OS-Copilot/OS-Atlas

Issues

• OS-Atlas/eval/eval_scripts/omniact /evaluation_unify_multi.py测评问题

  #52 opened 18 days ago by LianShuQuan
  2

• Different number of test samples of AndroidControl dataset

  #40 opened 3 months ago by NingLi670
  3

• ac数据测试相关问题，求解答

  #50 opened a month ago by lijianhong-code
  1

• Where can I find the data synthesis toolkit?

  #24 opened 4 months ago by QunixZ
  3

• 请教Offline评测数据集上的训练设置

  #45 opened a month ago by XuRui314
  6

• MLX Support

  #51 opened 22 days ago by prncvrm
  1

• Inference script for OS-Atlas-Base-4B on screenspot

  #46 opened 2 months ago by ZZZhr-1
  0

• android_control的评测相关问题

  #47 opened 2 months ago by lijianhong-code
  1

• question about IG_data in huggingface

  #37 opened 4 months ago by Kelatte
  2

• OSworld test code

  #18 opened 4 months ago by CarlHuangNuc
  3

• OS-ATLAS-7B-Pro sometimes fails to output low-level operation

  #44 opened 2 months ago by ZJULiHongxin
  0

• 为什么我们复现不了你们的训练过程？

  #35 opened 3 months ago by CarlHuangNuc
  3

• Inference Results on ScreenSpotv2

  #39 opened 3 months ago by xuzq23
  3

• Training Resources

  #26 opened 3 months ago by CarlHuangNuc
  10

• 请问模型输出的bbox坐标如何还原到原图上？

  #41 opened 3 months ago by cGy147
  2

• 关于训练的一些疑问?

  #25 opened 3 months ago by pange1802703882
  29

• Problem on cal Avg Acc in ScreenSpot

  #38 opened 3 months ago by shaotao-art
  2

• training data detail format

  #29 opened 3 months ago by CarlHuangNuc
  0

• 7B模型训练，数据准备中试用到point /OCR吗？

  #33 opened 3 months ago by CarlHuangNuc
  1

• 环境依赖是多少？

  #32 opened 3 months ago by MachineDora
  2

• Qwen2-VL-7B Finetuing 需要打开VIT 参数训练吗？

  #34 opened 3 months ago by CarlHuangNuc
  1

• HuggingFace上的4B模型似乎错误调用了Phi3ForCausalLM的generate方法

  #19 opened 5 months ago by Lucas-lyh
  2

• Catastrophic forgetting when trying to fine-tune OS-ATLAS-Base-7B on grounding

  #36 opened 4 months ago by PranitChawla
  0

• 训练集：point_grounding和ocr组织格式

  #30 opened 4 months ago by zhoumenghan
  3

• fineweb_3m dataset

  #31 opened 4 months ago by mrswang1
  0

• Whether IG data will public in future ?

  #20 opened 4 months ago by CarlHuangNuc
  3

• Plans to release training code?

  #12 opened 4 months ago by jasonlee-sf
  3

• 关于prompt.json

  #28 opened 4 months ago by zhoumenghan
  4

• 训练数据集格式

  #27 opened 4 months ago by zhoumenghan
  2

• grounding test code

  #23 opened 4 months ago by CarlHuangNuc
  0

• Training server

  #22 opened 4 months ago by CarlHuangNuc
  1

• grounding benchmark test code

  #21 opened 4 months ago by CarlHuangNuc
  0

• Plans to release training data except ScreenSpot-V2?

  #13 opened 4 months ago by jasonlee-sf
  0

• What is the difference between the two model ?

  #10 opened 4 months ago by CarlHuangNuc
  1

• action fine-tuning model

  #7 opened 5 months ago by aptsunny
  2

• License

  #4 opened 5 months ago by fakerybakery
  1

• 期待早日开源

  #1 opened 5 months ago by shuqingjinse
  2

• 请问团队预期什么时间放出数据？

  #3 opened 5 months ago by XinrunXu
  2

• 数据什么时候开源

  #5 opened 5 months ago by wuzhizhige
  4

• 期待着早日开源，造福科研界

  #11 opened 5 months ago by CarlHuangNuc
  1

• 期待早日开源

  #16 opened 5 months ago by lr389032
  1

• Conversational Model

  #6 opened 5 months ago by aswad546
  5

• Batched inference for `OS-Atlas-Base-7B` is broken with `attn_implementation="flash_attention_2"`

  #17 opened 5 months ago by jasonlee-sf
  0

• How to Generate Plan for the given User instruction Task?

  #15 opened 5 months ago by anonymousRj7
  0

• Onnx inference

  #14 opened 5 months ago by thewh1teagle
  0

• the 7b demo gives the wrong cooridinates

  #9 opened 5 months ago by zxxmxd
  1

• Average Score Clarification

  #8 opened 5 months ago by ant-8
  1

• 请问你们的模型有在中文场景上做优化吗？

  #2 opened 5 months ago by xyfZzz
  1