yemount / pose-animator

Apache License 2.0
8.73k stars 964 forks source link

Loading FaceMesh model takes very long time! #21

Open CyberFork opened 4 years ago

CyberFork commented 4 years ago

static_img.html camera.html both stoped at Loading FaceMesh model...

chenweigh commented 4 years ago

I have same problem now. do you solve it ?

lonely-root-number2 commented 4 years ago

have same problem. And console have 3 error about network.

CyberFork commented 4 years ago

I have same problem now. do you solve it ?

Used lantern

csingh27 commented 4 years ago

It's a GPU issue. Get a more powerful GPU

lonely-root-number2 commented 4 years ago

I have found that it is caused by the network,there are 4 files load error for the bad network. Can we solve the problems by dowdloading the file in advance then modify the source code and load the local file? I am not good at the frant, can't do it well. the 4 file are following: https://storage.googleapis.com/tfjs-models/savedmodel/posenet/mobilenet/quant2/100/group1-shard1of2.bin https://storage.googleapis.com/tfjs-models/savedmodel/posenet/mobilenet/quant2/100/group1-shard2of2.bin https://tfhub.dev/tensorflow/tfjs-model/blazeface/1/default/1/model.json?tfjs-format=file https://tfhub.dev/mediapipe/tfjs-model/facemesh/1/default/1/model.json?tfjs-format=file

mczha commented 4 years ago

I have found that it is caused by the network,there are 4 files load error for the bad network. Can we solve the problems by dowdloading the file in advance then modify the source code and load the local file? I am not good at the frant, can't do it well. the 4 file are following: https://storage.googleapis.com/tfjs-models/savedmodel/posenet/mobilenet/quant2/100/group1-shard1of2.bin https://storage.googleapis.com/tfjs-models/savedmodel/posenet/mobilenet/quant2/100/group1-shard2of2.bin https://tfhub.dev/tensorflow/tfjs-model/blazeface/1/default/1/model.json?tfjs-format=file https://tfhub.dev/mediapipe/tfjs-model/facemesh/1/default/1/model.json?tfjs-format=file

how to modify to local file

chenweigh commented 4 years ago

I have same problem now. do you solve it ?

Used lantern

thank you

lonely-root-number2 commented 4 years ago

I have found that it is caused by the network,there are 4 files load error for the bad network. Can we solve the problems by dowdloading the file in advance then modify the source code and load the local file? I am not good at the frant, can't do it well. the 4 file are following: https://storage.googleapis.com/tfjs-models/savedmodel/posenet/mobilenet/quant2/100/group1-shard1of2.bin https://storage.googleapis.com/tfjs-models/savedmodel/posenet/mobilenet/quant2/100/group1-shard2of2.bin https://tfhub.dev/tensorflow/tfjs-model/blazeface/1/default/1/model.json?tfjs-format=file https://tfhub.dev/mediapipe/tfjs-model/facemesh/1/default/1/model.json?tfjs-format=file

how to modify to local file

I don't know. It is just an idea and I am not sure it works.

xmy1024 commented 4 years ago

I use proxy can solve this issues

Sba-Stuff commented 4 years ago

This Project Requires Internet Connection, Won't work offline.

wzhqwq commented 4 years ago

Hey guys, I've found a way to fix the network issue. You can try my forked repository out.

Gjts commented 4 years ago

hey it's ok

------------------ 原始邮件 ------------------ 发件人: "yemount/pose-animator" <notifications@github.com>; 发送时间: 2020年8月12日(星期三) 下午4:14 收件人: "yemount/pose-animator"<pose-animator@noreply.github.com>; 抄送: "Subscribed"<subscribed@noreply.github.com>; 主题: Re: [yemount/pose-animator] Loading FaceMesh model takes very long time! (#21)

Hey guys, I've found a way to fix the network issue. You can try my forked repository out.

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

Sba-Stuff commented 4 years ago

Hey guys, I've found a way to fix the network issue. You can try my forked repository out.

Will It work without internet? Offline? It downloads models from tensorflow every time.

How did you fixed it?

wzhqwq commented 4 years ago

Hey guys, I've found a way to fix the network issue. You can try my forked repository out.

Will It work without internet? Offline? It downloads models from tensorflow every time.

How did you fixed it?

I changed the URLs of two models into some URLs which is accessible to Chinese. I discovered two packages in npm that download model without tfhub.dev and copied their URLs into facemesh module. Finally I moved this module to "third_party" to keep the changes.

Naveeie commented 3 years ago

I have same problem now. do you solve it ?

I have the same problem, was there any solution to that ?