OpenPPL/ppl.pmx

Issues

• 为什么使用llama脚本转换qwen1.5权重文件，转出来的文件是model.pth和opmx_params.json两个文件

  #90 opened 6 months ago by Flynn-Zh
  6

• Good first issue

  #80 opened 7 months ago by syheliel
  3

• 文档内显示支持qwen1.5，但实际转换报错。

  #77 opened 7 months ago by poemon1
  1

• [LLaMA2-7b] Warning: The shape inference of opmx::ParallelEmbedding type is missing, so it may result in wrong shape inference for the exported graph. Please consider adding it in symbolic function. (function UpdateReliable)

  #82 opened 7 months ago by syheliel
  1

• baichuan13B static/dynamic batching 生成一定长度的句子之后结果出现分歧

  #28 opened 9 months ago by Alcanderian
  1

• 补充RoPE2D dynamic batching算子文档

  #30 opened a year ago by Alcanderian
  0

• 转换llama2 70B模型时遇到的问题

  #25 opened a year ago by zbtrs
  4

• 怎么转换其它模型

  #15 opened a year ago by maiquanshen
  3

• TYPO in “ppl.pmx/docs/operators/dynamic_batching/MultiHeadAttention.md”

  #10 opened a year ago by wplf
  2

• Use A10 to export llama 13B model OOM.

  #3 opened a year ago by shanshanpt
  1