Tongjilibo/bert4torch

Issues

• could not broadcast input array from shape (2,13) into shape (2,)

  #173 opened 3 months ago by dengqiqi123
  3

• 大佬请教个问题

  #172 opened 3 months ago by wuxiaohui0
  1

• gridlabels的构建

  #170 opened 7 months ago by wuxiaohui0
  1

• NER嵌套实体识别bug

  #169 opened 7 months ago by tianchiguaixia
  4

• Text2Vec加载BAAI_bge-large-zh-v1.5进行相似度计算得到的结果有误

  #168 opened 9 months ago by sslalxh
  2

• 在CMeIE1.0数据集上使用TPLinker实现信息抽取任务，precision很高，recall极低

  #142 opened 10 months ago by JasonHI7
  4

• 请求大佬写个NER指针网络的预测代码

  #167 opened 9 months ago by Echo-she
  4

• self.random_sample和 self.beam_search返回值tensor未统一转化numpy，导致两函数同时调用报错

  #165 opened 10 months ago by ocyisheng
  2

• 加载Yi-6B进行lora微调报错

  #166 opened 9 months ago by zhadandeiren
  1

• bert4torch/examples/llm /task_chatglm2_lora.py中的config_path = dir_path + 'bert4torch_config.json' 在哪里？

  #163 opened 10 months ago by zysNLP
  3

• callback变红

  #164 opened a year ago by lili-li-cpu
  3

• bert4torch中，使用单卡多GPU训练，使用accelerate后，不能使用梯度剪裁。

  #162 opened a year ago by wjj97172
  3

• Trainer.compile与torch2.0新加的model.compile重名

  #159 opened a year ago by FairyOwO
  3

• 分类算法sentence句子编码的时候，没理解到mask处理逻辑

  #158 opened a year ago by cwqJim2023
  1

• bert4torch版本0.2.8升级到0.3.4问题

  #157 opened a year ago by kakaxzc
  2

• LLM：chatGLM2推理加速

  #156 opened a year ago by Lxhnnn
  1

• simbert微调问题

  #155 opened a year ago by Lxhnnn
  2

• XLnet分类报错

  #154 opened a year ago by Lxhnnn
  4

• chinese-xlnet-mid载入

  #153 opened a year ago by chosenone75
  1

• 请问怎么调整最后的输出结果的小数点位数，目前最后的测试结果是两位，在哪里可以调整一下啊？

  #150 opened a year ago by xinjiesun-ustc
  7

• 显示每个epoch的已用时间和预估剩余时间

  #152 opened a year ago by zhouyiyuan-mt
  3

• 请问在用bart bert等模型生成标题时，如果修改标题的最大长度？

  #151 opened a year ago by kakaxzc
  3

• DataLoader的num_workers设置为大于0时出错

  #149 opened a year ago by zhouyiyuan-mt
  3

• 预训练模型加载

  #148 opened a year ago by Smile-L-up
  5

• 询问一个bert4torch和bert4keras的差异

  #145 opened a year ago by pql123
  1

• 单机多卡BaseModelDDP初始化错误，AttributeError: can't set attribute

  #146 opened a year ago by jjRen-xd
  12

• 训练进度条中，loss一直显示Nan的问题

  #144 opened a year ago by Yikai-Liao
  5

• tokenizer_encode_config和tokenizer_decode_config问题

  #147 opened a year ago by YiTouch
  3

• convert sequence labeling model to onnx

  #143 opened a year ago by feng-1985
  1

• TypeError: forward() takes 2 positional arguments but 3 were given

  #141 opened a year ago by zhudongwork
  1

• 使用AutoModel替代build_transformer_model，发现其参数在训练过程中不会被更新

  #140 opened a year ago by ZayIsAllYouNeed
  14

• deberta_v2保存的模型非常大

  #118 opened 2 years ago by leizhu1989
  2

• 模型推理标签与字符错位，不确定是哪里出了问题

  #131 opened 2 years ago by ykallan
  7

• 有关提取token向量的探讨

  #116 opened a year ago by leizhu1989
  6

• Sohu_2022_ABSA sequence_padding的疑惑

  #130 opened a year ago by huaweiboy
  1

• RDropLoss函数实现 bug

  #138 opened a year ago by ZayIsAllYouNeed
  1

• print_trainable_parameters 函数缺失

  #139 opened 2 years ago by ZayIsAllYouNeed
  5

• 其他example的单机多卡适配问题

  #134 opened 2 years ago by yFhope
  4

• 安装时报错：Could not find a version that satisfies the requirement

  #133 opened 2 years ago by WenjingBao
  10

• GPLinker的中的MyLoss类下出现的问题

  #132 opened 2 years ago by liyao345496280
  8

• W2NER模型加载不连续实体的data

  #122 opened 2 years ago by Bureaux-Tao
  3

• 量化后的模型可以保存并 重新加载使用吗？

  #127 opened 2 years ago by Minami-su
  8

• 有关模型微调

  #108 opened 2 years ago by leizhu1989
  7

• DP模式gradient_penalty下Loss起步无限大

  #120 opened 2 years ago by skykiseki
  4

• 训练模型的时候在eval阶段就会oom

  #121 opened 2 years ago by wjddd
  2

• 基于mt5+amp的模型微调loss直接走向nan

  #107 opened 2 years ago by skykiseki
  1

• 权重转换问题

  #112 opened 2 years ago by Clearloveplus7
  1

• Global Pointer 的bug?

  #102 opened 2 years ago by CheungZeeCn
  2

• accelerate 不支持对抗训练

  #113 opened 2 years ago by skykiseki
  1

• example 过期失效链接

  #103 opened 2 years ago by skykiseki
  1