PersiaML/PERSIA

Issues

• The throughput is extremely low

  #233 opened 2 years ago by zxgx
  4

• tracking issue: initial version

  #7 opened 2 years ago by NOBLES5E
  1

• more examples

  #214 opened 2 years ago by NOBLES5E
  1

• Improve compile time of Rust Extension

  #206 opened 2 years ago by williamstar
  1

• feat: embedding model definition

  #201 opened 2 years ago by williamstar
  1

• feat: suppport fp16 embedding

  #90 opened 2 years ago by williamstar
  1

• PS features require access rules

  #73 opened 2 years ago by nealgavin
  1

• PS needs to be evicted by business policy

  #72 opened 2 years ago by nealgavin
  1

• terminating training by nats server signal

  #33 opened 2 years ago by williamstar
  1

• 用honcho启动卡在`SingleMachine training context init done`这里

  #232 opened 2 years ago by zxgx
  2

• npy is used for small dataset in the example, what data format should be used for large-size data?

  #216 opened 2 years ago by xpai
  1

• Where is EmbeddingWorkerNatsServicePublisher?

  #218 opened 2 years ago by zxgx
  2

• 请问这套架构为啥用rust写呢？

  #217 opened 2 years ago by NHZlX
  2

• Does it support Tensorflow?

  #215 opened 2 years ago by rabintang
  1

• Refactor persia.data module to provide more functionality

  #185 opened 2 years ago by williamstar
  0

• API documentation

  #23 opened 2 years ago by NOBLES5E
  2

• organize tests with pytest and pass CI

  #15 opened 3 years ago by NOBLES5E
  0

• docs: add cite persia section in readme and tutorial

  #4 opened 3 years ago by NOBLES5E
  0

• Generate e2e test from examples

  #70 opened 3 years ago by williamstar
  0

• More readable PersiaBatch construction

  #183 opened 3 years ago by snowpeakz
  0

• Generate criteo syn dataset

  #196 opened 3 years ago by BLue1881euLB
  3

• CI: multi-machine multi-gpu system test

  #187 opened 3 years ago by snowpeakz
  0

• Missing .env file?

  #188 opened 3 years ago by Tokkiu
  2

• Load exists checkpoint to continue training

  #182 opened 3 years ago by williamstar
  0

• expose configuration for max message size in NATS

  #22 opened 3 years ago by NOBLES5E
  0

• docs: Tutorial cloudflare pages with an unaccessible footer

  #153 opened 3 years ago by yzs981130
  1

• support pure CPU mode

  #21 opened 3 years ago by NOBLES5E
  0

• add timeline tools to analyze bottleneck

  #92 opened 3 years ago by snowpeakz
  1

• feat: accelerate server model dump and load

  #36 opened 3 years ago by NOBLES5E
  0

• Create appropriate CI image (persiaml/persia-ci) and use it for CI

  #37 opened 3 years ago by NOBLES5E
  4

• remove intent in config file

  #61 opened 3 years ago by snowpeakz
  0

• configurable start_deadlock_detection_thread

  #118 opened 3 years ago by snowpeakz
  0

• support low memory adagrad sparse optimizer

  #91 opened 3 years ago by NOBLES5E
  0

• torch ddp launch with tcp instead of env_file

  #20 opened 3 years ago by williamstar
  1

• feat: dump/load dense model use hdfs

  #59 opened 3 years ago by snowpeakz
  1

• provide shutdown interface to ensure rust thread handler exit normally

  #32 opened 3 years ago by williamstar
  1

• ci: add python test and benchmark

  #3 opened 3 years ago by NOBLES5E
  1

• ci: add pytype check

  #12 opened 3 years ago by NOBLES5E
  1

• ci: add semver auto release and pypi upload

  #1 opened 3 years ago by NOBLES5E
  0

• docs: create API documentation website

  #2 opened 3 years ago by NOBLES5E
  0