This is a learning experiment/project.

This LoRa aims to stabilize the generation of the Glock 17 firearm and how it is held.

Version 0.1 is just a base, and if the experiment goes well, there will be V1, V2, and V3 with the goal of making the weapon customizable (color, modifications, condition). Although I’m not entirely sure if this is really possible, part of the reason I wanted to publish this was to keep a “record” and also with the hope of writing an article about it in the future.

If this goes well, I'll probably make versions of other weapons.

V1:

After 3 redone datasets and about 10 training runs, I managed to get a more stable version.

It is recommended to use hires.fix.

Style bleeding has been resolved, at least for the most part.
Overall, the Glock 17 looks better proportioned in wide shots.
The Glock’s geometry is more stable from different angles (except for the two front holes—this version actually gave the best results).

In this version, I tried to separate trigger discipline and finger on trigger. I must say it turned out so-so… trigger discipline works the best, while finger on trigger works only occasionally, so I’ll leave trigger discipline as the default. Whoever reads this can still try using finger on trigger.

I tested some examples from different angles—side view, from above, from below worked correctly, from behind works only somewhat.

There are some complementary tags you can try, though they don’t work perfectly:

• two-handed

• aiming

• aiming at viewer

Made it this far?

If you like the model, let me know! Comments, feedback, suggestions, even insults—everything’s welcome.

I also really enjoy seeing what people create with my models. You’ve got my like guaranteed!

Also, feel free to explore my model library — I’ve got many LoRAs in various styles that you might enjoy.

And if you’d like to support me, you can do so on Ko-fi. Even the smallest contribution means a lot and truly motivates me to keep making models and tools for the community.
.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.