Speculative Decoding Modules - a nvidia Collection

nvidia 's Collections

NVIDIA Nemotron

Cosmos-Predict2.5

Inference Optimized Checkpoints (with Model Optimizer)

Speculative Decoding Modules

GEN3C

Nemotron-Pre-Training-Dataset

BioNeMo

ViPE

OpenReasoning-Nemotron

Cosmos-Predict2

Cosmos-Transfer2.5

Describe Anything

OpenMathReasoning

OpenCodeReasoning

OpenCodeReasoning-II

Llama Nemotron Feedback-Edit Inference-Time Scaling

Scoring Verifiers

Nemotron-UltraLong

Cosmos-Transfer1

Cosmos-Tokenize1

Cosmos-Predict1

Cosmos-Tokenizer

Llama-3.1-Nemotron-70B

NVILA-Speech-Audio-Setups

QLIP

Cosmos

DMC

AceMath

Eagle

NeMo Audio Codecs

Hymba

Optimized ONNX models for NVIDIA RTX GPUs

Riva

Nemotron 4 340B

SteerLM

Canary

RLHF

Llama3-ChatQA-1.5

SSMs

BigVGAN

PS3: Scaling Vision Pre-Training to 4K Resolution

RADIO

Llama3-ChatQA-2

NeMo Curator - Classifier Models

Speculative Decoding Modules

updated about 12 hours ago

A collection of speculative decoding modules created using TensorRT Model Optimizer.