t5-pytorch
performer-pytorch
| t5-pytorch | performer-pytorch | |
|---|---|---|
| 1 | 2 | |
| 41 | 1,055 | |
| - | - | |
| 3.1 | 1.8 | |
| 7 months ago | over 2 years ago | |
| Python | Python | |
| MIT License | MIT License |
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.
t5-pytorch
-
[P] T5 Implementation in PyTorch
Link to the repository: https://github.com/conceptofmind/t5-pytorch
performer-pytorch
-
[R] Rotary Positional Embeddings - a new relative positional embedding for Transformers that significantly improves convergence (20-30%) and works for both regular and efficient attention
Performer is the best linear attention variant, but linear attention is just one type of efficient attention solution. I have rotary embeddings already in the repo https://github.com/lucidrains/performer-pytorch and you can witness this phenomenon yourself by toggling it on / off
-
Why has Google's Performer model not replaced traditional softmax attention?
Here's an PyTorch implementation if you want to play around with it: lucidrains/performer-pytorch: An implementation of Performer, a linear attention-based transformer, in Pytorch (github.com)
What are some alternatives?
t5x
long-range-arena - Long Range Arena for Benchmarking Efficient Transformers
text-to-text-transfer-transformer - Code for the paper "Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer"
Perceiver - Implementation of Perceiver, General Perception with Iterative Attention in TensorFlow
RETRO-pytorch - Implementation of RETRO, Deepmind's Retrieval based Attention net, in Pytorch
memory-efficient-attention-pytorch - Implementation of a memory efficient multi-head attention as proposed in the paper, "Self-attention Does Not Need O(n²) Memory"
flamingo-pytorch - Implementation of 🦩 Flamingo, state-of-the-art few-shot visual question answering attention net out of Deepmind, in Pytorch
reformer-pytorch - Reformer, the efficient Transformer, in Pytorch
x-transformers - A simple but complete full-attention transformer with a set of promising experimental features from various papers
vit-pytorch - Implementation of Vision Transformer, a simple way to achieve SOTA in vision classification with only a single transformer encoder, in Pytorch
soundstorm-pytorch - Implementation of SoundStorm, Efficient Parallel Audio Generation from Google Deepmind, in Pytorch
deep-implicit-attention - Implementation of deep implicit attention in PyTorch