ymcui/Chinese-LLaMA-Alpaca-2

Issues

• 单机多卡训练，加载数据集时卡住，大概是卡在training_args.main_process_first(desc="dataset map tokenization and grouping")，请问如何解决，谢谢

  #570 opened 4 months ago
  0

• binascii.Error: Incorrect padding：How to solve it?

  #569 opened 4 months ago
  2

• 什么导致chinese-alpaca-2-7b推理存在大量重复生成情况 呢

  #568 opened 6 months ago
  6

• 请问reward模型怎么部署推理？

  #567 opened 5 months ago
  3

• 训练数据和测试数据开源了么?

  #566 opened 5 months ago
  6

• 模型预训练时的labels问题

  #565 opened 5 months ago
  2

• 模型微调

  #564 opened 5 months ago
  2

• HELP!!!!!!!!!!!!!!!!!!!!!!!

  #562 opened 6 months ago
  1

• 使用transformer命令行进行交互时推理报错

  #561 opened 6 months ago
  2

• 模型，做了屏蔽词管理么？

  #560 opened 6 months ago
  1

• 预训练数据以及微调数据会开源吗？

  #559 opened 6 months ago
  2

• 微调后的lora模块

  #558 opened 6 months ago
  9

• 权重合并后重新加载训练时出现错误

  #556 opened 6 months ago
  30

• 训练垂直领域大模型应该基于哪个版本？

  #555 opened 6 months ago
  3

• 通过openai_server_demo/openai_api_server_vllm.py 运行，输出出现自问自答

  #554 opened 6 months ago
  2

• ImportError: /usr/local/lib/python3.10/dist-packages/transformer_engine_extensions.cpython-310-x86_64-linux-gnu.so: undefined symbol:

  #553 opened 6 months ago
  2

• 多卡训练卡在加载模型

  #552 opened 6 months ago
  7

• 无法从checkpoint恢复训练

  #551 opened 6 months ago
  3

• 指令精调

  #550 opened 6 months ago
  4

• 指令精调

  #549 opened 6 months ago
  2

• 预训练完成后模型的使用

  #548 opened 6 months ago
  4

• 6卡指令精调，报错oom

  #547 opened 7 months ago
  4

• finetune之后的模型使用

  #546 opened 7 months ago
  3

• 'padding_value' (position 3) must be float, not NoneType

  #545 opened 7 months ago
  3

• 在精调的时候，如何让模型在指定的GPU上运行，而不是只在cuda:0上

  #544 opened 7 months ago
  4

• 词汇表扩充并且增量训练的具体流程和修改哪些部分？

  #543 opened 7 months ago
  7

• 词汇表扩充后出现错误？

  #542 opened 8 months ago
  1

• How can I output generation scores(logits)?

  #541 opened 7 months ago
  2

• The model's performance is poor when using the merged tokenizer.

  #540 opened 6 months ago
  5

• 扩充词表后对新添加token初始化的方式

  #538 opened 8 months ago
  2

• 卡在加载数据集这一步

  #537 opened 7 months ago
  5

• 运行模型时output norm.weight' notfound如何解决

  #534 opened 8 months ago
  2

• ceval的zero-shot测评，原生的llama-2-7b比本仓库的中文llama-2-7b效果要好

  #533 opened 8 months ago
  1

• 访问次数多了以后显存不释放

  #532 opened 7 months ago
  6

• 请教一个问题。如何才能喂饱多个GPU

  #531 opened 7 months ago
  3

• 如何调整 Batch Size

  #530 opened 7 months ago
  3

• 1.3B模型是如何训练的？

  #529 opened 7 months ago
  6

• Knowledge updation

  #527 opened 8 months ago
  1

• 运行时显存占用过大和没有获取json返回体

  #525 opened 7 months ago
  17

• 请问本仓库能否基于YaRN进行sft？

  #524 opened 7 months ago
  5

• “基座模型”和“指令模型”该怎么使用？

  #522 opened 9 months ago
  1

• model will broken when i start pretraining

  #521 opened 8 months ago
  3

• llama.cpp部署出现格式错误

  #519 opened 9 months ago
  3

• load_in_8bit 推理耗时比fp16长

  #516 opened 8 months ago
  3

• 请教如何更换Tokenizer进行训练，Tokenizer大小不匹配问题

  #514 opened 9 months ago
  6

• 报错import name 'insecure_hashlib' from 'huggingface_hub.utils'

  #508 opened 9 months ago
  0

• 关于强化学习

  #506 opened 9 months ago
  3

• 指令精调模型后生成的文件中只有adapter_model.safetensors，没有adapter_model.bin

  #504 opened 9 months ago
  4

• 奖励模型如何进行推理

  #503 opened 9 months ago
  3

• 使用flash attention会报错

  #502 opened 9 months ago
  2