What are ML Models? ML Models in Cyberwave are AI models registered in your workspace that can process various inputs—video, images, audio, text, or robot actions. They integrate with workflows and can run in the cloud or on edge devices.
ML Models define what the model can do (input types) and where it runs (provider). The actual inference happens through the model provider’s API or on your edge device.
Model Capabilities Each ML Model specifies what inputs it can process:
Capability Description Example Use Cases can_take_video_as_inputProcess video streams Surveillance, teleoperation can_take_image_as_inputProcess single images Quality inspection, object detection can_take_audio_as_inputProcess audio data Voice commands, anomaly detection can_take_text_as_inputProcess text prompts Natural language commands can_take_action_as_inputProcess robot actions Behavior cloning, RL policies
Registering a Model Via the SDK from cyberwave import Cyberwave client = Cyberwave() # Register a vision model model = client.api.create_mlmodel({ "name" : "Object Detector" , "description" : "YOLOv8 model for detecting warehouse items" , "model_external_id" : "yolov8n" , "model_provider_name" : "local" , "can_take_image_as_input" : True , "can_take_video_as_input" : True , "visibility" : "private" , "tags" : [ "vision" , "detection" , "warehouse" ] }) print ( f "Model UUID: { model.uuid } " )
Via the Dashboard
Navigate to ML Models in your workspace
Click Add Model
Fill in the model details:
Name and description
External ID (model identifier for the provider)
Provider name (e.g., “openai”, “local”, “huggingface”)
Input capabilities
Click Create
Model Providers Models can run through different providers:
Local / Edge Run on your edge devices using ONNX, TensorRT, or custom inference
Cloud APIs Use OpenAI, Anthropic, or other cloud AI services
Hugging Face Deploy models from Hugging Face Hub
Custom Your own inference servers and endpoints
Using Models in Workflows ML Models integrate with workflow nodes for automated processing:
┌─────────────┐ ┌─────────────┐ ┌─────────────┐ │ Capture │───▶│ ML Model: │───▶│ Condition: │ │ Image │ │ Detect │ │ Found? │ └─────────────┘ │ Objects │ └─────────────┘ └─────────────┘ │ ┌───────┴───────┐ ▼ ▼ ┌───────────┐ ┌───────────┐ │ Pick │ │ Alert │ │ Object │ │ Operator │ └───────────┘ └───────────┘
Example: Vision-Language Model Register a VLM for natural language robot control:
from cyberwave import Cyberwave client = Cyberwave() # Register GPT-4 Vision for robot commands vlm = client.api.create_mlmodel({ "name" : "Robot Commander" , "description" : "Interprets visual scenes and generates robot commands" , "model_external_id" : "gpt-4-vision-preview" , "model_provider_name" : "openai" , "can_take_image_as_input" : True , "can_take_text_as_input" : True , "visibility" : "private" , "metadata" : { "max_tokens" : 1000 , "temperature" : 0.2 } })
Listing Models from cyberwave import Cyberwave client = Cyberwave() # List all models in your workspace models = client.api.list_mlmodels() for model in models: print ( f " { model.name } ( { model.model_provider_name } )" ) print ( f " Video: { model.can_take_video_as_input } " ) print ( f " Image: { model.can_take_image_as_input } " ) print ( f " Text: { model.can_take_text_as_input } " )
Model Visibility Visibility Who Can Access privateOnly your workspace members workspaceAll workspace members publicAnyone (admin-only to create)
Running Inference Cloud Models For cloud-based models, Cyberwave routes requests to the provider:
# Via workflow nodes or API response = client.api.vlm_generation({ "model_uuid" : vlm.uuid, "prompt" : "What objects do you see? How should the robot pick them up?" , "image_url" : "https://..." })
Edge Models For local models, run inference on your edge device:
import asyncio from cyberwave import Cyberwave async def run_local_inference (): client = Cyberwave() # Your local model import torch model = torch.hub.load( 'ultralytics/yolov8' , 'yolov8n' ) twin_uuid = "your-twin-uuid" def on_frame ( frame_data ): """Process each incoming frame.""" # Run detection results = model(frame_data) # Publish results back to Cyberwave detections = results.pandas().xyxy[ 0 ].to_dict() client.mqtt.publish( f "twins/ { twin_uuid } /detections" , detections ) # Subscribe to video frames client.mqtt.subscribe( f "twins/ { twin_uuid } /video/frames" , on_frame ) while True : await asyncio.sleep( 1 ) asyncio.run(run_local_inference())
Next Steps 
