deepinsight / deepinsight/insightface
Using ArcFace as a feature extractor
- Dominant language
- Python
- Stars
- 29.7k
- Forks
- 6.1k
- PR merge metrics
- No merged PRs in 30d
Description
Hi,
I have downloaded the `LResNet100E-IR, ArcFace@ms1m-refine-v2` model and I am planning to use it as a feature extractor.
When I list the last 10 layers in the pre-trained network I see the following:
```
['bn1_output',
'dropout0_output',
'pre_fc1_weight',
'pre_fc1_bias',
'pre_fc1_output',
'fc1_gamma',
'fc1_beta',
'fc1_moving_mean',
'fc1_moving_var',
'fc1_output']
```
I need to remove the fully connected layers and I am not sure what do `pre_fc1_weight`, `pre_fc1_bias`, and
`pre_fc1_output` do? Is there a sketch of the network available anywhere? Are these layers fully connected layers?
I would really appreciate if you could help me with my questions.
Thanks
Contributor guide
No contributing guide indexed for this repository
Assessment
This issue has not been assessed yet.