graphcore-research Goto Github PK

.github

bess-kge

A PyTorch library for Knowledge Graph Embedding on Graphcore IPUs implementing the distribution framework BESS

flash-attention-ipu

Poplar implementation of FlashAttention for IPU

gfloat

Generic floating-point types in Python

graphcore-research.github.io

graphium-smg

Graphium fork for Scaling Molecular GNNs project at Graphcore

hydronet-gnn

Fast and Accurate Predictions from 3D Molecular Structures

ipu-ray-lib

Path-tracer with Neural HDRI for Graphcore IPUs.

jax-experimental

JAX for Graphcore IPU (experimental)

jit-dynamic-lookup

An experimental dynamic tensor slice operation, using JIT-compiled data exchanges

llama.cpp

LLM inference in C/C++

llm-inference-research

An experimentation platform for LLM inference optimisation

making-efficientnet-more-efficient

Supplementary materials for: Making EfficientNet More Efficient: Exploring Batch-Independent Normalization, Group Convolutions and Reduced Resolution Training

mess

MESS: Modern Electronic Structure Simulations

notebooks

Small, standalone, machine learning & IPU notebooks

out-of-the-box-fp8-training

Demo of the unit_scaling library, showing how a model can be easily adapted to train in FP8.

poptorch-experimental-addons

IPU-specific extensions to PopTorch

pyscf-ipu

PySCF on IPU

pytorch-tensor-tracker

Flexibly track outputs and grad-outputs of torch.nn.Module.

qm1b-dataset

one billion quantum mechanical simulations containing 9-11 heavy atoms

random-bases

Improving Neural Network Training in Low Dimensional Random Bases

sparsity-benchmarks

Benchmarking code for “PopSparse: Accelerated block sparse matrix multiplication on IPU”

tensorflow-jax-experimental

TensorFlow XLA backend of experimental JAX on Mk2

tessellate-ipu

TessellateIPU: low level Poplar tile programming from Python

unit-scaling

A library for unit scaling in PyTorch

unit-scaling-demo

Unit Scaling demo and experimentation code