Pinned Repositories
RWKV-LM
RWKV is an RNN with transformer-level LLM performance. It can be directly trained like a GPT (parallelizable). So it's combining the best of RNN and transformer - great performance, fast inference, saves VRAM, fast training, "infinite" ctx_len, and free sentence embedding.
rwkv-cpp-accelerated
A torchless, c++ rwkv implementation using 8bit quantization, written in cuda/hip/vulkan for maximum compatibility and minimum dependencies
linear-attention-transformer
Transformer based on a variant of attention that is linear complexity in respect to sequence length
rwkv.cpp
INT4/INT5/INT8 and FP16 inference on CPU for RWKV language model
DTW_co_processor_front_end
A pipelined DTW accelerator
EDAProject
FastSV-display
Show the advantages of FastSV in the context of Parallel computing
scatter_gather_aes_cuda
A High-Performance Side-Channel-Resistant AES on GPUs
slate-ipdps19
This serves as reproducibility for "Slate: Enabling Workload-Aware Efficient Multiprocessing for Modern GPGPUs" at IPDPS 2019.
SPICE_Programmable_Extension
This repository explores the possibility of making SPICE more programable.
zzczzc20's Repositories
zzczzc20/EDAProject
zzczzc20/FastSV-display
Show the advantages of FastSV in the context of Parallel computing
zzczzc20/DTW_co_processor_front_end
A pipelined DTW accelerator
zzczzc20/scatter_gather_aes_cuda
A High-Performance Side-Channel-Resistant AES on GPUs
zzczzc20/slate-ipdps19
This serves as reproducibility for "Slate: Enabling Workload-Aware Efficient Multiprocessing for Modern GPGPUs" at IPDPS 2019.
zzczzc20/SPICE_Programmable_Extension
This repository explores the possibility of making SPICE more programable.