bytedance/ByteMLPerf

Issues

• Debian 12 - LLM inference perf: [Error 24] Too many open files

  #121 opened a month ago by tdbyte
  5

• No such file or directory: 'general_perf/datasets/open_squad/dev-v1.1.json'

  #123 opened a month ago by yaoyixuejava
  1

• AMD GPU

  #114 opened a month ago by trinashih
  1

• llm_perf 厂商接入时torch的版本问题

  #116 opened 2 months ago by yizhou-xu
  5

• general_perf厂商接入的环境问题

  #118 opened 2 months ago by Yi-sir
  2

• 请教下目前 model_zoo 中提供的 fp32 的模型，有提供工具或者方法快速转换为 fp

  #38 opened 2 months ago by eric-yq
  2

• [CPU]activate_this.py missing

  #115 opened 2 months ago by zhengy001
  0

• 【llm_perf】 GPU backend chatglm2-6b，hardcode max_seq_len=4096 when init_kvcache

  #113 opened 2 months ago by wufann
  0

• 【llm_perf issue】using byte_infer_perf/llm_perf/launch.py to test chatglm, but meet multi-process competing

  #112 opened 3 months ago by danielhua23
  3

• llm perf with vllm server

  #104 opened 4 months ago by At1a8
  0

• 【Issue Help】 chatglm2-6b has some cases dismatch with golden

  #77 opened 7 months ago by DeepTecher
  2

• 【Issue Help】 run bs=8, input_len=2048 perf error

  #78 opened 6 months ago by DeepTecher
  2

• 【Bug】Bug Need fix -- dismatch output dimension with gloden

  #76 opened 7 months ago by DeepTecher
  3

• 有没有可能支持llama3推理测试

  #71 opened 7 months ago by incomingflyingbrick
  1

• 显存溢出

  #73 opened 7 months ago by incomingflyingbrick
  3

• python 3.10.12无法启动

  #72 opened 7 months ago by incomingflyingbrick
  2

• llm_perf not running

  #67 opened 8 months ago by ronysun
  1

• 运行gpt2-torch-fp32 task 遇到的2个问题

  #49 opened 10 months ago by longerzone
  1

• SD-Vae-decoder 输入是否有误？

  #44 opened 10 months ago by Ivorfeng
  1

• Deberta model json config file does not match the input of the model.

  #39 opened a year ago by abelleeye
  1

• can not run gp2 on habana gaudi2

  #34 opened a year ago by ZhaiFeiyue
  2

• are the bert tensorflow and bert pytorch the same model?

  #31 opened a year ago by Fizzbb
  2

• Re-batch not called if only the numeric checker set true

  #4 opened 2 years ago by jackzipu
  1

• 'PerfEngine' object has no attribute 'pre_compile_config' when the get_best_batch_size() implemented in the backend

  #3 opened 2 years ago by jackzipu
  1

• The contents missed in the byte_mlperf/datasets folder

  #1 opened 2 years ago by jackzipu
  2