IEIT-Yuan/Yuan-2.0

Issues

• 使用fastchat训练的标签问题

  #150 opened 3 months ago by tianbo-che
  1

• yuan官方魔改的vllm不支持加载lora

  #149 opened 3 months ago by air158
  0

• 关于pp/tp/dp通信时间的推导

  #80 opened a year ago by Rememberz
  4

• PP不均匀切分

  #148 opened 5 months ago by JinXiaozhao
  1

• 在云上部署Yuan2-2B-Mars-hf时，HuggingFace模型推理API部署失败

  #146 opened 5 months ago by zzzEureka
  0

• FastChat推理报错

  #140 opened 8 months ago by robin-programmer
  1

• [Feature Request] 安卓佈署方案 Android Deployment

  #139 opened 8 months ago by DakeQQ
  2

• 请问下预训练完成之后的模型，在使用LLaMA进行微调的时候是怎么调用预训练的模型的？

  #141 opened 7 months ago by lijiecheng
  1

• 联合向量数据库和 Llama-index 的报错

  #136 opened 8 months ago by iChuck-W
  0

• 请问一下，yuan2.0的51B模型在预训练的时候用了多少数据量，多少GPU，花费了多长时间呢？

  #138 opened 8 months ago by listwebit
  0

• 请问huggingface版本如何在intel CPU和intel GPU上跑呢

  #92 opened 8 months ago by pengyb2001
  2

• 请问源2.0-102B（Windows）是否支持英特尔arc独立显卡

  #133 opened 8 months ago by tonyxx1
  3

• 执行milvus.py脚本import llama_index有路径错误

  #134 opened 8 months ago by zhengbingyang
  5

• 基于TensorRT-LLM的Yuan 2.0推理服务2B-hf版编译试用问题

  #127 opened 9 months ago by 18842685792
  2

• “用 TensorRT-LLM & Triton Server 部署2B模型”请问这样部署是同步接口吗？貌似不能异步调用？

  #132 opened 8 months ago by dongteng
  1

• TypeError: LlamaRotaryEmbedding.forward() missing 1 required positional argument: 'position_ids'

  #131 opened 8 months ago by zhangxiying
  4

• 微调能否调整参数，使微调脚本可以在GPU T4或者CPU上进行微调

  #125 opened 9 months ago by dsscclg
  1

• 如何将2.0-51B-hf这个HuggingFace版的模型进行pp和tp转换

  #113 opened 10 months ago by 18842685792
  7

• FastChat 部署报错：TypeError: forward() missing 1 required positional argument: 'position_ids'

  #123 opened 9 months ago by maxoyed
  3

• 使用YUAN2.0-2B替换BLIP2中的T5-XL

  #120 opened 9 months ago by dszpr
  6

• LLaMA调参方式下，能否区分/指定训练集与测试集

  #118 opened 9 months ago by fredfany
  4

• Yuan2.0 推理 API 部署

  #122 opened 9 months ago by Emirateyang
  1

• LLama训练过程中,训练的语料对原有问题的回答有干扰时，这种情况该如何调参。

  #119 opened 9 months ago by lijiecheng
  1

• Mac 端可以部署吗？

  #117 opened 10 months ago by Napoleon-L
  0

• HuggingFace版本的模型可以在CPU和GPU上运行， 请问原始版本的模型可以在CPU上运行吗？

  #114 opened 10 months ago by 18842685792
  3

• 关于模型训练、调参相关的几个基础问题

  #100 opened 10 months ago by fredfany
  13

• Yuan-2.0 2.1B模型在进行模型微调的时候报显存容量不足的问题

  #87 opened 10 months ago by fredfany
  18

• 如何在容器方式下，替换YuanChat服务的logo及头像文件

  #106 opened 10 months ago by fredfany
  7

• 使用官方huggingface代码推理报错

  #108 opened 10 months ago by seanzhang-zhichen
  4

• 推理示例代码问题

  #105 opened 10 months ago by Iamlovingit
  3

• Generate Token in Yuan2.0

  #98 opened 10 months ago by hungptit123
  3

• 使用yuan-2.0容器方式启动推理服务失败，提示cuda获取不到GPU

  #99 opened 10 months ago by fredfany
  3

• 使用原生方式（megatron）进行参数微调，微调完成后新的数据集学习到，模型原由的泛化能力几乎完全丢失

  #103 opened 10 months ago by fredfany
  2

• 请问，yuan-2.0的51B模型可以通过pretrain_yuan.py加载你们训练的模型权重，继续持续预训练吗？

  #101 opened 10 months ago by listwebit
  4

• [BUG]Huggingface版本推理流式输出报错

  #94 opened 10 months ago by cauwulixuan
  3

• raise Exception(f"Unexpected tensor name: {name}")

  #97 opened 10 months ago by juexing1981
  2

• Yuan2-2B-Janus-hf 推理速度很慢

  #96 opened 10 months ago by yaleimeng
  2

• yuan-2.0的51B模型支持增量预训练(持续预训练)吗，请大佬回复？

  #88 opened 10 months ago by listwebit
  3

• yuanchat本地部署报错

  #93 opened 10 months ago by Legionee
  1

• 您们能尝试训练开源的芯片生成模型吗?

  #90 opened 10 months ago by win10ogod
  10

• Huggingface 版必须要安装flash attention？

  #70 opened a year ago by Armod-I
  8

• 51B模型ckpt转换的疑问，模型iter_0000000可以直接使用吗？

  #82 opened a year ago by yuki772049
  3

• Yuan-2.0-2.1B的多机分布式预训练问题：FileNotFoundError: [Errno 2] No such file or directory:'home/Yuan-2.0-main/megatron/fused_kernels/build/lock'

  #86 opened a year ago by BoyCO3
  5

• windows端部署

  #75 opened a year ago by Longleaves
  3

• 推理出现错误No CUDA runtime is found, using CUDA_HOME='/usr/local/cuda'

  #69 opened a year ago by chesp
  3

• 微调可以使用多个数据集吗？

  #74 opened a year ago by yuki772049
  2

• 请问支持多GPU部署吗?

  #76 opened a year ago by Ca11MeE
  2

• 请问支持VLLM或者TRT部署么？

  #85 opened a year ago by qibao77
  1

• ModuleNotFoundError: No module named 'transformers_modules.IEITYuan.Yuan2'

  #72 opened a year ago by hungptit123
  2

• 2.1B微调报错：torch.distributed.elastic.multiprocessing.api:failed (exitcode: 1)

  #67 opened a year ago by chesp
  9