Tf-pose-estimation: how can I get the three mobilenet models?

Created on 18 Jul 2018  Â·  10Comments  Â·  Source: ildoonet/tf-pose-estimation

Hi, there are three mobilenet models you mentioned in project, but I only get one mobilenet models which is in models\graph\mobilenet_thin\graph.pb。 And I want to try the other two models,how can I get them?

Most helpful comment

@kimdwkimdw hello,can you provide the other two mobilenet models?
mobilenet_accurate and mobilenet_fast
thx

All 10 comments

I downloaded a CMU model by executing the bash file. You could find it under 'cmu/'
https://github.com/ildoonet/tf-pose-estimation/blob/master/models/graph/cmu/download.sh

I downloaded CMU and mobilenet_thin, but I want to get the other two mobilenet models: mobilenet_accurate and mobilenet_fast. How Can I get them???

@ildoonet @arlcurten
I've transplanted openpose code to embedded platform use mobilenet_thin model, It cost 1s in 256*192 net-in resolution。 I can accept the time consuming,but not satisfied with the accuracy, especially for some specified pose such as arms raised overhead. So, I want to try mobilenet_accurate model which mentioned in the project, And how can I get it , or How can I train it ???

@zjbabylove what embedded platform did you use??

@freAkiee
I use qualcomm platform based on opencv4.0。

Hi,
I want to test mobilenet_accurate on my platform, where can I get the weights. Please share the path.

thanks
regards
arun

@arun-kumark
sorry, I don't have it. It seems that the author does not provide it directly. I don't know if we can train it by using "models/pretrained/mobilenet_v1_1.0_224_2017_06_14/" pretrain weight.

@kimdwkimdw hello,can you provide the other two mobilenet models?
mobilenet_accurate and mobilenet_fast
thx

@ildoonet

Hi @ildoonet , I'd also appreciate if you could include the two other models, namely mobilenet_accurate and mobilenet_fast.

Thank you!

Was this page helpful?
0 / 5 - 0 ratings

Related issues

Minotaur-CN picture Minotaur-CN  Â·  4Comments

arnabdeypolimi picture arnabdeypolimi  Â·  5Comments

gen21kawa picture gen21kawa  Â·  5Comments

khaerulumam42 picture khaerulumam42  Â·  6Comments

juciny picture juciny  Â·  3Comments