Connect LMM to other blocks to build a custom workflow
Ask a question to a Large Multimodal Model (LMM) with an image and text.
You can specify arbitrary text prompts to an LMMBlock.
The LLMBlock supports two LMMs:
- OpenAI's GPT-4 with Vision, and;
- CogVLM.
You need to provide your OpenAI API key to use the GPT-4 with Vision model. You do not
need to provide an API key to use CogVLM.
_If you want to classify an image into one or more categories, we recommend using the
dedicated LMMForClassificationBlock._
Connect pre-trained models, open source models, LLM APIs, advanced logic, and external applications. Deploy as an API endpoint, on-prem, or at the edge.
Learn how to use a low-code open source platform to simplify building and deploying vision AI applications.
1
Choose a Block
Choose from 40+ pre-built blocks that let you use custom models, open source models, LLM APIs, pre-built logic, and external applications. Blocks can be models from OpenAI or Meta AI, applications like Google Sheets or Pager Duty, and logic like filtering or cropping.
2
Connect Blocks
Each block can receive inputs, execute code, and send outputs to the next block in your Workflow. You can use the drag-and-drop UI to configure connections and see the JSON definitions of what’s happening behind the scenes.
3
Deploy Workflows
You’ll receive an output of the final result from your Workflow and the format you want it delivered in, like JSON. Once your Workflow produces sufficient results, you can use the Workflow as a hosted API endpoint or self-host in your own cloud, on-prem, or at the edge.
Roboflow powers millions of daily inferences for the world’s largest enterprises on-device and in the cloud
Deploy your Workflows directly on fully managed infrastructure through an infinitely-scalable API endpoint for high volume workloads
Run Workflows on-device, internet connection optional, without the headache of environment management, dependencies, and managing CUDA versions.
Isolate dependencies in your software by using the Python SDK or HTTP API to operate and maintain your Workflows separate from other logic within your codebase
Supported devices include ARM CPU, x86 CPU, NVIDIA GPU, and NVIDIA Jetson
Customize Your Pipeline
Connect models from OpenAI or Meta AI, applications like Slack or Pager Duty, and logic like filtering or cropping.
