Skip to content
View dnth's full-sized avatar
:electron:
Making models go 🚀 ⚡
:electron:
Making models go 🚀 ⚡

Block or report dnth

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Please don't include any personal information such as legal names or email addresses. Maximum 100 characters, markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
dnth/README.md

banner

🚀 I make models small, fast, and efficient. 💨

Fullstack computer vision engineer specializing in deploying models on edge devices for real-time inference.


Explore my webpage »
Projects · Blogs · LinkedIn · X · About

LinkedIn X Buy Me A Coffee

⭐ Featured Projects

Supercharge Your PyTorch Image Models

Supercharge Your PyTorch Image Models: Bag of Tricks to 8x Faster Inference with ONNX Runtime & Optimizations.

Accelerate inference speed for PyTorch image models using ONNX Runtime and TensorRT optimizations. Achieve up to 123x speedup over the original PyTorch model on CPU.

📅 September 30, 2024

Supercharge Your PyTorch Image Models

PyTorch at the Edge: Deploying Over 964 TIMM Models on Android with TorchScript and Flutter.

Deploy PyTorch models on Android using TIMM, Fastai, TorchScript, and Flutter. Select a model from TIMM's 900+ models, train with Fastai, export to TorchScript, and create an Android app with Flutter for inference.

📅 February 7, 2023

Supercharge Your PyTorch Image Models

Supercharging YOLOv5: How I Got 182.4 FPS Inference Without a GPU.

Optimize YOLOv5 model for CPU inference using Neural Magic's SparseML and DeepSparse. Train on custom data, apply sparsification techniques like pruning and quantization, and achieve up to 180+ FPS on a CPU with only 4 cores.

📅 June 7, 2022

Supercharge Your PyTorch Image Models

Faster than GPU: How to 10x your Object Detection Model and Deploy on CPU at 50+ FPS.

Optimize a YOLOX object detection model deploy on a CPU. Train with custom data, convert to ONNX and OpenVINO IR formats, and apply post-training quantization. This results in a 10x speed improvement, making real-time inference possible on CPU, even outperforming GPU performance.

📅 April 30, 2022

📝 Featured Blogs

GitHub Trending Developer

I Made It to GitHub Trending - My Open Source Journey

I was listed in GitHub's trending developers list for my open-source work on x.infer, a framework agnostic computer vision inference library. Thank you for supporting my work!

📅 October 28, 2024

Top 2% Scientists

Celebrating a Milestone in the Top 2% of Global Scientists

Honored to be recognized among the top 2% of global scientists by Stanford University in 2023. Reflecting on my 10-year journey from academia to industry in AI/ML.

📅 November 17, 2023

🚀 What I'm Building

x.infer Framework agnostic computer vision inference. Run inference on any 1000+ models with 3 lines of code.
GitHub stars
x.retrieval Evaluate your multimodal retrieval pipeline with any model.
GitHub stars
pgmmr Vector/Hybrid Search & Retrieval on PostgreSQL database your favorite Vision Language Model.
GitHub stars

🛠️ Tech Stack

Deep Learning
Hyperparameter
Optimization
Experiment
Management
Model
Deployment
Hardware
Software
Engineering
Data
Frontend

📈 Github Stats

GitHub Profile Summary
Top Languages by Repo Top Languages by Commit
Stats Commits (UTC +8.00)

❤️ Support Me

Creating free machine learning contents doesn't pay my bills. Support me in creating more free contents like these. Consider buying me a coffee. Your support means a lot to me.

Buy Me A Coffee

Pinned Loading

  1. x.infer x.infer Public

    Framework agnostic computer vision inference. Run 1000+ models by changing only one line of code. Supports models from transformers, timm, ultralytics, vllm, ollama and your custom model.

    Jupyter Notebook 125 12

  2. yolov5-deepsparse-blogpost yolov5-deepsparse-blogpost Public

    By the end of this post, you will learn how to: Train a SOTA YOLOv5 model on your own data. Sparsify the model using SparseML quantization aware training, sparse transfer learning, and one-shot qua…

    Jupyter Notebook 55 13

  3. timm-flutter-pytorch-lite-blogpost timm-flutter-pytorch-lite-blogpost Public

    PyTorch at the Edge: Deploying Over 964 TIMM Models on Android with TorchScript and Flutter.

    Jupyter Notebook 43 5

  4. supercharge-your-pytorch-image-models-blogpost supercharge-your-pytorch-image-models-blogpost Public

    Supercharge Your PyTorch Image Models: Bag of Tricks to 8x Faster Inference with ONNX Runtime & Optimizations

    Jupyter Notebook 20

  5. huggingface-timm-mobile-blogpost huggingface-timm-mobile-blogpost Public

    Bringing High-Quality Image Models to Mobile: Hugging Face TIMM Meets Android & iOS

    Dart 5 4

  6. x.retrieval x.retrieval Public

    Evaluate your multimodal retrieval system with any models and datasets.

    Jupyter Notebook 1