"Hortus: The Growers Toolbox App"! Released! care to test?

Way to go buddy keep up the good work you’ll get where you want to be​:grin::+1::+1:

6 Likes

Happy anniversary @kaptain3d

I hope you have a wonderful weekend!

5 Likes

Thank you Sir @Growmoredank :grin:
Have a great weekend yourself my friend :victory_hand:

5 Likes

Great success! @Krista

I trained the Ai model for Ai Plant Sex Analyzer…updated the ai app for it…

I scanned images from my PC screen…should be better in person.

Also fixed the Google flagged 4kb page size error they had flagged that they deprecated and upped to 16kb…so no more Google complaints about this again.

Even though it gets great results now…once the dataset gets 1000s of images properly labeled per class…it will be amazing. It will take years to get that much good data and process it…so I’ll do my best to update these as I have time.

5 Likes

Next up is the Ai Trichomes Analyzer!

:crossed_fingers::crossed_fingers:t2::crossed_fingers:t3::crossed_fingers:t4::crossed_fingers:t5::crossed_fingers:t6:

This one will take more time.

I am going in order of ease for me…the mof and trichome already had a lot of work from old versions…the pest and health analyzers are much more everything and will each take the most time…also I feel those last two have a lot more riding on their accuracy.

5 Likes

@kaptain3d

Are there already threads that you would label as the go to plant pest threads and plant health threads on here?

Or should I create two that would mutually benefit the users by creating a place to get these diagnosed and me for the juicy images…for my datasets?

6 Likes

Went to a memorial today…an installation of a memorial. Two Union soldier relatives that were in the Civil War were found buried in a local cemetery were recognized with a gun salute ceremony and a new memorial for them among others found in the cemetery…total of 7 men from the war.

6 Likes

Not as far as I can remember :thinking:

6 Likes

Thank you kind sir!

6 Likes

I’ll tag you when I get tagged in for diagnosis

6 Likes

That’s awesome! :clap:

I hope you’re not pushing yourself too much tho! I just asked again because I thought that would specifically pull more interest and was something I would be interested in using but please don’t overdo it for me! :sweat_smile::face_holding_back_tears:

Remember to take it easy - Rome was not built in a day :wink::wink::hugs:

Some of these other guys don’t know how to relax either tho :rofl::rofl: whereas if I push myself I’d end up in a straight jacket :face_with_peeking_eye::see_no_evil_monkey:

5 Likes

one of the reasons i have not asked for help with labeling images…the above image is a great example. this image is of clear trichomes, as the three red boxed trichomes shop. however, if someone mistakenly circled all of the trichomes and labeled them clear and some cloudy, they would be wrong in this case…

why?.. you might ask…

When taking pictures of trichomes you need light and a good, magnified image Because of the light… doing so can blur trichomes and make them appear as cloudy, but it is an artificial cloudy not the cloudy we are all looking for that represents a mature ripened cloudy trichome at its highest THC level.

these tiny differences can make or break a dataset labeling…

also notice the first red box is tightly around the trichome head, the second not so much…this loose boxing also lets the background bleed through into the training and valuation steps contaminating it with “background noise”. due to this and the sheer number of images that must be gone through and the quantity of labels that could be on each image…I would possibly have to go through and recheck and correct labels.

so, I have decided after making these mistakes myself and having to go back and correct them…that adding more people and having to train and scrutinize their work is not in my nature…correction, it is totally in my nature…however, I like the relationship I have with you all, and I cannot pay anyone, so being a boss in this situation isn’t in my best interests. friends are the world to me after my family.

4 Likes

Going through the classes I have for the trichome Ai…

Clear

Cloudy

Amber

Of those the Amber category is still very low on images vs the other two categories…I have over 175 images of each of the others…but only about 25 amber images.

I have started a thread:

5 Likes

Well…bad things happened overnight.

I was working on the Ai yesterday and left the PC running overnight.

Woke this morning and it cannot connect to the Internet and is acting funky… ultimately found my network cards and wifi cards have died…possibly the motherboard is as well… however the PC starts…

This sucks!

6 Likes

PC starts so you could run diagnostics with a patch cable and laptop. Do you have another available PCI slot to test the board in? Maybe that slot failed.

6 Likes

@Growmoredank did you ever get that par meter you was looking at i have one similar to it ? And was you gona make it connect to the Hortus app just wondering :grin::+1::+1:

6 Likes

I figured it out…I pulled the wifi/Bluetooth card from inside the PC…it was the issue…I had only reseated it, but failed to remove it completely earlier. Once it was removed, and windows 11 restarted…the onboard 10/100/1000 network controller worked…I found a very long cat6e cable laying around…so I am all good now…I ordered a new card for when I am back in the basement.

6 Likes

@Crazy81 which one do you have? Got a pic and model/manufacturer?

I haven’t bought one yet…but hope to soon.

With the information you provide I may be able to find information.

The plan was to add it to the DLI calculator like the vpd calculator does but for light.

6 Likes

@Crazy81

I plan to put the “+” where the red dots are to set it for each sensor if there are many.

6 Likes


5 Likes