RSN icon indicating copy to clipboard operation
RSN copied to clipboard

pretrained model suppurt?

Open liyupeng-ing opened this issue 4 years ago • 7 comments

Hi, nice work! Could you share the pretrained model about the RSN?

liyupeng-ing avatar Jul 22 '20 08:07 liyupeng-ing

Hi. I am not one of the authors, but I can provide you my trained model here: https://drive.google.com/drive/folders/1atyNvjT_cIXq3_objomu6IJHm4K8CmET?usp=sharing. In there, you can find weights, the config file used, and the COCO score for the validation set (2017). I used the folder 4xRSN18 (in exps folders) and got AP 0.724. The model was trained for 1.15 million iterations.

Tiago-Roxo avatar Dec 06 '20 12:12 Tiago-Roxo

Hi. I am not one of the authors, but I can provide you my trained model here: https://drive.google.com/drive/folders/1atyNvjT_cIXq3_objomu6IJHm4K8CmET?usp=sharing. In there, you can find weights, the config file used, and the COCO score for the validation set (2017). I used the folder 4xRSN18 (in exps folders) and got AP 0.724. The model was trained for 1.15 million iterations.

Hello, do you have a trained RSN18 model?

samsara177 avatar Dec 20 '20 05:12 samsara177

Hi. I am not one of the authors, but I can provide you my trained model here: https://drive.google.com/drive/folders/1atyNvjT_cIXq3_objomu6IJHm4K8CmET?usp=sharing. In there, you can find weights, the config file used, and the COCO score for the validation set (2017). I used the folder 4xRSN18 (in exps folders) and got AP 0.724. The model was trained for 1.15 million iterations.

there is something wrong with the website you provided. Could you please provide it again? Thank you very much。 大佬,能把模型的链接再发一下吗?感激不尽!

ghost avatar Jul 28 '21 08:07 ghost

Hi. I am not one of the authors, but I can provide you my trained model here: https://drive.google.com/drive/folders/1atyNvjT_cIXq3_objomu6IJHm4K8CmET?usp=sharing. In there, you can find weights, the config file used, and the COCO score for the validation set (2017). I used the folder 4xRSN18 (in exps folders) and got AP 0.724. The model was trained for 1.15 million iterations.

the link is invalid now. Could you share it again?

ghost avatar Jul 28 '21 09:07 ghost

不好意思,我已经换工作了,目前没有这个模型的参数文件了。

------------------ 原始邮件 ------------------ 发件人: "caiyuanhao1998/RSN" @.>; 发送时间: 2021年7月28日(星期三) 下午4:58 @.>; @.@.>; 主题: Re: [caiyuanhao1998/RSN] pretrained model suppurt? (#10)

Hi. I am not one of the authors, but I can provide you my trained model here: https://drive.google.com/drive/folders/1atyNvjT_cIXq3_objomu6IJHm4K8CmET?usp=sharing. In there, you can find weights, the config file used, and the COCO score for the validation set (2017). I used the folder 4xRSN18 (in exps folders) and got AP 0.724. The model was trained for 1.15 million iterations.

there is something wrong with the website you provided. Could you please provide it again? Thank you very much。 大佬,能把模型的链接再发一下吗?感激不尽!

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

samsara177 avatar Jul 28 '21 09:07 samsara177

Hi. I am not one of the authors, but I can provide you my trained model here: https://drive.google.com/drive/folders/1atyNvjT_cIXq3_objomu6IJHm4K8CmET?usp=sharing. In there, you can find weights, the config file used, and the COCO score for the validation set (2017). I used the folder 4xRSN18 (in exps folders) and got AP 0.724. The model was trained for 1.15 million iterations.

Hello, can you provide it again

lofyol avatar Oct 26 '21 03:10 lofyol

Hi. I am not one of the authors, but I can provide you my trained model here: https://drive.google.com/drive/folders/1atyNvjT_cIXq3_objomu6IJHm4K8CmET?usp=sharing. In there, you can find weights, the config file used, and the COCO score for the validation set (2017). I used the folder 4xRSN18 (in exps folders) and got AP 0.724. The model was trained for 1.15 million iterations.

Hello @Tiago-Roxo, @wql20000111, @yes-ai-do, and @samsara177 Does any of you have the weights above backed up and can share them with us?

orestis-z avatar Oct 29 '21 15:10 orestis-z