Model Optimizer Changelog
0.19 (2024-10-23)
Backward Breaking Changes
Deprecated the summarize task in the
llm_ptq
example.Deprecated the
type
flag in the huggingface_example.shDeprecated Python plugin support in ONNX.
Support TensorRT-LLM 0.13. Examples not compatible with TensorRT-LLM 0.12.
mtq.auto_quantize
API has been updated. The API now acceptsforward_step
andforward_backward_step
as arguments instead ofloss_func
andcollect_func
. Please see the API documentation for more details.
New Features
ModelOpt is compatbile for SBSA aarch64 (e.g. GH200) now! Except ONNX PTQ with plugins is not supported.
Add
effective_bits
as a constraint formtq.auto_qauntize
.lm_evaluation_harness
is fully integrated to modelopt backed by TensorRT-LLM.lm_evaluation_harness
benchmarks are now available in the examples for LLM accuracy evaluation.A new
--perf
flag is introduced in themodelopt_to_tensorrt_llm.py
example to build engines with max perf.Users can choose the execution provider to run the calibration in ONNX quantization.
Added automatic detection of custom ops in ONNX models using TensorRT plugins. This requires the
tensorrt
python package to be installed.Replaced
jax
withcupy
for faster INT4 ONNX quantization.mtq.auto_quantize
now supports search based automatic quantization for NeMo & MCore models (in addition to HuggingFace models).Add
num_layers
andhidden_size
pruning support for NeMo / Megatron-core models.
0.17 (2024-09-11)
Backward Breaking Changes
Deprecated
torch<2.0
support.modelopt.torch.utils.dataset_utils.get_dataset_dataloader()
now returns a key value pair instead of the tensor.
New Features
New APIs and examples:
modelopt.torch.prune
for pruning Conv, Linear, and Attention heads for NVIDIA Megatron-core GPT-style models (e.g. Llama 3), PyTorch Computer Vision models, and HuggingFace Bert/GPT-J models.New API:
modelopt.torch.distill
for knowledge distillation, along with guides and example.New Example: HF BERT Prune, Distill & Quantize showcasing how to chain pruning, distillation, and quantization to achieve the best performance on a given model.
Added INT8/FP8 DQ-only support for ONNX model.
New API:
modelopt.torch.speculative
for end-to-end support of Medusa models.Added Medusa QAT and End-to-end examples.
Modelopt now supports automatic save/restore of
modelopt_state
with the.save_pretrained
and.from_pretrained
APIs from Huggingface libraries, such astransformers
anddiffusers
. This feature can be enabled by callingmto.enable_huggingface_checkpointing()
.ONNX FP8 quantization support with amax calibration.
TensorRT-LLM dependency upgraded to 0.12.0. Huggingface tokenizer files are now also stored in the engine dir.
The unified model export API
modelopt.torch.export.export_hf_checkpoint
supports exportingfp8
andint4_awq
quantized checkpoints with packed weights for Hugging Face models with namings aligned with its original checkpoints. The exportedfp8
checkpoints can be deployed with both TensorRT-LLM and VLLM.Add int8 and fp8 quantization support for the FLUX.1-dev model.
Add a Python-friendly TensorRT inference pipeline for diffusion models.
Misc
Added deprecation warning for
set_data_parallel_group
andset_tensor_parallel_group
. These APIs are no longer needed for supporting distributed data and tensor parallelism in quantization. They will be removed in a future release.
0.15 (2024-07-25)
Backward Breaking Changes
Deprecated
QuantDescriptor
. UseQuantizerAttributeConfig
to configureTensorQuantizer
.set_from_attribute_config
can be used to set the quantizer attributes from the config class or attribute dictionary. This change applies only to backend APIs. The change is backward compatible if you are using only themtq.quantize
API.
New Features
Added quantization support for torch
RNN, LSTM, GRU
modules. Only available fortorch>=2.0
.modelopt.torch.quantization
now supports module class based quantizer attribute setting formtq.quantize
API.Added new LLM PTQ example for DBRX model.
Added new LLM (Gemma 2) PTQ and TensorRT-LLM checkpoint export support.
Added new LLM QAT example for NVIDIA NeMo framework.
TensorRT-LLM dependency upgraded to 0.11.0.
(Experimental):
mtq.auto_quantize
API which quantizes a model by searching for the best per-layer quantization formats.(Experimental): Added new LLM QLoRA example with NF4 and INT4_AWQ quantization.
(Experimental):
modelopt.torch.export
now supports exporting quantized checkpoints with packed weights for Hugging Face models with namings aligned with its original checkpoints.(Experimental) Added support for quantization of ONNX models with TensorRT plugin.
Misc
Added deprecation warning for
torch<2.0
. Support will be dropped in next release.
0.13 (2024-06-14)
Backward Breaking Changes
PTQ examples have been upgraded to use TensorRT-LLM 0.10.
New Features
Adding TensorRT-LLM checkpoint export support for Medusa decoding (official
MedusaModel
and Megatron CoreGPTModel
).Enable support for mixtral, recurrentgemma, starcoder, qwen in PTQ examples.
Adding TensorRT-LLM checkpoint export and engine building support for sparse models.
Import scales from TensorRT calibration cache and use them for quantization.
(Experimental) Enable low GPU memory FP8 calibration for the Hugging Face models when the original model size does not fit into the GPU memory.
(Experimental) Support exporting FP8 calibrated model to VLLM deployment.
(Experimental) Python 3.12 support added.
0.11 (2024-05-07)
Backward Breaking Changes
[!!!] The package was renamed from
ammo
tomodelopt
. The new full product name is Nvidia TensorRT Model Optimizer. PLEASE CHANGE ALL YOUR REFERENCES FROMammo
tomodelopt
including any paths and links!Default installation
pip install nvidia-modelopt
will now only install minimal core dependencies. Following optional dependencies are available depending on the features that are being used:[deploy], [onnx], [torch], [hf]
. To install all dependencies, usepip install "nvidia-modelopt[all]"
.Deprecated
inference_gpus
arg inmodelopt.torch.export.model_config_export.torch_to_tensorrt_llm_checkpoint
. User should useinference_tensor_parallel
instead.Experimental
modelopt.torch.deploy
module is now available asmodelopt.torch._deploy
.
New Features
modelopt.torch.sparsity
now supports sparsity-aware training (SAT). Both SAT and post-training sparsification supports chaining with other modes, e.g. SAT + QAT.modelopt.torch.quantization
natively support distributed data and tensor parallelism while estimating quantization parameters. The data and tensor parallel groups needs to be registered withmodelopt.torch.utils.distributed.set_data_parallel_group
andmodelopt.torch.utils.distributed.set_tensor_parallel_group
APIs. By default, the data parallel group is set as the default distributed group and the tensor parallel group is disabled.modelopt.torch.opt
now supports chaining multiple optimization techniques that each require modifications to the same model, e.g., you can now sparsify and quantize a model at the same time.modelopt.onnx.quantization
supports FLOAT8 quantization format with Distribution calibration algorithm.Native support of
modelopt.torch.opt
with FSDP (Fully Sharded Data Parallel) fortorch>=2.1
. This includes sparsity, quantization, and any other model modification & optimization.Added FP8 ONNX quantization support in
modelopt.onnx.quantization
.Added Windows (
win_amd64
) support for ModelOpt released wheels. Currently supported formodelopt.onnx
submodule only.
Bug Fixes
Fixed the compatibility issue of
modelopt.torch.sparsity
with FSDP.Fixed an issue in dynamic dim handling in
modelopt.onnx.quantization
with random calibration data.Fixed graph node naming issue after opset convertion operation.
Fixed an issue in negative dim handling like dynamic dim in
modelopt.onnx.quantization
with random calibration data.Fixed allowing to accept
.pb
file for input file.Fixed copy extra data to tmp folder issue for ONNX PTQ.