tlkh/tf-metal-experiments

Issues

• TFLOPS is off by a factor of 2 in conv_benchmark.py?

  #20 opened a year ago by geohot
  0

• Graph execution error

  #19 opened a year ago by stevebanik
  0

• Which CPU and RAM in the Reference Benchmarks 3090 ?

  #5 opened 2 years ago by Chizkiyahu
  1

• Adding pytorch benchmark?

  #18 opened 2 years ago by thipokKub
  0

• Disable Eager Execution

  #17 opened 2 years ago by my-other-github-account
  0

• Issues from "Big Sur" to "Monterey"

  #16 opened 2 years ago by leoeduardo69
  0

• does M1 max can use all of his memory in training???

  #10 opened 3 years ago by wildbrother
  2

• Did anyone try M1 Pro

  #13 opened 2 years ago by alin256
  2

• 2x Faster Inferencing

  #14 opened 2 years ago by TortoiseHam
  0

• smaller batch sizes

  #11 opened 2 years ago by invoxiaehu
  3

• M1 Max Thermal Throttling

  #4 opened 3 years ago by yadamonk
  9

• adding A100, other configs comparison for fun

  #6 opened 3 years ago by ngam
  30

• typo and probably do all stuff in conda?

  #7 opened 3 years ago by ngam
  3

• Low results for 3090?

  #8 opened 3 years ago by Randl
  5

• High Power Mode

  #3 opened 3 years ago by TortoiseHam
  5