OFA-Sys/Chinese-CLIP

Issues

• 文本行为特征定位坐标

  #374 opened a month ago by shengruicai
  1

• Add tqdm to finetune process

  #373 opened a month ago by rujiangyouyou
  0

• 使用下列代码，如何加载自己训练的 'ViT-L-14'模型指定路径的.pt文件。（比如加载Epoch为2的./data/model/epoch2.pt模型）

  #345 opened 5 months ago by laitianan
  7

• 怎样批量encode图片？

  #372 opened a month ago by Royhuiy
  0

• 关于调用cn_clip进行特征提取报错格式错误的问题

  #371 opened a month ago by Seing-yu
  0

• 报错：ModuleNotFoundError: No module named 'flash_attn.flash_attention'

  #370 opened a month ago by Charimanhua
  0

• Insecure called of torch.load in the function of load_from_name

  #369 opened 2 months ago by alpha731
  0

• 模型微调异常

  #362 opened 3 months ago by darkprices
  1

• onnx部署没问题，tensorrt部署有问题

  #350 opened 4 months ago by paulpaul91
  2

• 构建数据的流程能不能更具体点，或者提供原始转tsv和jsonl的脚本

  #366 opened 2 months ago by ctgushiwei
  0

• 单机四卡训练报错

  #367 opened 2 months ago by ctgushiwei
  0

• 关于clip score的计算方式

  #368 opened 2 months ago by thbupt
  0

• welcome to clip.cpp project, support openai clip and chinese clip model, quantization type(q4,q8_0,f16,f32), metal gpu and cuda speed

  #359 opened 3 months ago by yysu-888
  0

• 多卡微调多线程报错：Exception in thread

  #352 opened 4 months ago by fanshuaiyao
  6

• No model called 'RoBERTa-wwm-Base'

  #365 opened 3 months ago by ranck626
  0

• ncnn

  #364 opened 3 months ago by Heiheicxx
  0

• Fluctuating embedding value

  #363 opened 3 months ago by blizzardwj
  0

• The protocol error while running the build the build_lmdb_dataset.py

  #361 opened 3 months ago by darkprices
  0

• 微调好的finetune模型，做图片分类，用api方法不行

  #360 opened 3 months ago by xiaoAugenstern
  0

• 不兼容gguf格式？

  #356 opened 4 months ago by aiportor
  1

• 保存微调后模型，效果没有微调过程中好

  #358 opened 4 months ago by wangly1998
  0

• MUGE微调报错：ERROR:torch.distributed.elastic.multiprocessing.api:failed (exitcode: -7)

  #357 opened 4 months ago by Guan-Y
  0

• cuDNN mismatch tensorRT（onnx，tensorRT）

  #355 opened 4 months ago by ranck626
  0

• 有简单的办法生成热力图吗，很需要

  #354 opened 4 months ago by saymyname77
  0

• 支持英文吗？

  #353 opened 4 months ago by Phoenix724
  0

• get_image_features 段错误(吐核)

  #351 opened 4 months ago by yanqiangmiffy
  0

• ViT-H/14训练脚本、部署能够提供吗？（使用ViT-B/16更改为ViT-H/14脚本，训练起来，模型各个方面参数都是下降的，但是部署成tensorRT，用图片预测，效果堪忧）

  #322 opened 6 months ago by laitianan
  3

• 可以在fiftyone中实现文搜图跨模态检索吗

  #348 opened 5 months ago by seasnakes
  0

• AttributeError: 'ChineseCLIPConfig' object has no attribute 'hidden_size'

  #347 opened 5 months ago by Manjiz
  0

• 大佬可以训练一个像jina-clip-v1的中文版这样的话，以后做搜索或是 构建多模态 RAG 应用 就牛了~~

  #346 opened 5 months ago by gg22mm
  0

• 通过pip安装的cn_clip库方法load_from_name不能加载本地模型

  #342 opened 5 months ago by qiuchen001
  1

• 翻译有点糟糕

  #344 opened 5 months ago by ccl-private
  0

• The context length to use; all baseline models use 52 as the context length

  #343 opened 5 months ago by qiuchen001
  0

• 请教下，提取图文特征json数据后，应该怎么配合clip-retrieval使用搭建自己的检索demo？

  #341 opened 5 months ago by JingMo
  0

• 微调太复杂 简单的东西搞得这么复杂

  #338 opened 5 months ago by wangsheng1991
  0

• 进行 rn50 model-arch 转 onnx 的时候会报如下错误

  #337 opened 5 months ago by pisces365
  1

• Is there a CLIP model that is smaller than ResNet50 and RBT3?

  #336 opened 5 months ago by guanfuchen
  0

• 运行，报错了，bash run_scripts/muge_finetune_vit-b-16_rbt-base.sh ${DATAPATH}

  #335 opened 5 months ago by Renjunhua30
  0

• 模型加载问题

  #334 opened 5 months ago by dayuanjie
  0

• 使用MUGE微调时,训练日志里面的Image2Text Acc与评测时的R@1的召回指标不一致？

  #333 opened 5 months ago by dulibubai
  0

• onnx模型转换报错

  #328 opened 5 months ago by guanxiaoguan
  1

• COCO-CN数据集申请

  #332 opened 6 months ago by zlw1113
  0

• CN-CLIPViT-B/16模型在输入与向量化图片无关的输入时，如！@#￥，仍能有结果匹配

  #330 opened 6 months ago by wangyan828
  1

• 能否通过文本向量化中加入指令优化图片召回率

  #331 opened 6 months ago by Amphetaminewei
  0

• “参数命名不同”：Huggingface上的CLIP模型和GITHUB上的 CLIP模型 参数命名不同

  #329 opened 6 months ago by ranck626
  0

• 各位大佬，专家，请问该模型能否在Windows系统下进行运行呢？欢迎来交流

  #327 opened 6 months ago by LJK12581
  0

• Question: gradient accumulate

  #326 opened 6 months ago by zwhus
  0

• how to continue training with self-data? or Full Parameter Fine-Tuning

  #325 opened 6 months ago by StatYuDoc
  1

• 手动修改ChineseCLIPVisionModel to ChineseCLIPVisionModelWithProjection 失败。

  #324 opened 6 months ago by ranck626
  1

• ChineseCLIPVisionModel in huggingface can used?

  #323 opened 6 months ago by ranck626
  3