search

1adrianb / face-alignment

:fire: 2D and 3D Face alignment library build using pytorch
https://www.adrianbulat.com
BSD 3-Clause "New" or "Revised" License
6.99k stars 1.34k forks source link

Smaller pretrained models #124

Closed SergeiSamuilov closed 3 years ago

SergeiSamuilov commented 5 years ago

Adrian, are fewer stacks models available at the moment? Large model is great, but it would be really nice to speed things up a bit.

mibrahimy commented 5 years ago

Same query. How can we use the NetworkSize.SMALL option? From where can we download the required smaller landmark detector?

ierezell commented 5 years ago

Hi, thanks a lot for this beautiful work, it run perfectly fine !

Same question here, I tried to modify the api to gain speed with NetworkSize.TINY or to call it with fa = face_alignment.FaceAlignment(face_alignment.LandmarksType._3D, flip_input=False, device='cpu', network_size=1)

but when loading the model I got a urllib.error.HTTPError: HTTP Error 404: Not Found

Traceback here : Traceback (most recent call last): File "getLandmarks3D.py", line 12, in <module> flip_input=False, device='cpu', network_size=1) File "/home/pedro/.local/lib/python3.7/site-packages/face_alignment/api.py", line 86, in __init__ network_name, os.path.join(fan_temp_path)) File "/usr/lib/python3.7/urllib/request.py", line 247, in urlretrieve with contextlib.closing(urlopen(url, data)) as fp: File "/usr/lib/python3.7/urllib/request.py", line 222, in urlopen return opener.open(url, data, timeout) File "/usr/lib/python3.7/urllib/request.py", line 531, in open response = meth(req, response) File "/usr/lib/python3.7/urllib/request.py", line 641, in http_response 'http', request, response, code, msg, hdrs) File "/usr/lib/python3.7/urllib/request.py", line 569, in error return self._call_chain(*args) File "/usr/lib/python3.7/urllib/request.py", line 503, in _call_chain result = func(*args) File "/usr/lib/python3.7/urllib/request.py", line 649, in http_error_default raise HTTPError(req.full_url, code, msg, hdrs, fp) urllib.error.HTTPError: HTTP Error 404: Not Found

can you please put the link for the weights of a smaller network ?

Selimonder commented 4 years ago

+1 to this. Execution with smaller network size yields 404.

danilopeixoto commented 4 years ago

Hi Adrian, Do you have pretrained weights for tiny, small or medium network size hosted online?

The face_alignment package for Python is great! Thanks.

moncio commented 3 years ago

I have the same questions than the other colleagues, do we have available pretrained models for smaller networks? I'm pretty interested to use TINY or SMALL to gain some fps. Great job and amazing library @1adrianb !!!

1adrianb commented 3 years ago

Duplicate of #4