Dear Edge Impulse Community,
I need a little help, I’m very new at experimenting with this platform and have been inspired to use it for my science project. I was recreating a project with an E-nose, and cloned that project data. When I began to collect data on my Seed Wio Terminal unfortunately I was guided to write all axes points with an _
This _ made it so when I try to flash the device it does not work I believe properly with the Wio Terminal.
I ran the original public project I cloned and it works fine. I also found that when I started the project over again with axes that did NOT include _ that the new firmware somewhat flashed ok to the Wio. terminal.
Question #1 is there a way for me to transfer all of the collected data from my old project to the new project when the 4 axes are written differently? For Example Nitrogen_Dioxide vs. Nitrogen Dioxide
Question #2 is there a way to rename my samples without affecting data collection? When I tried to change the sample/label name edge impulse treated it like a new sample
Thank you in advance
Question #1 is there a way for me to transfer all of the collected data from my old project to the new project when the 4 axes are written differently?
You can easily export your data from your old project to restore it in a new project.
The name of the samples and the labels will be the same but you can change it afterward:
Question #2 is there a way to rename my samples without affecting data collection?
To change the labels of your samples do the following for each class that you want to rename.
What do you mean by “without affecting data collection”?
Dear Mr. Louis,
Thank you so much for your response. I very much appreciate your help.
I apologize ill try to be more specific. In regards to question #1. I believe I can import the data, but my concern is I assigned 4 axes sensors for smell detection, and if I export from the old project will it import the old axes on my new project, or will the export use the axes of the current project assigned?
#2. "What do you mean by “without affecting data collection.”
I attempted to change the names of one of the samples, and instead of simply just changing the name in data acquisition it treated it like a new sample., between training and testing data.
I hope that makes sense. Again thank you for helping me, I am a student, and I’m just beginning to learn and understand this.
Oh ok, I see what you mean!
No at the moment, there is no easy way to rename the axis.
What I would do is, export all the data which will be files in a json format, write a small script to find and replace
Nitrogen Dioxide and do that for every labels.
Here is a small script I wrote for you, just place it in your parent of your export (where there is the
and execute the script doing:
python3 replace_axis.py --input training --old_axis accX --new_axis acc_X
def find_and_place(input, old_axis, new_axis):
for filename in glob.glob(os.path.join(input, '*.json')):
with open(filename) as file:
data = json.load(file)
for item in data['payload']['sensors']:
with open(filename, 'w') as file:
print("Find a replace labels")
a = argparse.ArgumentParser()
a.add_argument("--input", help="input folder")
a.add_argument("--old_axis", help="old axis name")
a.add_argument("--new_axis", help="new axis name")
args = a.parse_args()
find_and_place(args.input, args.old_axis, args.new_axis)
#exemple: python3 replace_axis.py --input training --old_axis accX --new_axis acc_X
I hope that helps,
Thank you so much for all of your help. I really appreciate this.
Apologies for my delay! Regardless, I’m recording my opinion on question/answer #1.
In a similar situation here, the import process turned out to be very easy: just select the exported parent folder and it easily recognizes the division and labels.
However, I would like to suggest an interesting feature: sharing data between projects.
Excellent platform! Congratulations!