PyTorch-Guide
TransformerEngine
| PyTorch-Guide | TransformerEngine | |
|---|---|---|
| 2 | 2 | |
| 23 | 1,450 | |
| - | 6.4% | |
| 1.8 | 9.5 | |
| over 2 years ago | 6 days ago | |
| Python | Python | |
| - | 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.
PyTorch-Guide
- Useful Tools and Programs for Deep Learning with PyTorch
-
Cool PyTorch Guide/Wiki
PyTorch Guide/Wiki: https://github.com/mikeroyal/PyTorch-Guide
TransformerEngine
-
Benchmarking Large Language Models on NVIDIA H100 GPUs with CoreWeave (Part 1)
4090 now has its 8-bit float enabled as well, see the [transformer engine issue](https://github.com/NVIDIA/TransformerEngine/issues/15)
-
GPUs for Deep Learning in 2023 – An In-depth Analysis
Would be curious to see your benchmarks. Btw, Nvidia will be providing support for fp8 in a future release of CUDA - https://github.com/NVIDIA/TransformerEngine/issues/15
I think TMA may not matter as much for consumer cards given the disproportionate amount of fp32 / int32 compute that they have.
Would be interesting to see how close to theoretical folks are able to get once CUDA support comes through.
What are some alternatives?
halutmatmul - Hashed Lookup Table based Matrix Multiplication (halutmatmul) - Stella Nera accelerator
Whisper - High-performance GPGPU inference of OpenAI's Whisper automatic speech recognition (ASR) model
NeuralCDE - Code for "Neural Controlled Differential Equations for Irregular Time Series" (Neurips 2020 Spotlight)
autocvd - Tool to automatically set CUDA_VISIBLE_DEVICES based on GPU utilization. Usable from command line and code.
cog - Containers for machine learning
warp-drive - Extremely Fast End-to-End Deep Multi-Agent Reinforcement Learning Framework on a GPU (JMLR 2022)
bittensor - Internet-scale Neural Networks
ivy - The Unified AI Framework
nanoGPT - The simplest, fastest repository for training/finetuning medium-sized GPTs.
fastaudio - 🔊 Audio and fastai v2
liberate-fhe - A Fully Homomorphic Encryption (FHE) library for bridging the gap between theory and practice with a focus on performance and accuracy.
FastFold - Optimizing AlphaFold Training and Inference on GPU Clusters