Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / FlagOpen/FlagEmbedding issues and pull requests

#99 - 对单个词的embedding计算不符合预期

Issue - State: closed - Opened by Awezome about 1 year ago - 5 comments

#98 - 关于fine-tune涉及到的参数

Issue - State: closed - Opened by runwean about 1 year ago - 1 comment

#97 - finetune调参相关咨询

Issue - State: closed - Opened by PNightOwlY about 1 year ago - 1 comment

#96 - 关于预训练的一些细节咨询

Issue - State: closed - Opened by stanpcf about 1 year ago - 3 comments

#95 - 预训练与微调两阶段，哪个阶段效果收益更大

Issue - State: closed - Opened by Qzsl123 about 1 year ago - 1 comment

#94 - 关于构造1.2亿训练数据对

Issue - State: closed - Opened by JDZW2014 about 1 year ago - 2 comments

#93 - 关于预训练数据的问题

Issue - State: closed - Opened by Harmon-l about 1 year ago - 9 comments

#92 - 如何给Trainer加上评估，我想针对检索和分类进行评估。

Issue - State: closed - Opened by 1006076811 about 1 year ago - 5 comments

#91 - Sequence Length 512是最大token数还是最大字符长度

Issue - State: closed - Opened by ftufkc about 1 year ago - 7 comments

#90 - 微调阶段，loss 一般降到多少，算是不错的结果

Issue - State: closed - Opened by zouzoutingting about 1 year ago - 3 comments

#89 - 微调之后效果较差

Issue - State: closed - Opened by 34127chi about 1 year ago - 1 comment

#88 - 有无相关论文链接

Issue - State: closed - Opened by weiiWill about 1 year ago - 1 comment

#87 - 请问使用BGE做分类Fintune的方法

Issue - State: closed - Opened by 1006076811 about 1 year ago - 4 comments

#86 - 预训练方法选择

Issue - State: closed - Opened by wing7171 about 1 year ago - 5 comments

#85 - Error when running the finetune script

Issue - State: closed - Opened by FlyingBird11 about 1 year ago - 2 comments

#84 - 关于微调后下游任务

Issue - State: closed - Opened by Yyy11181 about 1 year ago - 5 comments

#83 - finetune model

Issue - State: closed - Opened by Cloopen-ReLiNK about 1 year ago - 4 comments

#82 - 1

Issue - State: closed - Opened by walker313504 about 1 year ago

#81 - 预训练报错

Issue - State: closed - Opened by zzxxxl about 1 year ago - 18 comments

#80 - Fix data parallelism

Pull Request - State: closed - Opened by Muennighoff about 1 year ago - 4 comments

#79 - 关于normalized参数的使用方法

Issue - State: closed - Opened by Muennighoff about 1 year ago - 1 comment

#78 - 关于C-MTEB评测

Issue - State: closed - Opened by Saicat about 1 year ago - 1 comment

#77 - 微调报错：RuntimeError: Could not infer dtype of NoneType

Issue - State: closed - Opened by caimanjing about 1 year ago - 6 comments

#75 - N

Issue - State: closed - Opened by rashlab about 1 year ago

#73 - 我遇到了微调的列表越界错误

Issue - State: closed - Opened by haldon98 about 1 year ago - 4 comments

#72 - 怎么提交自己训练模型预测的结果到MTEB上

Issue - State: closed - Opened by Young1993 about 1 year ago - 1 comment

#71 - 关于基于retroMAE pre-train 的问题

Issue - State: closed - Opened by runwean about 1 year ago - 3 comments

#70 - 关于loss的计算

Issue - State: closed - Opened by zhaobinNF about 1 year ago - 1 comment

#69 - Add multilingual-e5 models & additional MTEB zh datasets

Pull Request - State: closed - Opened by Muennighoff about 1 year ago

#67 - 数据引用错误

Issue - State: closed - Opened by campuslifeceo about 1 year ago - 2 comments

#66 - 请问下，为什么预训练和微调的数据格式不一致？

Issue - State: closed - Opened by asenasen123 about 1 year ago - 3 comments

#65 - 关于Finetune的一些小问题，没有early stop和best_model save

Issue - State: closed - Opened by PNightOwlY about 1 year ago - 3 comments

#64 - 增量预训练后，相似度变得很高

Issue - State: closed - Opened by zouzoutingting about 1 year ago - 7 comments

#63 - BAAI/bge-large-zh和BAAI/bge-large-zh-noinstruct两个版本的区别是什么？

Issue - State: closed - Opened by Zado-Tan about 1 year ago - 2 comments

#62 - 训练集构成

Issue - State: closed - Opened by Wenze7 about 1 year ago - 3 comments

#61 - error：FlagEmbedding/C_MTEB/zh/train.jsonl doesn't exist

Issue - State: closed - Opened by jllliang about 1 year ago

#60 - Fix typos & improve readability

Pull Request - State: closed - Opened by Muennighoff about 1 year ago - 1 comment

#58 - 关于微调参数“tempeture”等

Issue - State: closed - Opened by Yyy11181 about 1 year ago - 33 comments

#57 - 想问下检索任务在评测时使用的检索类型

Issue - State: closed - Opened by Wenze7 about 1 year ago - 1 comment

#56 - 关于微调的问题

Issue - State: closed - Opened by bombia about 1 year ago - 3 comments

#54 - Schedule里的Multilingual model是考虑中英双语的吗？

Issue - State: closed - Opened by firezym about 1 year ago - 2 comments

#53 - 请问你们的的原理之类的有在Arxiv上面发论文吗

Issue - State: closed - Opened by bombia about 1 year ago - 1 comment

#51 - Error while trying to use custom DistilBERT model in the pretraining script

Issue - State: closed - Opened by debashisai about 1 year ago - 11 comments

#50 - 增大模型最长长度限制

Issue - State: closed - Opened by knowledge-llz about 1 year ago - 5 comments

#48 - 相似度的归一化相比于OpenAI的Embedding差距还是不小的

Issue - State: closed - Opened by aierwiki about 1 year ago - 8 comments

#47 - 可以介绍一下如何从wudao中挖掘训练样本对的思路吗？

Issue - State: closed - Opened by campuslifeceo about 1 year ago - 5 comments

#45 - 微调数据咨询

Issue - State: closed - Opened by Yyy11181 about 1 year ago - 4 comments

#43 - 请问在微调的过程中，数据会打乱吗？

Issue - State: closed - Opened by NLPJCL about 1 year ago - 1 comment

#42 - Domain tokens and pretrain or fine-tuning

Issue - State: closed - Opened by erlakshmi123 about 1 year ago - 1 comment

#41 - 文本长度和相似度的关系

Issue - State: closed - Opened by xuyungit about 1 year ago - 8 comments

#40 - hf上c-mteb的文件编码格式

Issue - State: closed - Opened by Wenze7 about 1 year ago - 1 comment

#39 - 各个小项的指标

Issue - State: closed - Opened by Wenze7 about 1 year ago - 1 comment

#38 - 关于FineTune

Issue - State: closed - Opened by BrownTen about 1 year ago - 3 comments

#37 - 关于negatives_cross_device的疑问

Issue - State: closed - Opened by zhaobinNF about 1 year ago - 1 comment

#36 - [finetune]对文档检索任务应该怎么调优

Issue - State: closed - Opened by kouhigh55 about 1 year ago - 6 comments

#35 - [Instructions] Any other query instructions for other tasks?

Issue - State: closed - Opened by jlia0 about 1 year ago - 3 comments

#33 - 能否开放api服务,给其他应用调用.

Issue - State: closed - Opened by guyuexue about 1 year ago - 1 comment

#32 - 如果使用bge_large_en，instruction该如何写？

Issue - State: closed - Opened by ybalbert001 about 1 year ago - 2 comments

#31 - 文档检索任务，在query上加instruction的效果反而没有不加的好

Issue - State: closed - Opened by wanggdnju about 1 year ago - 3 comments

#30 - 关于预训练和微调数据量的问题

Issue - State: closed - Opened by EliverQ about 1 year ago - 1 comment

#29 - 请问这报错什么原因，mac跟linux都一样不行

Issue - State: closed - Opened by msute about 1 year ago - 5 comments

#28 - 硬件要求

Issue - State: closed - Opened by Donglong about 1 year ago - 2 comments

#27 - 依据任务特性修改网络的两个问题

Issue - State: closed - Opened by KekeWa about 1 year ago - 1 comment

#26 - Pre-train 示例疑问

Issue - State: closed - Opened by CuteMing about 1 year ago - 5 comments

#25 - parameter: batch size per gpu

Issue - State: closed - Opened by hjq133 about 1 year ago - 4 comments

#24 - Cross-encoder Rerank？

Issue - State: closed - Opened by knowledge-llz about 1 year ago - 1 comment

#22 - 这个可以用CPU跑吗

Issue - State: closed - Opened by jzhelp99 about 1 year ago - 1 comment

#21 - 开心和石头的相似度得分都0.8，这也太高了吧

Issue - State: closed - Opened by NaviQuant about 1 year ago - 27 comments

#20 - 请问输出的embedding是 encoder的还是decoder的？

Issue - State: closed - Opened by dotsonliu about 1 year ago - 1 comment

#19 - 请问fine-tuning的训练数据是如何收集的

Issue - State: closed - Opened by BiggyBing about 1 year ago - 4 comments

#18 - self-hosting: what are the estimated GPU size requirements?

Issue - State: closed - Opened by skeptrunedev about 1 year ago - 1 comment

#17 - hybrid search using both BGE and BM25

Issue - State: closed - Opened by bugface about 1 year ago - 5 comments

#15 - FlagModel 没办法使用cpu吗

Issue - State: closed - Opened by ReverseSystem001 about 1 year ago - 2 comments

#14 - Hope to get in touch

Issue - State: closed - Opened by JimmyMa99 about 1 year ago

#13 - Update README_zh.md

Pull Request - State: closed - Opened by ng-fukgin about 1 year ago

#12 - huggingface上使用的Sentence Similarity是什么方法

Issue - State: closed - Opened by coolmian about 1 year ago - 4 comments

#11 - 线上推理性能咨询

Issue - State: closed - Opened by KekeWa about 1 year ago - 12 comments

#10 - 支持中英文双语的模型？

Issue - State: open - Opened by knowledge-llz about 1 year ago - 22 comments

#9 - Langchain BGE implementation

Issue - State: closed - Opened by manmax31 about 1 year ago - 4 comments

#8 - readme为空

Issue - State: closed - Opened by cristianohello about 1 year ago - 1 comment

#7 - 怎么扩展输入文本的长度

Issue - State: closed - Opened by yuanjie-ai about 1 year ago - 10 comments

#6 - 难负例的挖掘

Issue - State: closed - Opened by NLPJCL about 1 year ago - 4 comments

#5 - Multilingual Models

Issue - State: open - Opened by Siegi96 about 1 year ago - 11 comments

#3 - C-MTEB部分数据集评测分数为0

Issue - State: closed - Opened by izhx about 1 year ago - 4 comments

#2 - MTEB榜单没看懂，为什么说BGE是最佳？

Issue - State: closed - Opened by yucc-leon about 1 year ago - 1 comment

#1 - 不相关的句子的相似度得分很高

Issue - State: closed - Opened by goggeryang about 1 year ago - 18 comments