COMPAS Lab
- 9 followers
- United States of America
- https://compas.cs.stonybrook.edu/
Popular repositories Loading
-
sparsity-transformers-analyzer
sparsity-transformers-analyzer Publicpython script extracting the attentions from existing BERT-like NLP models and analyzing the sparsity of the multi-head attention.
Jupyter Notebook 1
-
sparsity-compas-forked-kascade
sparsity-compas-forked-kascade PublicForked from microsoft/kascade
Forked from Kascade: A Practical Sparse Attention Method for Long-Context LLM Inference, for the attn out statistical recovery project
Python 1
-
-
sparsity-compas-forked-transformers
sparsity-compas-forked-transformers Publictransformers package forked from huggingface
Python
-
sparsity-nx10-matmul-gidel-proj
sparsity-nx10-matmul-gidel-proj PublicGidel and Quartus project for COMPAS SpMM core
C++
-
sparsity-intel-tensor-core-transformers-accel
sparsity-intel-tensor-core-transformers-accel PublicSpMM core using Intel/Altera Tensor Blocks and Block Aggregation
SystemVerilog
Repositories
- sparsity-intel-tensor-core-transformers-accel Public
SpMM core using Intel/Altera Tensor Blocks and Block Aggregation
- sparsity-compas-forked-kascade Public Forked from microsoft/kascade
Forked from Kascade: A Practical Sparse Attention Method for Long-Context LLM Inference, for the attn out statistical recovery project
- sparsity-transformers-analyzer Public
python script extracting the attentions from existing BERT-like NLP models and analyzing the sparsity of the multi-head attention.
- fstack-riscv-pmd-documentation Public
A high-performance repository implementing a custom Poll Mode Driver (PMD) to bridge DPDK with proprietary RISC-V FPGA hardware. It features a ported FreeBSD network stack via F-Stack, specifically patched for RISC-V's weak memory ordering to enable zero-copy kernel-bypass for user-space applications.
- cqfrs Public
The counting quotient filter is a general purpose counting filter and approximate membership query data structure.
- sparsity-compas-forked-pytorch Public Forked from pytorch/pytorch
Tensors and Dynamic neural networks in Python with strong GPU acceleration, mainly for customizing flex attention
- sparsity-compas-forked-cocotb Public Forked from cocotb/cocotb
cocotb: Python-based chip (RTL) verification (COMPAS fork)
- elider-vSwarm-u Public
Top languages
Loading…
Most used topics
Loading…