BeyonderXX/InstructUIE

Issues

• NER task的具体形式

  #41 opened 2 months ago by pangpang-xuan
  0

• 请问可以提供'../output/flant5-11b-v8-zeroshot/predict_eval_predictions.jsonl'文件嘛？我想看看数据格式并跑通caculate_f1.py

  #43 opened 3 months ago by wyclike
  0

• BuildConfig does not work

  #40 opened 7 months ago by jdm4pku
  2

• Auxiliary Tasks 是怎么添加的？

  #42 opened 4 months ago by LIYHUI
  0

• 请问运行./scripts/eval_flan-t5.sh需要多大的显卡资源，resume_from_checkpoint是指你提供的chekpoint路径，那么model_name_or_path应该是原始模型flan-T5-11b的路径吗？

  #33 opened a year ago by jy-101361-1810897
  3

• 开源的checkpoint的评估结果与论文差异较大？

  #24 opened a year ago by xueyongfu11
  3

• 混合精度训练(fp16)

  #39 opened 9 months ago by charosen
  0

• 关于论文里的Event Trigger F1、Event Argument F1问题

  #38 opened 10 months ago by guihonghao
  0

• NER的ACE2004在论文中没有实验结果？

  #37 opened a year ago by jy-101361-1810897
  0

• eval fan-T5 11b 的参数该如何设置

  #36 opened a year ago by cosmicexotic
  0

• 您好，InstructUIE是11B全参数微调吗？有没有用微调工具，类似lora什么的？另外复现这个项目要的硬件环境是啥呢？我目前有4块3090，不知道能不能行

  #35 opened a year ago by orangezfj
  0

• 更小的模型效果如何

  #34 opened a year ago by powerful6
  2

• Performance of GENIA-Evt in paper missing

  #32 opened a year ago by raven44099
  1

• 如何在 demo 模式下运行 run_uie.py ?

  #31 opened a year ago by OStars
  0

• 请问8张A100 80GB怎么设置参数

  #30 opened a year ago by HypherX
  1

• 输入长度限制

  #28 opened a year ago by JunxiuChen
  1

• where can we find the InstructUIE model besides huggingface？

  #29 opened a year ago by HelloBroTan
  0

• zero-shot for EE datasets？

  #26 opened a year ago by JunxiuChen
  1

• 训练时数据集下载出现问题，data_dir是不是下载的ie_instruction的路径呢？

  #27 opened a year ago by vv521
  2

• How much GPU memory do I need for fine-tuning and inference?

  #5 opened 2 years ago by DYF-AI
  3

• DatasetGenerationError: An error occurred while generating the dataset

  #16 opened a year ago by xiaoxiaobai321
  0

• chinese field

  #25 opened a year ago by githubgtl
  1

• 如何改成支持中文的

  #15 opened a year ago by xiaoxiaobai321
  5

• 论文的NER等任务的实验结果都是用flan-t5-11b进行全量参数微调的结果吗？

  #23 opened a year ago by xueyongfu11
  0

• Bugs in Data Preprocess and may cause the wrong evaluation results: when there are multiple events in a sentence, the output data only contains one event and misses other events.

  #22 opened a year ago by Lee-zix
  0

• Question that whether this method support Chinese Task?

  #21 opened a year ago by 976230600
  7

• Instructions in the released dataset

  #20 opened a year ago by Sreyan88
  1

• 这个程序应该在什么目录下执行呢

  #14 opened a year ago by xiaoxiaobai321
  1

• deepspeed模式下如何debug

  #17 opened a year ago by xiaoxiaobai321
  1

• run eval_flan-t5.sh with do_demo

  #18 opened a year ago by dignjun
  2

• Want to get clarity on how to give the input for zero-shot NER.

  #12 opened a year ago by bhavuk-singhal
  1

• Zero-shot NER experiment settings

  #13 opened a year ago by Spico197
  2

• When will the dataset and the checkpoint files been released？

  #3 opened a year ago by Hjunchi
  3

• When can we expect the releasing of the checkpoint?

  #8 opened a year ago by bhavuk-singhal
  2

• 4块titan微调模型时显存不够了，用glm-6b换了flan-t5的模型（fp16），改小了batch参数，还是不行，请问有哪些参数还可以调一下，并且不怎么影响效果的

  #9 opened a year ago by Tom2576
  3

• question about evaluator.py

  #10 opened 2 years ago by Wangxh-07
  0

• what is the means for samlpe_15000 in dataset?

  #6 opened 2 years ago by zplovekq
  1

• The mismatch between InstructUIE datasets and UIE datasets

  #7 opened 2 years ago by gao-xiao-bai
  3

• the format of the dataset

  #4 opened 2 years ago by awhathell
  1

• Dataset

  #2 opened 2 years ago by gzw209
  2