Closed HeChengHui closed 5 months ago
Hello HeChengHui,
Let me know if the above are not clear enough.
@Vagver
I have tried to run the demo inference but there seems to be a mismatched of model?
infer_res34_x128_xgz.yaml
loads 'data/res18_x128_vert_xgz.pth'
but the link provided is res34_x128_vert_xgz
. Changing the name in the config causes an error: "Unexpected key(s) in state_dict:"
@HeChengHui
The links provided have been currently tested for the inference.py and test.py scripts in the root dir of the repo. The demo code will be updated soon.
The models provided by insightface have similar performance, but will be updated soon, too.
@Vagver
- Is the performance of both implementation the same as well? I see that yours is more recently updated.
I have tried to run the demo inference but there seems to be a mismatched of model?
infer_res34_x128_xgz.yaml
loads'data/res18_x128_vert_xgz.pth'
but the link provided isres34_x128_vert_xgz
. Changing the name in the config causes an error: "Unexpected key(s) in state_dict:"
Hi, Did you get data/res18_x128_vert_xgz.pth
?
@Echo-jyt Nope. I am using the InsughtFace version. Have yet to check this repo for any updates.
@Echo-jyt Nope. I am using the InsughtFace version. Have yet to check this repo for any updates.
Can InsightFace run directly on video?
@Echo-jyt Dont think so. Frames.
@Echo-jyt Dont think so. Frames.
OK , fine.Thanks!
Hello @Vagver, thank you for your work! I have a few questions to clarify.
1) what is the difference between this inference script and this inference script? 2) Do i only need to do preprocessing for images? since it is not mentioned in the demo folder. 3) Is the method implemented by Insightface the same as this?