Zheng Xin comments

Results 69 comments of


                                            Zheng Xin

How can I download the dataset

i meet the same question, have you got the dataset? can you share it? thanks.

Request for preprocessed dataset

have you got the preprocessed dataset?

请问一下，你是怎么转onnx的呢？可以共享一下脚本吗？

> ``` > def forward(self, x): > # x = x.copy() # for profiling > z = [] # inference output > # self.training |= self.export > if self.export: >...

thanks for your repository. how can i run the code and prepare the dataset? thanks for your help!

should i prepare my own dataset as MOT16 dataset format? or any other dataset? if i should prepare my own dataset as the MOT16 format, how can i create det.txt,should...

how to visiual the pred result , need help

> 你好！你发的文件我已经收到了。我会尽快回复你的~ hi, can you help me? thanks.

[BUG] ES connection Error with unexpected keyword argument 'host'

> **问题描述 / Problem Description** > > 2024-10-18 11:27:27.077 | WARNING | chatchat.server.knowledge_base.kb_service.es_kb_service:do_init:45 - ES未配置用户名和密码 2024-10-18 11:27:27.077 | ERROR | chatchat.server.knowledge_base.kb_service.es_kb_service:do_init:59 - Error 发生 : Elasticsearch.**init**() got an unexpected keyword...

[BUG] ES connection Error with unexpected keyword argument 'host'

> > 你好，我也遇到了这个问题，请问后来解决了吗 > > I suppose change 'host' key to 'hosts' will fix the issue. connection_info = dict(host=f"{self.scheme}://{self.IP}:{self.PORT}") > > Not sure why it happens. thanks,I will try.

Qwen3 VL 模型 "data:video/jpeg;base64," 方式传多个视频帧 vllm serve报错 Error code: 500 - {'error': {'message': '', 'type': 'Internal Server Error', 'param': None, 'code': 500}}

> > > [@wulipc](https://github.com/wulipc) 代码参考你在另外一个issue post的方法 [#806 (comment)](https://github.com/QwenLM/Qwen3-VL/issues/806#issuecomment-2667817885) > > > > > > [@chuengleo](https://github.com/chuengleo) hi，我昨天看了下 vLLM 预处理部分的代码，根据传入的 base64 有 2 中情况： > > > > * data:video/jpeg;base64：如果使用的是这个，一定会报错，因为返回的video metadata (参考[这里](https://github.com/vllm-project/vllm/blob/5f6cbf60d667408ef1bf3871c0b7563effa6094a/vllm/multimodal/video.py#L248C9-L248C17))是空字典，后续初始化会报错，代码详见[这里](https://github.com/vllm-project/vllm/blob/5f6cbf60d667408ef1bf3871c0b7563effa6094a/vllm/multimodal/video.py#L296)....

使用VLLM的openai接口进行视频推理时出现异常

vllm启动代码：CUDA_VISIBLE_DEVICES=0,1,2,3 vllm serve /mnt/disk1/DATA/llm/Qwen3-VL-32B-Instr Qwen3-VL-32B-Instruct --dtype float16 --tensor-parallel-size 4 --gpu-memory-utilization 0.8 --port 9996 --max-model-ling --media-io-kwargs '{"video": {"num_frames": -1}}' 调用代码：import base64 import numpy as np from PIL import Image from io...