Nemotron 3.5 Content Safety: Customizable Multimodal Safety for Global Enterprise AI
• 11
AI & ML interests
None defined yet.
Recent Activity
View all activity
Papers
ENPIRE: Agentic Robot Policy Self-Improvement in the Real World
Adaptive Volumetric Mechanical Property Fields Invariant to Resolution
Articles
NVIDIA Cosmos
NVIDIA OmniDreams
NVIDIA OmniDreams model checkpoints and sample datasets.
Nemotron-Labs-Diffusion
A Tri-Mode Language Model Family Unifying Autoregressive, Diffusion, and Self-Speculation Decoding
Nemotron-Labs-Elastic
Cosmos1
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/cosmos3
AnyFlow
Any-Step Video Diffusion Model with On-Policy Flow Map Distillation
Nemotron Vision-Language
Image-text paired datasets for building vision-language models (VLMs).
Nemotron Safety & Content Moderation
Datasets for building safe models with refusals, content moderation, PII detection, agentic safety, and audio safety capabilities.
Nemotron Code & SWE
Datasets for building models that write, debug, and reason about code. Covers competitive programming, software engineering, and code pretraining.
Nemotron Reward Modeling
Human preference data, reward model training sets, and generative reward modeling data for training Nemotron reward models.
NVIDIA Ising
NVIDIA Ising is a new Model Family to enable building useful Quantum Computers with AI.
PixelDiT
NvPanoptix-3D
3D panoptic reconstruction segmentation model
BioNeMo - Design
NVIDIA BioNeMo Models for Design
GR00T-N1.6
NVIDIA Isaac GR00T N1.6 open vision-language-action (VLA) model for generalized humanoid
Cosmos3
Omnimodal World Models for Physical AI
Nemotron Speech
Open, state-of-the-art, production‑ready enterprise speech models from the NVIDIA Speech research team for ASR, TTS, Speaker Diarization and S2S
Nemotron OCR and Object Detection
Steering Reasoning VLAs
Steering Reasoning VLA in robotics manipulation https://www.arxiv.org/abs/2510.16281
Cosmos2
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/cosmos3
Nemotron-Post-Training-v3
Collection of datasets used in the post-training phase of Nemotron Nano, Super, and Ultra v3.
Nemotron RAG
Set of tools to build retrieval-augmented generation (RAG) systems, improve search and ranking accuracy, and extract structured data from complex docs
Cosmos Policy
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/c
NVIDIA Nemotron V2
Open, Production-ready Enterprise Models. Nvidia Open Model license.
ChronoEdit
ChronoEdit: Towards Temporal Reasoning for Image Editing and World Simulation
Cosmos-Drive-Dreams
A collection of tokenizers, diffusion models, and datasets relevant to the cosmos-drive-dreams platform.
Llama Nemotron
Open, Production-ready Enterprise Models
BioNeMo - Understand
NVIDIA BioNeMo Models for Understanding Biology
Cosmos-Predict2.5
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/cosmos3
Nemotron-Personas
A collection of multilingual, region-specific synthetic persona datasets that support sovereign AI development across many countries and regions.
GEN3C
3D-Informed World-Consistent Video Generation with Precise Camera Control
BioNeMo
Accelerated models for digital biology by the NVIDIA BioNeMo team. https://www.nvidia.com/en-us/clara/biopharma/
OpenReasoning-Nemotron
Collection of models for OpenReasoning-Nemotron which are trained on 5M reasoning traces for Math, Code and Science.
Audio2Face-3D
Open-weight Audio2Face-3D and Audio2Emotion networks and a sample dataset for training and evaluation
Cosmos-Transfer2.5
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/cosmos3
Nemotron-H
Mamba-Transformer hybrid models
Describe Anything
Multimodal Large Language Models for Detailed Localized Image and Video Captioning
OpenMathReasoning
Models and datasets from "AIMO-2 Winning Solution: Building State-of-the-Art Mathematical Reasoning Models with OpenMathReasoning dataset"
OpenCodeReasoning-II
Reasoning data for supervised finetuning of LLMs to advance code generation and critique
Scoring Verifiers
Benchmarks for evaluating synthetic verifiers like test case generation and code reward models (as found in https://www.arxiv.org/abs/2502.13820).
Cosmos-Reason1
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/cosmos3
Cosmos-Predict1
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/cosmos-predict2
Llama-3.1-Nemotron-70B
SOTA models on Arena Hard and RewardBench as of 1 Oct 2024.
QLIP
QLIP is a family of image tokenizers with SOTA reconstruction quality and zero-shot image understanding.
DMC
LLMs equipped with Dynamic Memory Compression to accelerate generation.
NemoGuard
Essential datasets and models for content safety, topic-following, and security guardrails
NeMo Audio Codecs
A series of Neural Audio Codecs
Optimized ONNX models for NVIDIA RTX GPUs
Collection of optimized ONNX model checkpoints for NVIDIA RTX GPUs
OpenMath-2
A collection of models and datasets introduced in "OpenMathInstruct-2: Accelerating AI for Math with Massive Open-Source Instruction Data"
SteerLM
A collection of models and datasets relating to SteerLM and HelpSteer.
Canary ASR/AST
A collection of multilingual and multitask speech to text models from NVIDIA NeMo 🐤
OpenMath
A collection of models and datasets introduced in "OpenMathInstruct-1: A 1.8 Million Math Instruction Tuning Dataset"
NV-Embed
NV-Embed is a generalist embedding model encompassing retrieval, reranking, classification, clustering, STS tasks.
SSMs
A collection of Mamba-2-based research models with 8B parameters trained on 3.5T tokens for comparison with Transformers.
BigVGAN
BigVGAN is a universal neural vocoder that generates audio waveform using mel spectrogram as input.
PS3: Scaling Vision Pre-Training to 4K Resolution
Enabling 4k resolution for VLMs, CVPR 2025, https://nvlabs.github.io/PS3/
RADIO
A collection of Foundation Vision Models that combine multiple models (CLIP, DINOv2, SAM, etc.).
NeMo Curator - Classifier Models
Classifier models that can be used in NeMo Curator for labelling/filtering datasets.
NVIDIA Nemotron v3
Open, Production-ready Enterprise Models
Open-SWE-Traces
Open-SWE-Traces: Advancing Dual-Mode Multilingual Distillation for Software Engineering Agents
Inference Optimized Checkpoints (with Model Optimizer)
A collection of generative models quantized and optimized for inference with Model Optimizer.
swe-zero-to-swe-hero
Datasets and Models for SWE-ZERO to SWE-HERO paper (https://arxiv.org/abs/2604.01496)
Efficient-DLM
Nemotron Supervised Fine-Tuning
SFT datasets covering math, code, chat, safety, agentic, VLM, multilingual, and specialized domains.
Nemotron Agentic & Tool-Use
Datasets for building models capable of function calling, multi-step agentic tasks, terminal use, and SWE workflows.
Nemotron Chat & Instruction Following
Datasets for building helpful, multi-turn, instruction-following conversational models across single and multi-turn settings.
Nemotron Math & Reasoning
Datasets for building models that excel at math reasoning, proofs, and quantitative problem-solving. Covers SFT, RL, and pretraining data.
Lyra
Project Lyra: Open Generative 3D World Models
NVIDIA EGM
Efficient Grounding Models
Nemotron-Cascade 2
Post-Training LLMs with Cascade RL and Multi-Domain On-Policy Distillation
Kimodo-v1
Models for human(oid) motion generation
GR00T-N1.7
NVIDIA Isaac GR00T N1.7 open vision-language-action (VLA) model for generalized humanoid
MedTech Open Models
Open models for physical AI and medical imaging — robot control, surgical simulation, segmentation, reconstruction, generation, and reasoning.
Nemotron-Terminal
We are releasing Nemotron-Terminal models and training datasets.
Speculative Decoding Modules
A collection of speculative decoding modules created using Model Optimizer.
Nemotron ColEmbed V2
State-of-the-Art Late Interaction Vision-Language Embedding Models
Earth-2
Open, state of the art models for Climate and Weather forecasting. Nowcasting, Medium range, S2S range, Downscaling.
Nemotron-Cascade
Scaling Cascaded Reinforcement Learning for General-Purpose Reasoning Models
Nemotron-Pre-Training-Datasets
Large scale pre-training datasets used in the Nemotron family of models.
NeMo Gym
Collection of RL verifiable data for NeMo Gym
KVzap
Alpamayo
A collection related to the Alpamayo ecosystem, containing Reasoning VLA models, Physical AI data, simulation frameworks, training utilities, and more
PhysicsNeMo
Framework of PyTorch composable modules for developing physics guided machine learning training pipelines. https://github.com/NVIDIA/physicsnemo
-
Earth2 Inference Demo
🚀10Visualize weather forecasts for any date and time range
-
DrivAerML Aero Surrogates Demo
🏎5Predict and visualize car surface pressure and shear stress
-
DoMINO with Ahmed Body Dataset - Multi-Scale Neural Operator for CFD
🟢4Access JupyterLab for interactive coding
-
Modeling Magnetohydrodynamics with PhysicsNeMo
🟢4Access JupyterLab for interactive coding
Reward Models 10-2025
A collection of great reward models for research and production
Clara Medical
NVIDIA Clara Open Models for medical imaging AI: segment, generate, and reason across CT, MRI, and X-ray. Built on MONAI by NVIDIA.
BioNeMo - Optimize
NVIDIA BioNeMo Models for Optimization
Cosmos-Reason2
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/cosmos3
Llama-Embed-Nemotron-8B
State-of-the-Art Text Embedding Model
Reasoning Efficiency Research
Ultra-efficient reasoning model! SOTA Accuracy / CoT Length trade-offs
ViPE
Cosmos-Predict2
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/cosmos-predict25
Reward Models 06-2025
Nemotron reward models. For use in RLHF pipelines and LLM-as-a-Judge
AceReason
Math and Code reasoning model trained through reinforcement learning (RL)
Cosmos-Embed1
Joint video-text embedding for physical AI
AceMath-RL
Math reasoning models trained through reinforcement learning (RL)
OpenCodeReasoning
Reasoning data for supervised finetuning of LLMs to advance data distillation for competitive coding
Llama Nemotron Feedback-Edit Inference-Time Scaling
Novel ITS approach for open-ended tasks - No. 1 on Arena Hard on 18 Mar 2025
Nemotron-UltraLong
Cosmos-Transfer1
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/cosmos3
Cosmos-Tokenizer1
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/cosmos3
Physical AI
Collection of open, commercial-grade datasets for physical AI developers
Cosmos-Preidct1
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/cosmos3
AceMath
We are releasing math instruction models, math reward models, general instruction models, all training datasets, and a math reward benchmark.
Eagle
Eagle is a family of frontier vision-language models with data-centric strategies. The model supports both HD image and long-context video input.
Hymba
A series of Hybrid Small Language Models.
NVLM 1.0
A family of frontier-class multimodal large language models (LLMs) that achieve state-of-the-art results on vision-language tasks and text-only tasks.
Nemotron 4 340B
Nemotron-4: open models for Synthetic Data Generation (SDG). Includes Base, Instruct, and Reward models.
Parakeet ASR
NeMo Parakeet ASR Models attain strong speech recognition accuracy while being efficient for inference. Available in CTC and RNN-Transducer variants.
InstructRetro
InstructRetro is an autoregressive decoder-only language model (LM) with retrieval-augmented pretraining and instruction tuning.
RLHF
A collection of models trained with Reinforcement Learning from Human Feedback (RLHF).
Llama3-ChatQA-1.5
Llama3-ChatQA-1.5 models excel at conversational question answering (QA) and retrieval-augmented generation (RAG).
Nemotron 3 8B
The Nemotron 3 8B Family of models is optimized for building production-ready generative AI applications for the enterprise.
MambaVision
MambaVision: A Hybrid Mamba-Transformer Vision Backbone. Includes both 1K and 21K pretrained models.
Minitron
A family of compressed models obtained via pruning and knowledge distillation
Llama3-ChatQA-2
This is the collection that presents ChatQA-2, a suite of 128K long-context models, that also have exceptional RAG capabilities
Nemotron v3 Pre-Training
Large scale pre-training datasets used in the Nemotron family of models.
NVIDIA Cosmos
NVIDIA Nemotron v3
Open, Production-ready Enterprise Models
NVIDIA OmniDreams
NVIDIA OmniDreams model checkpoints and sample datasets.
Open-SWE-Traces
Open-SWE-Traces: Advancing Dual-Mode Multilingual Distillation for Software Engineering Agents
Nemotron-Labs-Diffusion
A Tri-Mode Language Model Family Unifying Autoregressive, Diffusion, and Self-Speculation Decoding
Inference Optimized Checkpoints (with Model Optimizer)
A collection of generative models quantized and optimized for inference with Model Optimizer.
Nemotron-Labs-Elastic
swe-zero-to-swe-hero
Datasets and Models for SWE-ZERO to SWE-HERO paper (https://arxiv.org/abs/2604.01496)
Cosmos1
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/cosmos3
Efficient-DLM
AnyFlow
Any-Step Video Diffusion Model with On-Policy Flow Map Distillation
Nemotron Supervised Fine-Tuning
SFT datasets covering math, code, chat, safety, agentic, VLM, multilingual, and specialized domains.
Nemotron Vision-Language
Image-text paired datasets for building vision-language models (VLMs).
Nemotron Agentic & Tool-Use
Datasets for building models capable of function calling, multi-step agentic tasks, terminal use, and SWE workflows.
Nemotron Safety & Content Moderation
Datasets for building safe models with refusals, content moderation, PII detection, agentic safety, and audio safety capabilities.
Nemotron Chat & Instruction Following
Datasets for building helpful, multi-turn, instruction-following conversational models across single and multi-turn settings.
Nemotron Code & SWE
Datasets for building models that write, debug, and reason about code. Covers competitive programming, software engineering, and code pretraining.
Nemotron Math & Reasoning
Datasets for building models that excel at math reasoning, proofs, and quantitative problem-solving. Covers SFT, RL, and pretraining data.
Nemotron Reward Modeling
Human preference data, reward model training sets, and generative reward modeling data for training Nemotron reward models.
Lyra
Project Lyra: Open Generative 3D World Models
NVIDIA Ising
NVIDIA Ising is a new Model Family to enable building useful Quantum Computers with AI.
NVIDIA EGM
Efficient Grounding Models
PixelDiT
Nemotron-Cascade 2
Post-Training LLMs with Cascade RL and Multi-Domain On-Policy Distillation
NvPanoptix-3D
3D panoptic reconstruction segmentation model
Kimodo-v1
Models for human(oid) motion generation
BioNeMo - Design
NVIDIA BioNeMo Models for Design
GR00T-N1.7
NVIDIA Isaac GR00T N1.7 open vision-language-action (VLA) model for generalized humanoid
GR00T-N1.6
NVIDIA Isaac GR00T N1.6 open vision-language-action (VLA) model for generalized humanoid
MedTech Open Models
Open models for physical AI and medical imaging — robot control, surgical simulation, segmentation, reconstruction, generation, and reasoning.
Cosmos3
Omnimodal World Models for Physical AI
Nemotron-Terminal
We are releasing Nemotron-Terminal models and training datasets.
Nemotron Speech
Open, state-of-the-art, production‑ready enterprise speech models from the NVIDIA Speech research team for ASR, TTS, Speaker Diarization and S2S
Speculative Decoding Modules
A collection of speculative decoding modules created using Model Optimizer.
Nemotron OCR and Object Detection
Nemotron ColEmbed V2
State-of-the-Art Late Interaction Vision-Language Embedding Models
Steering Reasoning VLAs
Steering Reasoning VLA in robotics manipulation https://www.arxiv.org/abs/2510.16281
Earth-2
Open, state of the art models for Climate and Weather forecasting. Nowcasting, Medium range, S2S range, Downscaling.
Cosmos2
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/cosmos3
Nemotron-Cascade
Scaling Cascaded Reinforcement Learning for General-Purpose Reasoning Models
Nemotron-Post-Training-v3
Collection of datasets used in the post-training phase of Nemotron Nano, Super, and Ultra v3.
Nemotron-Pre-Training-Datasets
Large scale pre-training datasets used in the Nemotron family of models.
Nemotron RAG
Set of tools to build retrieval-augmented generation (RAG) systems, improve search and ranking accuracy, and extract structured data from complex docs
NeMo Gym
Collection of RL verifiable data for NeMo Gym
Cosmos Policy
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/c
KVzap
NVIDIA Nemotron V2
Open, Production-ready Enterprise Models. Nvidia Open Model license.
Alpamayo
A collection related to the Alpamayo ecosystem, containing Reasoning VLA models, Physical AI data, simulation frameworks, training utilities, and more
ChronoEdit
ChronoEdit: Towards Temporal Reasoning for Image Editing and World Simulation
PhysicsNeMo
Framework of PyTorch composable modules for developing physics guided machine learning training pipelines. https://github.com/NVIDIA/physicsnemo
-
Earth2 Inference Demo
🚀10Visualize weather forecasts for any date and time range
-
DrivAerML Aero Surrogates Demo
🏎5Predict and visualize car surface pressure and shear stress
-
DoMINO with Ahmed Body Dataset - Multi-Scale Neural Operator for CFD
🟢4Access JupyterLab for interactive coding
-
Modeling Magnetohydrodynamics with PhysicsNeMo
🟢4Access JupyterLab for interactive coding
Cosmos-Drive-Dreams
A collection of tokenizers, diffusion models, and datasets relevant to the cosmos-drive-dreams platform.
Reward Models 10-2025
A collection of great reward models for research and production
Llama Nemotron
Open, Production-ready Enterprise Models
Clara Medical
NVIDIA Clara Open Models for medical imaging AI: segment, generate, and reason across CT, MRI, and X-ray. Built on MONAI by NVIDIA.
BioNeMo - Understand
NVIDIA BioNeMo Models for Understanding Biology
BioNeMo - Optimize
NVIDIA BioNeMo Models for Optimization
Cosmos-Predict2.5
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/cosmos3
Cosmos-Reason2
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/cosmos3
Nemotron-Personas
A collection of multilingual, region-specific synthetic persona datasets that support sovereign AI development across many countries and regions.
Llama-Embed-Nemotron-8B
State-of-the-Art Text Embedding Model
GEN3C
3D-Informed World-Consistent Video Generation with Precise Camera Control
Reasoning Efficiency Research
Ultra-efficient reasoning model! SOTA Accuracy / CoT Length trade-offs
BioNeMo
Accelerated models for digital biology by the NVIDIA BioNeMo team. https://www.nvidia.com/en-us/clara/biopharma/
ViPE
OpenReasoning-Nemotron
Collection of models for OpenReasoning-Nemotron which are trained on 5M reasoning traces for Math, Code and Science.
Cosmos-Predict2
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/cosmos-predict25
Audio2Face-3D
Open-weight Audio2Face-3D and Audio2Emotion networks and a sample dataset for training and evaluation
Reward Models 06-2025
Nemotron reward models. For use in RLHF pipelines and LLM-as-a-Judge
Cosmos-Transfer2.5
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/cosmos3
AceReason
Math and Code reasoning model trained through reinforcement learning (RL)
Nemotron-H
Mamba-Transformer hybrid models
Cosmos-Embed1
Joint video-text embedding for physical AI
Describe Anything
Multimodal Large Language Models for Detailed Localized Image and Video Captioning
AceMath-RL
Math reasoning models trained through reinforcement learning (RL)
OpenMathReasoning
Models and datasets from "AIMO-2 Winning Solution: Building State-of-the-Art Mathematical Reasoning Models with OpenMathReasoning dataset"
OpenCodeReasoning
Reasoning data for supervised finetuning of LLMs to advance data distillation for competitive coding
OpenCodeReasoning-II
Reasoning data for supervised finetuning of LLMs to advance code generation and critique
Llama Nemotron Feedback-Edit Inference-Time Scaling
Novel ITS approach for open-ended tasks - No. 1 on Arena Hard on 18 Mar 2025
Scoring Verifiers
Benchmarks for evaluating synthetic verifiers like test case generation and code reward models (as found in https://www.arxiv.org/abs/2502.13820).
Nemotron-UltraLong
Cosmos-Reason1
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/cosmos3
Cosmos-Transfer1
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/cosmos3
Cosmos-Predict1
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/cosmos-predict2
Cosmos-Tokenizer1
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/cosmos3
Llama-3.1-Nemotron-70B
SOTA models on Arena Hard and RewardBench as of 1 Oct 2024.
Physical AI
Collection of open, commercial-grade datasets for physical AI developers
QLIP
QLIP is a family of image tokenizers with SOTA reconstruction quality and zero-shot image understanding.
Cosmos-Preidct1
⚠️ This collection is archived.
👉 https://huggingface.co/collections/nvidia/cosmos3
DMC
LLMs equipped with Dynamic Memory Compression to accelerate generation.
AceMath
We are releasing math instruction models, math reward models, general instruction models, all training datasets, and a math reward benchmark.
NemoGuard
Essential datasets and models for content safety, topic-following, and security guardrails
Eagle
Eagle is a family of frontier vision-language models with data-centric strategies. The model supports both HD image and long-context video input.
NeMo Audio Codecs
A series of Neural Audio Codecs
Hymba
A series of Hybrid Small Language Models.
Optimized ONNX models for NVIDIA RTX GPUs
Collection of optimized ONNX model checkpoints for NVIDIA RTX GPUs
NVLM 1.0
A family of frontier-class multimodal large language models (LLMs) that achieve state-of-the-art results on vision-language tasks and text-only tasks.
OpenMath-2
A collection of models and datasets introduced in "OpenMathInstruct-2: Accelerating AI for Math with Massive Open-Source Instruction Data"
Nemotron 4 340B
Nemotron-4: open models for Synthetic Data Generation (SDG). Includes Base, Instruct, and Reward models.
SteerLM
A collection of models and datasets relating to SteerLM and HelpSteer.
Parakeet ASR
NeMo Parakeet ASR Models attain strong speech recognition accuracy while being efficient for inference. Available in CTC and RNN-Transducer variants.
Canary ASR/AST
A collection of multilingual and multitask speech to text models from NVIDIA NeMo 🐤
InstructRetro
InstructRetro is an autoregressive decoder-only language model (LM) with retrieval-augmented pretraining and instruction tuning.
OpenMath
A collection of models and datasets introduced in "OpenMathInstruct-1: A 1.8 Million Math Instruction Tuning Dataset"
RLHF
A collection of models trained with Reinforcement Learning from Human Feedback (RLHF).
NV-Embed
NV-Embed is a generalist embedding model encompassing retrieval, reranking, classification, clustering, STS tasks.
Llama3-ChatQA-1.5
Llama3-ChatQA-1.5 models excel at conversational question answering (QA) and retrieval-augmented generation (RAG).
SSMs
A collection of Mamba-2-based research models with 8B parameters trained on 3.5T tokens for comparison with Transformers.
Nemotron 3 8B
The Nemotron 3 8B Family of models is optimized for building production-ready generative AI applications for the enterprise.
BigVGAN
BigVGAN is a universal neural vocoder that generates audio waveform using mel spectrogram as input.
MambaVision
MambaVision: A Hybrid Mamba-Transformer Vision Backbone. Includes both 1K and 21K pretrained models.
PS3: Scaling Vision Pre-Training to 4K Resolution
Enabling 4k resolution for VLMs, CVPR 2025, https://nvlabs.github.io/PS3/
Minitron
A family of compressed models obtained via pruning and knowledge distillation
RADIO
A collection of Foundation Vision Models that combine multiple models (CLIP, DINOv2, SAM, etc.).
Llama3-ChatQA-2
This is the collection that presents ChatQA-2, a suite of 128K long-context models, that also have exceptional RAG capabilities
NeMo Curator - Classifier Models
Classifier models that can be used in NeMo Curator for labelling/filtering datasets.
Nemotron v3 Pre-Training
Large scale pre-training datasets used in the Nemotron family of models.