---
language:
- en
library_name: transformers
pipeline_tag: image-text-to-text
license: apache-2.0
datasets:
- ServiceNow/BigDocs-Sketch2Flow
base_model:
- mistralai/Pixtral-12B-2409
---
# Model Card for ServiceNow/Pixtral-12B-2409-StarFlow

Pixtral-12B-2409-StarFlow is a vision-language model finetuned for **structured workflow generation from sketch images**. It translates hand-drawn or computer-generated workflow diagrams into structured JSON workflows, including triggers, flow logic, and actions.

## Model Details

### Model Description

Pixtral-12B-2409-StarFlow is part of the **StarFlow** framework for automating workflow creation. It extends Pixtral-12B with domain-specific finetuning on workflow diagrams, enabling accurate sketch-to-workflow generation.

* **Developed by:** ServiceNow Research
* **Model type:** Transformer-based Vision-Language Model (VLM)
* **Language(s) (NLP):** English
* **License:** Apache 2.0
* **Finetuned from model:** [Pixtral-12B](https://huggingface.co/mistralai/Pixtral-12B-2409)

### Model Sources

* **Repository:** [ServiceNow/Pixtral-12B-2409-StarFlow](https://huggingface.co/ServiceNow/Pixtral-12B-2409-StarFlow)
* **Paper:** [StarFlow: Generating Structured Workflow Outputs From Sketch Images](https://arxiv.org/abs/2503.21889);

---

## Uses

### Direct Use

* Translating **sketches of workflows** (hand-drawn, whiteboard, or digital diagrams) into **JSON structured workflows**.
* Supporting **workflow automation** in enterprise platforms by removing the need for manual low-code configuration.

### Downstream Use

* Integration into **enterprise low-code platforms** for rapid prototyping of workflows by users.
* Used in **automation migration pipelines**, e.g., converting legacy workflow screenshots into JSON representations.

### Out-of-Scope Use

* General-purpose vision-language tasks (e.g., image captioning, OCR).
* Use on domains outside workflow automation (e.g., arbitrary diagram-to-code).
* Real-time handwriting recognition (StarFlow focuses on structured workflow translation, not raw OCR).

---

## Bias, Risks, and Limitations

* **Limited generalization**: Finetuned models perform poorly on out-of-distribution diagrams from unfamiliar platforms.
* **Sensitivity to input style**: Whiteboard/handwritten sketches degrade performance compared to digital or UI-rendered workflows.
* **Component naming mismatches**: Model may mispredict action definitions (e.g., “create\_user” vs. “create\_a\_user”), leading to execution errors.
* **Evaluation gap**: Current metrics don’t always reflect execution correctness of generated workflows.

### Recommendations

Users should:

* Validate outputs before deployment.
* Be cautious with **handwritten/ambiguous sketches**.
* Consider supplementing with **retrieval-augmented generation (RAG)** or **tool grounding** for robustness.

---

## How to Get Started with the Model

```python
from transformers import AutoProcessor, LlavaForConditionalGeneration
from PIL import Image

processor = AutoProcessor.from_pretrained("ServiceNow/Pixtral-12B-2409-StarFlow")
model = LlavaForConditionalGeneration.from_pretrained("ServiceNow/Pixtral-12B-2409-StarFlow")

image = Image.open("workflow_sketch.png")
inputs = processor(images=image, text="Generate workflow JSON", return_tensors="pt")

outputs = model.generate(**inputs, max_length=4096)
workflow_json = processor.decode(outputs[0], skip_special_tokens=True)

print(workflow_json)
```

---

## Training Details

### Training Data

The model was trained using the [ServiceNow/BigDocs-Sketch2Flow](https://huggingface.co/datasets/ServiceNow/BigDocs-Sketch2Flow) dataset, which includes the following data distribution:

* **Synthetic** (12,376 Graphviz-generated diagrams)
* **Manual** (3,035 sketches hand-drawn by annotators)
* **Digital** (2,613 diagrams drawn using software)
* **Whiteboard** (484 sketches drawn on whiteboard / blackboard)
* **User Interface** (373 screenshots from ServiceNow Flow Designer)

### Training Procedure

#### Preprocessing

* Synthetic workflows generated via **heuristics** (Scheduled Loop, IF/ELSE, FOREACH, etc.).
* Annotators recreated flows in digital, manual, and whiteboard formats.

#### Training Hyperparameters

* Optimizer: **AdamW** with β=(0.95,0.999), lr=2e-5, weight decay=1e-6.
* Scheduler: **cosine learning rate** with 30 warmup steps.
* Early stopping based on validation loss.
* Precision: **bf16 mixed-precision**.
* Sequence length: up to **32k tokens**.

#### Speeds, Sizes, Times

* Trained with **16× NVIDIA H100 80GB GPUs** across two nodes.
* Full Sharded Data Parallel (FSDP) training, no CPU offloading.

---

## Evaluation

### Testing Data

Same dataset distribution as training: synthetic, manual, digital, whiteboard, UI-rendered workflows.

### Factors

* **Source of sample** (synthetic, manual, UI, etc.)
* **Orientation** (portrait vs. landscape diagrams)
* **Resolution** (small <400k pixels, medium, large >1M pixels)

### Metrics

All Evaluation metrics can be found in the official [StarFlow repo](https://github.com/ServiceNow/StarFlow).

* **Flow Similarity (FlowSim)** – tree edit distance similarity.
* **TreeBLEU** – structural recall of subtrees.
* **Trigger Match (TM)** – accuracy of workflow triggers.
* **Component Match (CM)** – overlap of predicted vs. gold components.

### Results

* Proprietary models (GPT-4o, Claude-3.7, Gemini 2.0) outperform open-weights **without finetuning**.
* **Finetuned Pixtral-12B achieves SOTA**:

  * FlowSim w/ inputs: **0.919**
  * TreeBLEU w/ inputs: **0.950**
  * Trigger Match: **0.753**
  * Component Match: **0.930**

#### Summary

Finetuning yields **large gains over base Pixtral-12B and GPT-4o**, particularly in matching workflow components and triggers.

---

## Model Examination

* Finetuned models capture **naming conventions** and structured execution logic better.
* Failure modes include **missing ELSE branches** or **generic table names**.

---

## Technical Specifications

### Model Architecture and Objective

* Base: **Pixtral-12B**, a multimodal transformer (12B parameters).
* Objective: **Image-to-JSON structured workflow generation**.

### Compute Infrastructure

* **Hardware:** 16× NVIDIA H100 80GB (2 nodes)
* **Software:** FSDP, bf16 mixed precision, PyTorch/Transformers

---

## Citation

**BibTeX:**

```bibtex
@article{bechard2025starflow,
  title={StarFlow: Generating Structured Workflow Outputs from Sketch Images},
  author={B{\'e}chard, Patrice and Wang, Chao and Abaskohi, Amirhossein and Rodriguez, Juan and Pal, Christopher and Vazquez, David and Gella, Spandana and Rajeswar, Sai and Taslakian, Perouz},
  journal={arXiv preprint arXiv:2503.21889},
  year={2025}
}
```

**APA:**
Béchard, P., Wang, C., Abaskohi, A., Rodriguez, J., Pal, C., Vazquez, D., Gella, S., Rajeswar, S., & Taslakian, P. (2025). **StarFlow: Generating Structured Workflow Outputs from Sketch Images**. *arXiv preprint arXiv:2503.21889*.

---

## Glossary

* **FlowSim**: Metric based on tree edit distance for workflows.
* **TreeBLEU**: BLEU-like score using tree structures.
* **Trigger Match**: Correctness of predicted workflow trigger.
* **Component Match**: Correctness of predicted components (order-agnostic).

---

## More Information

* [ServiceNow Flow Designer](https://www.servicenow.com/products/platform-flow-designer.html)
* [StarFlow Blog](https://www.servicenow.com/blogs/2025/starflow-ai-turns-sketches-into-workflows)

---

## The StarFlow Team

* Patrice Béchard, Chao Wang, Amirhossein Abaskohi, Juan Rodriguez, Christopher Pal, David Vazquez, Spandana Gella, Sai Rajeswar, Perouz Taslakian

---

## Model Card Contact

* Patrice Bechard - [patrice.bechard@servicenow.com](mailto:patrice.bechard@servicenow.com)
* ServiceNow Research – [research.servicenow.com](https://research.servicenow.com)