Jar7

Could you please provide the log for training with the default configuration?

3

I train the model with the provided dataset and configuration. And the initial loss is huge, e.g. 8~9e+10. I think there is something wrong with my training. Could you please...

re-implementation with tensorflow cannot achieve the same accuracy on mxnet.

1

Hi, I apply insightface for face recognition with my own dataset by simply add 2 fc layers and output the index of the class. Firstly I trained it on mxnet,...

Help! core dump: trained on gpu, search on disk

4

UPDATE: Sorry I found it occurs even with faiss-cpu. There is nothing to do with gpu.(Although I did uninstall faiss-gpu and reinstall faiss-cpu). Here is the code. You can run...

run test_widerface.py failed

1

``` Traceback (most recent call last): File "test_widerface.py", line 134, in detect(opt=opt) File "test_widerface.py", line 43, in detect model(torch.zeros(1, 3, imgsz, imgsz).to(device).type_as(next(model.parameters()))) # run once File "/opt/conda/lib/python3.6/site-packages/torch/nn/modules/module.py", line 1102, in...

大佬您好， 我测了下您的pretrained model，分数都是稳定的0.9+，包括小人脸。希望能复现您的结果。 想请教下， 1. 您用的是4*4还是9*9的拼图，训练的时候拼图占比多大呀？ 2. 分数range不够稳定，会是什么原因呢？我看原作者biubug6的模型，分数range也很大，请问您有做其他参数修改吗？还是说pose分支很关键？ 3. 目前我是对已经增强的图拼在一起，再resize，过滤掉边缘没有眼睛的人脸，这样对吗？ 谢谢！！

Hello, I am working on a celebrity identification task. I just found that the IQIYI-VID dataset is what I need, however, the download link is no longer available. The dataset...

memory leak when index.search

6

``` import faiss from facebank import * import pickle import time import gc #index = faiss.index_factory(512, "IVF16384, Flat") index = faiss.read_index('log/features_exp2/feature_00550220.index') index_ivf = faiss.extract_index_ivf(index) index_ivf.nprobe = 50 with open('log/features_exp2/records.add.00000000_00023729.pkl','rb') as...