Oneflow-Inc/OneFlow-Benchmark

Issues

• Tensor 与 tensor 可能误用的排查工作

  #231 opened 3 years ago by doombeaker
  0

• bert run_squad.py训练问题

  #217 opened 3 years ago by lpj0711
  1

• gpt 0.4.0 的完整支持和验证

  #210 opened 3 years ago by strint
  1

• A100集群测试2机16卡吞吐低于单机8卡吞吐

  #206 opened 3 years ago by iamweizhi
  1

• 使用oneflow.nn.conv2d_transpose需要group参数，目前group不可以由用户指定

  #184 opened 3 years ago by FinnShen
  0

• Would like to see more CTR benchmarks.

  #172 opened 4 years ago by EtoDemerzel0427
  0

• Bert Model Parallel

  #168 opened 4 years ago by sywang0111
  4

• relu6实现有误

  #166 opened 4 years ago by MARD1NO
  3

• 未来需要合并的算法

  #147 opened 4 years ago by JiaqingFu
  2

• AttributeError: module 'oneflow' has no attribute 'keras'

  #160 opened 4 years ago by gentelyang
  1

• run sh train.sh hangs in resnet50 benchmark with 4 and 8 gpus of single machine

  #152 opened 4 years ago by wuyujiji
  1

• Unable to complete model training

  #141 opened 4 years ago by fengyuchao97
  5

• CNN benchmark cannot run

  #130 opened 4 years ago by JF-D
  9

• Update the code to match the latest version API

  #92 opened 4 years ago by ZhichengHuang
  3

• PiecewiseConstantScheduler implementation in Benchmark.

  #129 opened 4 years ago by JiaojiaoYe1994
  0

• 性能测试结果

  #90 opened 4 years ago by nbcsm
  8

• cnns two nodes train came error: `Check failed: error == CUDNN_STATUS_SUCCESS (9 vs. 0) CUDNN_STATUS_NOT_SUPPORTED` but one nodes can work well

  #103 opened 4 years ago by qianzhang613
  5

• README中链接忘记更新了

  #89 opened 4 years ago by wyg1997
  1