Maria

Results 10 issues of Maria

I have trained model for chinese weibo data ,how could i predict single line?

I was training by chinese weibo data ,but when running decode.py,it failed to calculate rouge score .This is the issue which have the same problem but still remain to be...

训练预测的时候是正常的 ![image](https://user-images.githubusercontent.com/28859589/60876479-8c922100-a26e-11e9-8bfa-8914b0ea5983.png) ![image](https://user-images.githubusercontent.com/28859589/60876493-961b8900-a26e-11e9-847d-17d426bc4c74.png) 实际上2700也算是收敛了,效果不会这样吧。但是我使用demo_inference的时候就预测不出来了, python demo_inference.py --batch_size=1 \ --checkpoint=../logs/model.ckpt-2742\ --image_path_pattern=./datasets/data/fsns/temp/%02d.png 图片是000.png 最后只得到了444444,也就只有第一个数字,不知道是什么问题?哪里设置出错了么?

windows可以打开,但是在linux加载不了了,就是在fuseki_conf.ttl里面路径设置了之后打开3030网页看不到任何dataset

tuple_extractor.py这个函数抽取出来的candidates_tuple如下:{('', '', ''): ['', '莫妮卡', 0.00027995862]}, 导致tuple_filter.py里面的X.append([features[9][0][1]])索引出错。这里 score = [entity]+[s for s in candidate_entitys[entity][0:1]]是不是写错了? > def extract_tuples(self,candidate_entitys,question): '''''' candidate_tuples = {} for entity in candidate_entitys: #得到该实体的所有关系路径 starttime=time.time() relations = GetRelationPaths(entity)...

文字检测部分有时候出错,是不是得增加数据集训练?检测之后如何剔除掉非汉字的部分识别呢?

想复现一下论文的结果,求告知下参数呢? For reference, our modified baseline achieves 96.9% rank-1 and 82.0% mAP accuracy on Veri-776 benchmark 我自己跑了8个epoch,Acc就达到了0.97。但是验证的时候,mAP=56.7%,Rank-1=86.7%;而我在第4个epoch的时候,Acc也有0.92了,mAP =55.2% ,Rank-1=84.5%,我修改了一部分参数,不清楚是不是改错了,似乎相差挺大的。

![image](https://user-images.githubusercontent.com/28859589/71343623-6271b880-259b-11ea-8c37-2a4805d2a172.png) ![image](https://user-images.githubusercontent.com/28859589/71343661-7ae1d300-259b-11ea-84b3-1978c182aa2c.png) Test some images from midv datasets with the provided model,the above seems difficult to identify accurately,can someone explain?

华为的ascend 910机器,跑的是paddleocr的文本识别训练。 运行命令: /home/ma-user/anaconda3/envs/MindSpore/bin/python -u train.py -c configs/rec/ch_ppocr_v2.0/rec_chinese_common_train_v2.0.yml -o use_npu=True 系统: Linux notebook-2a7fcf5e-f8cc-473f-9414-a6ff8ebbf05a 4.19.36-vhulk1907.1.0.h619.eulerosv2r8.aarch64 #1 SMP Mon Jul 22 00:00:00 UTC 2019 aarch64 aarch64 aarch64 GNU/Linux Paddle: paddlepaddle_npu-0.0.0-cp37-cp37m-linux_aarch64.whl 完整报错: [2022/10/28...

recognition

env文件下填上了apikey: OPENAI_API_KEY='sk-MX8...Yk' 系统win10,安装[prm-cli](https://github.com/wangrongding/prm-cli)也要很久,所以我安装的cnpm镜像 npm install -g cnpm --registry=https://registry.npm.taobao.org 运行cnpm i没问题,运行cnpm run dev报错如下: ![image](https://user-images.githubusercontent.com/28859589/216377930-c415fb75-be71-4920-a93e-7dc81e17fe1b.png)