facebookresearch / diffq
DiffQ performs differentiable quantization using pseudo quantization noise. It can automatically tune the number of bits used per weight or group of weights, in order to achieve a given trade-off between model size and accuracy.
☆234Updated last year
Related projects ⓘ
Alternatives and complementary repositories for diffq
- Implementation of a Transformer, but completely in Triton☆248Updated 2 years ago
- Official code for "Distributed Deep Learning in Open Collaborations" (NeurIPS 2021)☆116Updated 2 years ago
- Implementation of fused cosine similarity attention in the same style as Flash Attention☆207Updated last year
- Memory Efficient Attention (O(sqrt(n)) for Jax and PyTorch☆179Updated last year
- Implementation of the specific Transformer architecture from PaLM - Scaling Language Modeling with Pathways - in Jax (Equinox framework)☆185Updated 2 years ago
- torch::deploy (multipy for non-torch uses) is a system that lets you get around the GIL problem by running multiple Python interpreters i…☆176Updated this week
- Accelerate PyTorch models with ONNX Runtime☆356Updated 2 months ago
- Simple and efficient RevNet-Library for PyTorch with XLA and DeepSpeed support and parameter offload☆124Updated 2 years ago
- Library for 8-bit optimizers and quantization routines.☆714Updated 2 years ago
- A library to inspect and extract intermediate layers of PyTorch models.☆470Updated 2 years ago
- graftr: an interactive shell to view and edit PyTorch checkpoints.☆110Updated 4 years ago
- ☆268Updated this week
- Named tensors with first-class dimensions for PyTorch☆322Updated last year
- A library that contains a rich collection of performant PyTorch model metrics, a simple interface to create new metrics, a toolkit to fac…☆216Updated this week
- PyTorch implementation of L2L execution algorithm☆106Updated last year
- Implementation of the Adan (ADAptive Nesterov momentum algorithm) Optimizer in Pytorch☆247Updated 2 years ago
- Amos optimizer with JEstimator lib.☆81Updated 6 months ago
- Swarm training framework using Haiku + JAX + Ray for layer parallel transformer language models on unreliable, heterogeneous nodes☆237Updated last year
- Fast Block Sparse Matrices for Pytorch☆545Updated 3 years ago
- Torch Distributed Experimental☆116Updated 3 months ago
- Butterfly matrix multiplication in PyTorch☆164Updated last year
- Context Manager to profile the forward and backward times of PyTorch's nn.Module☆83Updated last year
- Approximate nearest neighbor search with product quantization on GPU in pytorch and cuda☆214Updated 11 months ago
- Framework for writing deep learning training loops. Lightweight, and retaining full freedom to design as you see fits. It handles checkpo…☆103Updated 8 months ago
- [Prototype] Tools for the concurrent manipulation of variably sized Tensors.☆253Updated 2 years ago
- Block-sparse primitives for PyTorch☆148Updated 3 years ago
- Implementation of a memory efficient multi-head attention as proposed in the paper, "Self-attention Does Not Need O(n²) Memory"☆362Updated last year
- Learning to Initialize Neural Networks for Stable and Efficient Training☆136Updated 2 years ago
- Sequence modeling with Mega.☆298Updated last year
- Implementation of Flash Attention in Jax☆196Updated 8 months ago