
Guides fine-tuning of NVIDIA Cosmos3 reasoning models with Cosmos Framework or Cosmos-RL
- Open source
- 4K GitHub stars
- Design and media

Guides fine-tuning of NVIDIA Cosmos3 reasoning models with Cosmos Framework or Cosmos-RL

Fine-tunes HuggingFace vision, VLM and LLM models in an NVIDIA GPU container

Runs Kumo-Forecast time series forecasting, DARR retrieval blending and fine-tuning

Runs InternVideo2-CLIP video-text retrieval, fine-tuning, ONNX export and TensorRT deploy

Fine-tunes and runs the PAIDF AnomalyGen pipeline to generate synthetic anomaly images

Computes image embeddings from a parquet of file paths using CLIP, SigLIP or TAO

Auto-labels image-caption pairs with referring expressions and pixel bounding boxes

Generates referring-expression and grounding annotations from images with KITTI bounding

Turns raw videos into chain-of-thought QA training data using VLM and LLM teacher models

Gates TAO model launches behind preflight checks, image confirmation and a review step

Lists what the TAO Skill Bank plugin can do, including models, workflows and AutoML

Mines look-alike source images for VCN augmentation via DEFT embedding

Guides an agent to port a HuggingFace vision model into the NVIDIA TAO Toolkit

Prepares false-negative masks and clean-image inputs for AnomalyGenNext generation runs

Runs container-backed hyperparameter search for NVIDIA TAO networks via AutoMLRunner

Chains AutoML tuning, a DEFT data-improvement loop and a final AutoML pass for PCB AOI

Runs an iterative DEFT loop that improves Cosmos3 bare OK/NG AOI inspection accuracy

Runs an iterative mining-based DEFT loop to improve Cosmos Reason traffic-video

Runs an iterative retrain-and-evaluate loop to improve NVIDIA TAO CLIP image-text

Runs iterative DINOv3 self-supervised data selection and retraining from an unlabeled pool

Runs TAO model checkpoints as local inference microservices with start, query and stop

Lets an agent submit, monitor and cancel TAO GPU training jobs on Kubernetes clusters

Trains, evaluates, exports and runs NVIDIA TAO Deformable DETR object detectors

Trains, evaluates, exports and runs DINO 2D object detectors via TAO in Docker

Continues self-supervised DINOv3 pre-training on unlabeled images to adapt vision

Trains, distills, quantizes and runs RT-DETR real-time object detection models

Runs a standard TAO train, evaluate and export sequence without AutoML loops

Trains, evaluates and exports TAO Sparse4D multi-camera 3D detection and tracking models

Guides an agent through writing and debugging cuEquivariance JAX polynomials and Flax NNX

Guides building and debugging cuEquivariance PyTorch tensor products and equivariant GNNs
See every kind of AI for ai models and llm tooling: tools, agents, MCP servers and skills.Ranked: the best AI for ai models and llm tooling.