Mixtrain Documentation
The post-training platform for specialized multimodal models. Curate and manage datasets, train models, and run evaluations across image, video, audio, 3D, and text.
Getting Started
Quickstart
Get started with the SDK and CLI
Authentication
Configure API keys and credentials
Examples
Complete examples of models and workflows on Mixtrain
The Model Lifecycle
Datasets
View and explore multimodal data
Models
Run, train, and fine-tune AI models
Evaluations
Test and evaluate model outputs
Platform
Workflows
Build reusable ML pipelines
Routing
Route requests across models
Types
Common types used across Mixtrain
Open Source Models
Physical AI
GR00T N1.6, pi0.5, SmolVLA
Vision
SmolVLM 2, SigLIP 2, SAM 2.1
Audio & Speech
Parakeet TDT v3, Canary Qwen, Whisper