i think i know why this is happening: someone has made an assumption about aspect ratios. my phone is tall and narrow, so anything locked to 16:9 will be stretched or have gaps.
hydrofoiling catamarans, but yes
lime
joined 2 years ago
boooo +1
...you okay man?
sounds like you should transition to being an architect
it's melted snowflakes do try to keep up
not tried gemma yet, i've stayed away from google stuff. maybe i'll give it a shot.
yeah one of those framework machines with 128GB shared ram would have been amazing. shame they're sending money to racists.
one of my most recent fun activities came from discovering the "allow editing" button in koboldcpp. since the model is fed the entire conversation so far as its only context, and doesn't save data between iterations, you can basically re-write its memory on the fly. i knew this before but i'd never though to do it until there was an easy ui option for it, and it turned out to be a lot of fun, because when using a "thinking" model like qwen3.5 you can convince it that it's bypassing its own censorship.
basically you give the model a prompt to work off of, pause it in the middle of the thinking process, change previous thoughts to something it's been trained to filter out (like sex or violence or opinions critical of the ccp), and it will start second-guessing itself. sometimes it gets stuck in a loop, sometimes it overcomes the contradiction (at which point you can jump in again and tweak its memory some more) and sometimes it gets tied up in knots trying to prove a negative.
a previous experiment was about feeding stable diffusion images back into itself to see what happens. i was inspired by a talk at 37c3 where they demonstrated model collapse by repeatedly trying to generate the same image as they put in (i think this was how sora worked).
i will never understand that attitude. four hours of exploration, learning and puzzle-solving sounds like the best part of the job. an isolated, well-specified problem that can be completed in a day is like the most fun you can have programming. why swap that for an hour of code review?
i did my first machine learning course more than 10 years ago, so i'm not ashamed to admit that i bought beefier hardware to play around with local models in early 2023. i still like doing that. mostly because i know my gpu is powered entirely off of fossil-free energy and because i decided early on not to spew the output all over the internet unless it was poignant. or funny. not as in "the llm told a good joke", more as in "i compressed this poor thing to fit on a cd and now it can only talk about dolphins".
qwen3.5-12B really screams along on a 7900xtx. like, up to 70-100 tokens a second. perfect for seeing the results of your torture methods quickly.
no, robert deniro is not "in heat"
Note: the original thread title was "how do i fix overextrusion on infill in orcaslicer?". we've since deduced that's not what's happening. i'm leaving the rest of the op as is so you can follow the process.
So i'm doing a test print for a hot wheels track i'm making for a friend's kid on my snapmaker u1, and i'm hearing scraping noises. when i look at the in-progress print, i see this horribly mangled infill. obviously the nozzle is hitting the previous layer, right? so that's overextrusion, i think. too much material. but i let the print run, thinking maybe ironing will save it. but the surface finish is absolutely awful. all of the bumps and ridges of the infill pattern transfer up through the solid layers. not to mention now there's ringing from the nozzle hitting the bumps, so there's even more bumps. bummer.
also yes i fucked up the overhangs by trying to cheap out on supports. at least that one i know how to fix.
so, how do i deal with this? snapmaker ships a specialized version of orcaslicer (it's called snapmaker orca, it's on github) to deal with the u1s four separate print heads, and as far as i can tell there's no setting in there for infill flow? should i just try to slow everything down? i thought it might be vibration-related so i added a 20kg concrete slab and a thick anti-vibration rubber mat to the setup but nothing changed. i also dried the filament out for six hours. the hygrometer in the snapdryer got down to 12% i think.
i'd hate to not be able to print this for the kid, it's such a cool plaything.
Edit: to clarify, the grey filament is snapmaker matte PLA. the spool has an rfid chip in it so i've not changed any settings, the printer just detects it and sends it to orca.
Edit 2: i've done another test using gyroid infill and a lower flow rate, as recommended in the thread, but the surface finish is all bubbly. i cut a part out to check if the infill was the problem but it looks fine, while the surface is fucked. this is after ironing, by the way.
Edit 3: okay, i've now dried the spool out overnight and made a test disk with tweaked parameters, and i'm still seeing bubbles on top. here i increased the ironing flow from 8% to 20% so the surface is a lot smoother, but the bubbles are still visible. also getting some weird blobs on the side? could be related to the ironing. the middle circle is for a multicolor test but the dot was so small that the filament amount came out to 0.00 grams and the printer didn't really know what to do. it just put in a single dot of (the wrong) filament and avoided the area.
the changed parameters are
- flow ratio: 0.95 -> 1
- nozzle temp: 215C -> 220C
- max volumetric speed: 22mm^3^/s -> 15mm^3^/s
- seam position: nearest -> random
- scarf joint: off -> contour and hole
- ironing flow: 8% -> 20%
- infill: gyroid 15% -> TPMS-D 10%
- infill combination: off -> on, 80%
Growing up, portable cassette players were always called "freestyles" here. I never knew it was a marketing thing, or that some other countries also objected to the naming.
this is "original research", which means i dicked around on the internet archive for half an hour. it may be wrong.
view more: next ›
holy fuck the resolution to the cow thing had me in stitches.