search

SullyChen / Autopilot-TensorFlow

A TensorFlow implementation of this Nvidia paper: https://arxiv.org/pdf/1604.07316.pdf with some changes
MIT License
1.25k stars 425 forks source link

Getting same steering angle for "run.mp4" #39

Open BatoolNaqvi opened 4 years ago

BatoolNaqvi commented 4 years ago

I have trained your provided model on the same data set as you have used. But when I run "run_dataset" it is giving the same steering angle prediction for the whole video. Although the mse loss value I got was 0.2 approx. But still there is some issue with the output. What could possibly go wrong? If you can help me with this.

maxmstrmn commented 4 years ago

Same here. Have you solved your problem?

BatoolNaqvi commented 4 years ago

Same here. Have you solved your problem?

Not yet. Did you find any solution?

maxmstrmn commented 4 years ago

No, unfortunately not :(

coral-agc commented 4 years ago

I'm having that same issue, as well- let me know if you figure anything out!

theshuv commented 4 years ago

Did any one found the solution ? , plz share

Wanli-Jiang commented 4 years ago

same issue!

SullyChen commented 4 years ago

Hey everyone, so sorry for being absent from this project for a while (quite busy with undergrad now). Is everyone using the standard test scripts? I may add a seed to the python's rnd generator so that everyone can get consistent results from now on. Also, which (of the two) datasets are you using for training?

taimeile commented 1 year ago

使用的是6万多张图片的数据集，我也遇到了这样的问题，有人解决了吗？方向盘一直不动

Shruti11234 commented 1 week ago

@SullyChen in which file you have made that change?