bryanzhang

i-mdShanghai

Pinned Repositories

transformers-flashattention
🤗 Transformers: State-of-the-art Machine Learning for Pytorch, TensorFlow, and JAX.
Language:Python10
ufbp
Udp File Broadcasting/Multicasting Protocol
Language:C++10
bert
TensorFlow code and pre-trained models for BERT
Language:Python00
books
00
c3p0
a mature, highly concurrent JDBC Connection pooling library, with support for caching and reuse of PreparedStatements.
Language:Java00
cabinet
00
cook
a build tool try to cook everything
Language:Scala00
cpu_bitonicsort
Language:C++00
cuda_matmul
Language:Cuda00
cuda_reduce
Language:Cuda00

bryanzhang's Repositories

bryanzhang/LOBS_v2
Language:Python
bryanzhang/layernorm_cpu
Language:C++
bryanzhang/cuda_softmax
Language:Cuda
bryanzhang/hessian_torch_kernels
Boosted sparse hessian(or it pseduo-inverse) matmul implementation.
Language:Python
bryanzhang/LOBS_v1
layerwise optimal brain surgeon v1
Language:Python
bryanzhang/LOBS_v0
Language:Python
bryanzhang/cpu_bitonicsort
Language:C++
bryanzhang/cuda_matmul
Language:Cuda
bryanzhang/high_order_grads
求解高阶导数
Language:Python
bryanzhang/optimal-brain-damage
Implementation of the Optimal Brain Damage paper in PyTorch (using second order Taylor series to prune neural networks)
Language:Python
bryanzhang/obd_V2
optimal brain damage pytorch implement v2
Language:Python
bryanzhang/obd
optimal brain damage pytorch implement
Language:Python
bryanzhang/cuda_reduce
Language:Cuda
bryanzhang/cuda_transpose
Language:Cuda
bryanzhang/transpose
矩阵转置优化
Language:C++
bryanzhang/triton_fusedattention
使用triton实现FlashAttention V2推理功能(FP16).
Language:Python
bryanzhang/simple_skip_list
简单的跳表实现.
bryanzhang/matmul
Language:Cuda
bryanzhang/transformers-flashattention
🤗 Transformers: State-of-the-art Machine Learning for Pytorch, TensorFlow, and JAX.
Language:Python1
bryanzhang/smoothquant
SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Models
Language:Python
bryanzhang/FasterTransformer
Transformer related optimization, including BERT, GPT
bryanzhang/fastertransformer_backend
用于调试观察
Language:Jupyter Notebook
bryanzhang/threadstacks
ThreadStacks can be used to inspect thread stacktraces of live C/C++ processes.
Language:C++
bryanzhang/libbacktrace
A C library that may be linked into a C/C++ program to produce symbolic backtraces
Language:C
bryanzhang/gptj-chatbot
bryanzhang/kongmingqi
孔明棋强化学习
Language:Python
bryanzhang/bert
TensorFlow code and pre-trained models for BERT
Language:Python
bryanzhang/books
bryanzhang/ufbp
Udp File Broadcasting/Multicasting Protocol
Language:C++1
bryanzhang/c3p0
a mature, highly concurrent JDBC Connection pooling library, with support for caching and reuse of PreparedStatements.