PyPI page
Home page
Author:
NVIDIA Corporation
Summary:
Nvidia Model Optimizer: A unified library of SOTA model optimization techniques like quantization, pruning, Neural Architecture Search (NAS), distillation, speculative decoding, etc. It compresses deep learning models for downstream deployment frameworks like TensorRT-LLM, TensorRT, vLLM, etc. to optimize inference speed.
Latest version:
0.47.0
Required dependencies:
autodoc_pydantic
|
cupy-cuda12x
|
deepspeed
|
ninja
|
numpy
|
nvidia-ml-py
|
omegaconf
|
onnxruntime
|
onnxruntime-ep-nv-tensorrt-rtx-cu13
|
onnxruntime-gpu
|
packaging
|
pulp
|
pydantic
|
pyyaml
|
regex
|
rich
|
safetensors
|
scipy
|
setuptools
|
shibuya
|
sphinx
|
sphinx-argparse
|
sphinx-autobuild
|
sphinx-copybutton
|
sphinx-inline-tabs
|
sphinx-togglebutton
|
torch
|
tqdm
Optional dependencies:
accelerate
|
bandit
|
coverage
|
cppimport
|
datasets
|
diffusers
|
fire
|
huggingface_hub
|
hydra-core
|
immutabledict
|
lief
|
lru-dict
|
ml_dtypes
|
mlflow-skinny
|
mypy
|
nltk
|
nox
|
nvidia-modelopt
|
onnx
|
onnx-graphsurgeon
|
onnxconverter-common
|
onnxscript
|
onnxslim
|
pandas
|
peft
|
polygraphy
|
pre-commit
|
psutil
|
pytest
|
pytest-cov
|
pytest-instafail
|
pytest-timeout
|
ruff
|
sentencepiece
|
tiktoken
|
timm
|
torch-geometric
|
torchprofile
|
torchvision
|
transformers
|
typeguard
|
uv
|
wonderwords
Downloads last day:
8,356
Downloads last week:
58,471
Downloads last month:
266,206