics                                            | (n, )       | `classification`                                        | [Run](/api-reference/control-api/models/run) |
| `pinch-off-parameter-extractor-v0` | Extracts key parameters from pinch-off measurements (cut\_off\_voltage, transition\_voltage, saturation\_voltage) | (n, )       | `cut_off_index`, `transition_index`, `saturation_index` | [Run](/api-reference/control-api/models/run) |

You can see how to use the Pinch-off models on the [Pinch-off Models](/control/models/pinch-off) page.

### Turn-on Analysis

| Model                            | Description                                                          | Input Shape | Output Keys       | Endpoints                                    |
| -------------------------------- | -------------------------------------------------------------------- | ----------- | ----------------- | -------------------------------------------- |
| `turn-on-classifier-v0`          | Determines if a current measurement exhibits turn-on characteristics | (n, )       | `classification`  | [Run](/api-reference/control-api/models/run) |
| `turn-on-parameter-extractor-v0` | Extracts the threshold voltage from turn-on measurements             | (n, )       | `threshold_index` | [Run](/api-reference/control-api/models/run) |

You can see how to use the Turn-on models on the [Turn-on Models](/control/models/turn-on) page.

### Resonator Spectroscopy Analysis

| Model                     | Description                                             | Input Shape | Output Keys   | Endpoints                                    |
| ------------------------- | ------------------------------------------------------- | ----------- | ------------- | -------------------------------------------- |
| `resonator-dip-finder-v0` | Detects resonator dips in spectroscopy magnitude traces | (n,)        | `dip_indices` | [Run](/api-reference/control-api/models/run) |

You can see how to use the Resonator Dip Finder model on the [Resonator Dip Finder](/control/models/resonator-dip-finder) page.

### QEC Decoding (NVIDIA Ising Decoding)

| Model                        | Description                                                             | Input Shape         | Output Keys                       | Endpoints                                    |
| ---------------------------- | ----------------------------------------------------------------------- | ------------------- | --------------------------------- | -------------------------------------------- |
| `ising-decoding-v1-fast`     | Low-latency AI pre-decoder for surface code QEC (0.9M params, R=9)      | (batch, 4, T, D, D) | `logits`, `variant`, `batch_size` | [Run](/api-reference/control-api/models/run) |
| `ising-decoding-v1-accurate` | Higher-accuracy AI pre-decoder for surface code QEC (1.8M params, R=13) | (batch, 4, T, D, D) | `logits`, `variant`, `batch_size` | [Run](/api-reference/control-api/models/run) |

These models accept detector syndrome tensors as 5D NumPy arrays and return correction logits. See the [QEC Decoding](/control/models/qec-decoding) page for details and code examples.

### Agents

| Agent                  | Description                                                               | Input                               | Endpoints                                    |
| ---------------------- | ------------------------------------------------------------------------- | ----------------------------------- | -------------------------------------------- |
| `ising-calibration-v1` | Analyses quantum calibration plots using the NVIDIA Ising Calibration VLM | `image_base64`, `prompt` (optional) | [Run](/api-reference/control-api/agents/run) |

See the [NVIDIA Ising Calibration](/control/agents) page for details and code examples.