OFA-Sys Chinese-CLIP issues

OFA-Sys / Chinese-CLIP

Chinese version of CLIP which achieves Chinese cross-modal retrieval and representation generation.

MIT License

4k stars 418 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

windows单机单卡训练报错RuntimeError: Distributed package doesn't have NCCL built in

#279 bowenzc opened 3 months ago
2
关于执行sh脚本文件的报错 RuntimeError: NCCL error in: ../torch/lib/c10d/ProcessGroupNCCL.cpp:911, unhandled system error, NCCL version 2.7.8

#278 Fumon554 opened 3 months ago
0
几何图形特征

#277 songge25 opened 3 months ago
0
请问，imagenet上的数据集精度是如何得到的？

#276 xiguadong opened 3 months ago
0
在Huggingface上使用你们的模型的时候出现无法加载tokenizer的错误

#275 IF-chan opened 4 months ago
0
finetune时的问题

#274 DengyingFu opened 4 months ago
1
请问在chinese-clip获取到图文特征json数据后，如何在clip-retrieval使用呢，想搭建自己的demo

#273 IF-chan opened 4 months ago
8
使用flash-Attention训练，loss不下降，acc也不增加

#272 gobigrassland opened 4 months ago
2
finetune时如何运行sh文件

#271 DengyingFu closed 2 months ago
0
使用脚本怎么跑不了了，一直卡着什么输出也没有

#270 Elibeau closed 4 months ago
0
请问将cn_clip转为onnx格式后，有没有不使用cn_clip库进行预处理的思路，我们在尝试用C++部署

#269 Amphetaminewei closed 3 months ago
0
尝试使用CLIP模型来进行文搜文遇到的问题

#268 Amphetaminewei closed 3 months ago
1
转换torchscript报错

#267 istoneyou opened 4 months ago
0
关于训练过程一个batch内，text文本标签存在较多重复，是否会导致训练不收敛

#266 created-Bi opened 4 months ago
3
论文初始化image encoder的模型参数

#265 gobigrassland opened 4 months ago
2
关于跨模态检索的召回率疑问，请大佬解答一番

#264 954028118 opened 4 months ago
1
torch onnx tensorrt 推理速度和deployment.md评估效果差异较大

#263 GaoDongWan opened 4 months ago
1
load_from_name function logic problem

#262 JohnnMa opened 4 months ago
0
AttributeError: 'Namespace' object has no attribute 'use_flash_attention'

#261 IF-chan closed 3 months ago
4
ImportError: cannot import name 'convert_state_dict' from 'cn_clip.clip.model'

#260 IF-chan closed 4 months ago
2
这个支持词性标注吗

#259 hu394854434 closed 4 months ago
1
AttributeError: 'NoneType' object has no attribute 'get'

#258 5zjk5 opened 4 months ago
5
为什么使用gpu进行推理比cpu还慢？

#257 liumingzhu6060 closed 4 months ago
1
转换onnx失败

#256 bit1002lst opened 4 months ago
2
相似度计算实测下来完全不对

#255 yumianhuli1 closed 4 months ago
3
train.py中sampler.set_epoch的判断不够严谨

#254 ChesonHuang closed 4 months ago
0
源码中的一些bug，请确认下～

#253 ChesonHuang closed 4 months ago
0
使用默认超参数在COCO-CN上微调，无法收敛

#252 xiuxiuxius opened 5 months ago
0
无法识别图片里面的中文

#251 timchenxiaoyu opened 5 months ago
2
中文和英文同时匹配图片时中文得分非常低

#250 HiddenMarkovModel opened 5 months ago
0
请问CLIPTokenizer和Chinese-CLIP中的什么对应呢？

#249 AnnyOrange opened 5 months ago
5
inference速度慢

#248 wutongtong opened 6 months ago
0
请问这个项目有图片转成中文描述的功能吗，类似clip的interrogator

#247 mmix574 closed 5 months ago
1
训练时存在内存泄露的问题。

#246 liudongcheng opened 6 months ago
1
基于模型微调后得到epoch_latest.pt，但是为什么训练完推理不了，我试了好几个服务器跑完都不行

#245 Luhuanz closed 6 months ago
2
Text-to-Image Retrieval & Image-to-Text Retrieval速度太慢，怎么进行合理加速

#244 chengjianjie opened 6 months ago
0
你好，我想在自己的数据集上微调，两张3090够吗

#243 biedaxiaohua closed 5 months ago
1
论文无法复现

#242 nlpplus opened 6 months ago
1
转换后的onnx 模型，在运行同一张图像时，输出不一致

#241 MengTanOwn opened 6 months ago
1
使用自己构建的数据集，微调loss不下降

#240 LiqunW closed 6 months ago
1
关于自己训练数据集模型推理的问题

#239 ipeaking opened 6 months ago
1
请教下多机多卡的配置应该如何配置

#238 chengjianjie opened 6 months ago
0
训练数据集图文多对多的情况，直接采样分类交叉熵是不是有问题？训练集应该怎么处理？

#237 FoolishMao opened 7 months ago
0
loss 为0

#236 wwangxinhao opened 7 months ago
1
hugging face demo访问出错了

#235 xiuxiuxius closed 4 months ago
1
预训练模型授权洽詢

#234 fann1993814 closed 4 months ago
1
自建数据集微调的疑问

#233 MarcoZuo closed 7 months ago
1
相同的文本词，使用encode_text产出的结果不一样

#232 m3n78amv2 opened 7 months ago
2
您好，我想进行文搜图的任务。在notebook的演示流程中有需要8G显存的微调流程，但是我的笔记本只有6G，请问是否可以不进行微调，直接进行尝试？

#231 qinghuanyyz opened 7 months ago
3
Add pytorch to coreml conversion script

#230 manymuch closed 7 months ago
2

Previous Next