jetson-inference
TensorRT
Our great sponsors
jetson-inference | TensorRT | |
---|---|---|
11 | 22 | |
7,294 | 9,031 | |
- | 3.6% | |
8.5 | 5.0 | |
about 1 month ago | 16 days ago | |
C++ | C++ | |
MIT License | Apache License 2.0 |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
jetson-inference
-
Can this NVIDIA Jetson Nano handle advanced machine learning tasks?
Jetson Nano’s are obsolete and no longer supported; but to answer your question, this might be a good place to start.
- help with project involving object detection and tracking with camera
-
Jetson Nano 2GB Issues During Training (Out Of Memory / Process Killed) & Other Questions!
I’m trying to do the tutorial, where they retrain the neural network to detect fruits (jetson-inference/pytorch-ssd.md at master · dusty-nv/jetson-inference · GitHub 1)
-
Jetson Nano
Jetson-Inference is another amazing resource to get started on. This will allow you to try out a number of neural networks (classification, detection, and segmentation) all with your own data or with sample images included in the repo.
-
Pretrained image classification model for nuts and bolts (or similar)
Hello! I'm looking for some pre trained image classification models to use on a Jetson Nano. I already know about the model zoo and the pre trained models included in the https://github.com/dusty-nv/jetson-inference repo. For demonstration purposes, however, I need a model trained on small objects from the context of production, ideally nuts, bolts, and similar small objects. Does anyone happen to know a source for this? Thanks a lot!
-
PyTorch 1.8 release with AMD ROCm support
> They provide some SSD-Mobilenet-v2 here: https://github.com/dusty-nv/jetson-inference
I was aware of that repository but from taking a cursory look at it I had thought dusty was just converting models from PyTorch to TensorRT, like here[0, 1]. Am I missing something?
> I get 140 fps on a Xavier NX
That really is impressive. Holy shit.
[0]: https://github.com/dusty-nv/jetson-inference/blob/master/doc...
[1]: https://github.com/dusty-nv/jetson-inference/issues/896#issu...
- NVIDIA DLSS released as a plugin for Unreal Engine 4
-
Help getting started
If you have a screen and keyboard and mouse plugged into the Nano, I would recommend starting with Hello AI World on https://github.com/dusty-nv/jetson-inference#hello-ai-world
-
I'm tired of this anti-Wayland horseshit
Well, don't get me wrong. I do like my Jetson Nano. For a hobbyist who likes to tinker with machine learning in their spare time it's definitely a product cool and there are quite a few repositories on Github[0, 1] with sample code.
Unfortunately… that's about it. There is little documentation about
- how to build a custom OS image (necessary if you're thinking about using Jetson as part of your own product, i.e. a large-scale deployment). What proprietary drivers and libraries do I need to install? Nvidia basically says, here's a Ubuntu image with the usual GUI, complete driver stack and everything – take it or leave it. Unfortunately, the GUI alone is eating up a lot of the precious CPU and GPU resources, so using that OS image is no option.
- how deployment works on production modules (as opposed to the non-production module in the Developer Kit)
- what production modules are available in the first place ("Please refer to our partners")
- what wifi dongles are compatible (the most recent Jetson Nano comes w/o wifi)
- how to convert your custom models to TensorRT, what you need to pay attention to etc. (The official docs basically say: Have a look at the following nondescript sample code. Good luck.)
- … (I'm sure I'm forgetting many other things that I've struggled with over the past months)
Anyway. It's not that this information isn't out there somewhere in some blog post, some Github repo or some thread on the Nvidia forums[2]. (Though I have yet to find a reliably working wifi dongle…) But it usually takes you days orweeks to find it. From a product which is supposed to be industry-grade I would have expected more.
-
Basic Teaching
https://github.com/dusty-nv/jetson-inference#system-setup
TensorRT
-
AMD MI300X 30% higher performance than Nvidia H100, even with optimized stack
> It's not rocket science to implement matrix multiplication in any GPU.
You're right, it's harder. Saying this as someone who's done more work on the former than the latter. (I have, with a team, built a rocket engine. And not your school or backyard project size, but nozzle bigger than your face kind. I've also written CUDA kernels and boy is there a big learning curve to the latter that you gotta fundamentally rethink how you view a problem. It's unquestionable why CUDA devs are paid so much. Really it's only questionable why they aren't paid more)
I know it is easy to think this problem is easy, it really looks that way. But there's an incredible amount of optimization that goes into all of this and that's what's really hard. You aren't going to get away with just N for loops for a tensor rank N. You got to chop the data up, be intelligent about it, manage memory, how you load memory, handle many data types, take into consideration different results for different FMA operations, and a whole lot more. There's a whole lot of non-obvious things that result in high optimization (maybe obvious __after__ the fact, but that's not truthfully "obvious"). The thing is, the space is so well researched and implemented that you can't get away with naive implementations, you have to be on the bleeding edge.
Then you have to do that and make it reasonably usable for the programmer too, abstracting away all of that. Cuda also has a huge head start and momentum is not a force to be reckoned with (pun intended).
Look at TensorRT[0]. The software isn't even complete and it still isn't going to cover all neural networks on all GPUs. I've had stuff work on a V100 and H100 but not an A100, then later get fixed. They even have the "Apple Advantage" in that they have control of the hardware. I'm not certain AMD will have the same advantage. We talk a lot about the difficulties of being first mover, but I think we can also recognize that momentum is an advantage of being first mover. And it isn't one to scoff at.
-
Getting SDXL-turbo running with tensorRT
(python demo_txt2img.py "a beautiful photograph of Mt. Fuji during cherry blossom"). https://github.com/NVIDIA/TensorRT/tree/release/8.6/demo/Diffusion
-
Show HN: Ollama for Linux – Run LLMs on Linux with GPU Acceleration
- https://github.com/NVIDIA/TensorRT
TVM and other compiler-based approaches seem to really perform really well and make supporting different backends really easy. A good friend who's been in this space for a while told me llama.cpp is sort of a "hand crafted" version of what these compilers could output, which I think speaks to the craftmanship Georgi and the ggml team have put into llama.cpp, but also the opportunity to "compile" versions of llama.cpp for other model architectures or platforms.
-
Nvidia Introduces TensorRT-LLM for Accelerating LLM Inference on H100/A100 GPUs
https://github.com/NVIDIA/TensorRT/issues/982
Maybe? Looks like tensorRT does work, but I couldn't find much.
-
Train Your AI Model Once and Deploy on Any Cloud
highly optimized transformer-based encoder and decoder component, supported on pytorch, tensorflow and triton
TensorRT, custom ml framework/ inference runtime from nvidia, https://developer.nvidia.com/tensorrt, but you have to port your models
- A1111 just added support for TensorRT for webui as an extension!
-
WIP - TensorRT accelerated stable diffusion img2img from mobile camera over webrtc + whisper speech to text. Interdimensional cable is here! Code: https://github.com/venetanji/videosd
It uses the nvidia demo code from: https://github.com/NVIDIA/TensorRT/tree/main/demo/Diffusion
-
[P] Get 2x Faster Transcriptions with OpenAI Whisper Large on Kernl
The traditional way to deploy a model is to export it to Onnx, then to TensorRT plan format. Each step requires its own tooling, its own mental model, and may raise some issues. The most annoying thing is that you need Microsoft or Nvidia support to get the best performances, and sometimes model support takes time. For instance, T5, a model released in 2019, is not yet correctly supported on TensorRT, in particular K/V cache is missing (soon it will be according to TensorRT maintainers, but I wrote the very same thing almost 1 year ago and then 4 months ago so… I don’t know).
-
Speeding up T5
I've tried to speed it up with TensorRT and followed this example: https://github.com/NVIDIA/TensorRT/blob/main/demo/HuggingFace/notebooks/t5.ipynb - it does give considerable speedup for batch-size=1 but it does not work with bigger batch sizes, which is useless as I can simply increase the batch-size of HuggingFace model.
- demoDiffusion on TensorRT - supports 3090, 4090, and A100
What are some alternatives?
openpose - OpenPose: Real-time multi-person keypoint detection library for body, face, hands, and foot estimation
DeepSpeed - DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.
onnx-tensorrt - ONNX-TensorRT: TensorRT backend for ONNX
FasterTransformer - Transformer related optimization, including BERT, GPT
tensorflow - An Open Source Machine Learning Framework for Everyone
yolov5-deepsort-tensorrt - A c++ implementation of yolov5 and deepsort
vllm - A high-throughput and memory-efficient inference and serving engine for LLMs
obs-studio - OBS Studio - Free and open source software for live streaming and screen recording
openvino - OpenVINO™ is an open-source toolkit for optimizing and deploying AI inference
trt_pose_hand - Real-time hand pose estimation and gesture classification using TensorRT
stable-diffusion-webui - Stable Diffusion web UI