Skip to content

Latest commit

 

History

History
 
 

papers

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
ISPD-2022
ISPD-2022
 
 
NLP
NLP
 
 
attribute-recognition
attribute-recognition
 
 
classification
classification
 
 
cluster_scheduler
cluster_scheduler
 
 
communication
communication
 
 
data
data
 
 
fasterMoe
fasterMoe
 
 
frameworks
frameworks
 
 
imgs
imgs
 
 
knownledge-distillation
knownledge-distillation
 
 
large-model-inference
large-model-inference
 
 
moe
moe
 
 
object-detection
object-detection
 
 
parallelism
parallelism
 
 
pruning
pruning
 
 
quantization
quantization
 
 
recommender
recommender
 
 
sparsity
sparsity
 
 
text-image-synthesis
text-image-synthesis
 
 
transformer
transformer
 
 
8-bit-floating-point-numbers.md
8-bit-floating-point-numbers.md
 
 
ActNN-Theorem-Prove-HaotianHe.pdf
ActNN-Theorem-Prove-HaotianHe.pdf
 
 
ActNN-source-code.md
ActNN-source-code.md
 
 
ActNN.md
ActNN.md
 
 
GSPMD.md
GSPMD.md
 
 
GShard：Scaling Giant Models with Conditional Computation and Automatic Sharding.md
GShard：Scaling Giant Models with Conditional Computation and Automatic Sharding.md
 
 
ImageNet.md
ImageNet.md
 
 
Inception-GoogLeNet.md
Inception-GoogLeNet.md
 
 
LeNet.md
LeNet.md
 
 
PyTorch Distributed-Data Parallel Training.md
PyTorch Distributed-Data Parallel Training.md
 
 
README.md
README.md
 
 
Rammer-Enabling-Holistic-DL-Optimizations-with-rTasks.md
Rammer-Enabling-Holistic-DL-Optimizations-with-rTasks.md
 
 
ZeRO-Offload.md
ZeRO-Offload.md
 
 
ZeRO.md
ZeRO.md
 
 
batch-normalization.md
batch-normalization.md
 
 
capuchin.md
capuchin.md
 
 
characterizing-deep-learning-training-workloads-on-alibaba-pai.md
characterizing-deep-learning-training-workloads-on-alibaba-pai.md
 
 
cnn-concepts.md
cnn-concepts.md
 
 
deformable-convolutional-networks.md
deformable-convolutional-networks.md
 
 
designing-a-profiling-and-visualization-tool-for-scalable-and-in-depth-analysis-of-high-performance-gpu-clusters.md
designing-a-profiling-and-visualization-tool-for-scalable-and-in-depth-analysis-of-high-performance-gpu-clusters.md
 
 
efficient-large-scale-language-model-training.md
efficient-large-scale-language-model-training.md
 
 
layer-normalization.md
layer-normalization.md
 
 
mixed-precision-training.md
mixed-precision-training.md
 
 
pathways.md
pathways.md
 
 
persisten-threads.md
persisten-threads.md
 
 
rammer-source-code.md
rammer-source-code.md
 
 
stegastamp.md
stegastamp.md
 
 

README.md

  1. 在解决的是什么问题？
  2. 为何成功，标志/准是什么？
  3. 在前人基础上的关键创新是什么？
  4. 关键结果有哪些？
  5. 有哪些局限性？如何优化？
  6. 这个工作可能有什么深远的影响？