QwenLM/Qwen2-VL

Issues

• Some weights were not used after GPTQ quantization

  #416 opened 23 days ago by chongkuiqi
  0

• code of visual embedding adapter?

  #336 opened a month ago by MartinYYYYan
  1

• vllm和hf推理的结果差异较大

  #481 opened 18 days ago by yangte3518880
  3

• 量化模型GPTQ推理速度慢

  #483 opened 17 days ago by thuBingo
  1

• Bug in beam_search: memory allocation

  #334 opened a month ago by greeksharifa
  2

• missing dependency

  #480 opened 18 days ago by jpcam
  2

• 7b-awq模型在A10上使用VLLM镜像部署，显存总是oom。

  #482 opened 18 days ago by ZhanNicky
  1

• High training loss when using M-ROPE

  #406 opened 18 days ago by Wiselnn570
  0

• [Qwen2-VL SFT training implemention ! ] Qwen2-VL has been implemented in PaddleMIX, welcome to use it !

  #479 opened 19 days ago by nemonameless
  0

• 请问Qwen2-VL支持多轮对话微调吗？

  #417 opened 23 days ago by lcs0215
  2

• AWQ output result is incorrect, return [the .]

  #478 opened 19 days ago by lxiaoxiaoxing
  0

• Why Qwen2-VL do not release Qwen2-VL (pretrain model), but Qwen-VL release that one(Qwen-VL + Qwen-VL-Char)?

  #477 opened 19 days ago by CarlHuangNuc
  0

• Birthday wishes in marathi

  #426 opened 20 days ago by brahmamudi-serial-today-episode
  0

• 使用AutoGPTQ量化lora微调后的72b模型

  #347 opened a month ago by ChinChyi
  2

• 按照要求安装transformers和vllm后使用vllm部署报错

  #348 opened 22 days ago by xuyifan-0731
  6

• qwen2-Vl如何封装API，官网有提供吗 还是需要自己写fastAPI呢

  #414 opened 22 days ago by ops-xin
  0

• A unkown Bug about `asyncio.exceptions.CancelledError`

  #418 opened 22 days ago by Jalen-Zhong
  0

• How did the length extrapolation perform in Figure 5 ?

  #415 opened 23 days ago by SCZwangxiao
  0

• multi turn issue

  #413 opened 23 days ago by codeMonkey-shin
  0

• VQA Score support

  #408 opened 24 days ago by Dratlan
  1

• 基于box提问效果较差，是prompt或者调用上有问题吗？请帮确认下

  #411 opened 23 days ago by qinzhenyi1314
  0

• 多轮对话模板

  #404 opened 25 days ago by THU-Kingmin
  1

• The vocab_size=152064, why len(tokenizer) = 151657 ?

  #350 opened a month ago by zhang9302002
  4

• Where can I find the implementation of M-RoPE?

  #409 opened 24 days ago by Backdrop9019
  0

• I use the key "fps" instead of "video_fps." ,but it still doesn't work !!!

  #351 opened a month ago by jackhexiangyu12
  1

• How many samples in calibration dataset should i use when AWQ quantization ?

  #407 opened 24 days ago by chongkuiqi
  0

• moov atom not found

  #405 opened 25 days ago by cuizhiling
  0

• 是否各个阶段的模型release (stage1,stage2)

  #403 opened 25 days ago by CarlHuangNuc
  0

• 多模态大模型grounding/Detection的时候如何加入空的bbox

  #394 opened 25 days ago by xingmimfl
  0

• how about data release plan on stage2,stage3?

  #353 opened 25 days ago by CarlHuangNuc
  0

• refcoco系列指标无法对齐

  #352 opened a month ago by 2018211801
  0

• Cannot call any model using api key

  #349 opened a month ago by HumayounMustafa
  0

• How to train on both image data and text-only data together?

  #346 opened a month ago by dooinee-bom
  1

• Qwen2VLForConditionalGeneration.from_pretrained 出现问题

  #344 opened a month ago by ChinChyi
  0

• vllm error

  #345 opened a month ago by liulixin29
  0

• 72b 模型对显卡的硬件要求说明

  #340 opened a month ago by V-yw
  2

• @Trapper4888 you need to specify `--video_fps 1.0`

  #343 opened a month ago by pingbowen23
  1

• Can anyone confirm if this is how qwen2-vl prompting for grounding task with multiple bounding boxes?

  #342 opened a month ago by 5thGenDev
  0

• Does Qwen2-VL-72B-Instruct have text generation capabilities equivalent to Qwen2-72B-Instruct ？

  #341 opened a month ago by Valdanitooooo
  1

• 请问当前支持的图像格式有哪些，svg格式是不是不支持？

  #339 opened a month ago by yangxin60-tal
  0

• vllm部署，在请求时报这个错误“TypeError: Unable to evaluate type annotation 'Required[Union[str, Iterable[ChatCompletionContentPartTextParam]]]'.”

  #338 opened a month ago by liukun2007
  5

• vllm离线推理正常，在线推理利用官方格式报错400

  #337 opened a month ago by BIAO-ac
  0

• how to run 7B BF16 in A10(24G)?

  #335 opened a month ago by MultiChen
  0

• eval scripts

  #309 opened a month ago by mfarre
  1

• cv2.imshow not working after install qwen_vl_utils

  #333 opened a month ago by SEUNGHYUN-PARK
  0

• 请问下是否支持 qwen2-vl-72b 纯本文和图片指令数据的混合微调？

  #332 opened a month ago by Luccadoremi
  0

• [Typo]For quantized models, we offer two types of quantization: AWQ and GPQ(🤗🤖).

  #331 opened a month ago by PaulWongDlut
  0

• When I upload more than one image at the same time, the VLLM server reports the following error.

  #308 opened a month ago by Sushaokun
  0

• OpenAI API server for macOS

  #307 opened a month ago by ChristianWeyer
  0

• 官方 qwen-xl-max 出現 InvalidParameter 但 qwen-xl-plus 可以處理

  #306 opened a month ago by coding101-hk
  0