I really wanted to make a system that would recognize if it’s seen the same 30-second clip before (since ads are always repeats) by a shared signature that would just play something else or silence for the length of the ad on the client side, especially for live sports streams.
Yeah I don’t wanna exaggerate but it was nearly the worst environment I could think of to just leave bread in. If you had a party or a brunch or something and didn’t want the bread sitting out, sure, but for having in there for a week and a half unmonitored, we’ve been taken for absolute fools. :)
Had one, hated it, forget the bread was in there* and it doesn’t have some sort of magical bread preservation properties, it’s just a spot taking up counter space to hold a plastic bag.
* (I recognize this is my specific problem but it’s on the list of why it doesn’t work for me)
This is a great point, not sure what kind of bias it is, but you’d literally see thousands of people agreeing (the upvotes) and then 10 people circling around in a knife fight. Did we need science to tell us Reddit is full of trolls? Trolls existed on Reddit before LLMs became popular.
smiles at Queen Amidala
They’ve been terrible for a long time. After Anthem, honestly never again. Sorry.
Putendo
Yeah I got mind refurbished also, so someone else took the first hit on driving it off the lot (and waiting for it to be built). I guess they didn’t use it to its full extent though. That didn’t make it “cheap” though.
It’s sort of a niche within a niche and I appreciate your sharing some knowledge with me, thanks!
Hmm maybe in the new year I’ll try and update my process. I’m in the middle of a project though so it’s more about reliability than optimization. Thanks for the info though.
I usually run batches of 16 at 512x768 at most, doing more than that causes bottlenecks, but I feel like I was also able to do that on the 3070ti also. I’ll look into those other tools though when I’m home, thanks for the resources. (HF diffusers? I’m still using A1111)
(ETA: I have written a bunch of unreleased plugins to make A1111 work better for me, like VSCode-like editing for special symbols like (/[, and a bunch of other optimizations. I haven’t released them because they’re not “perfect” yet and I have other projects to be working on, but there’s reasons I haven’t left A1111)
I just run SD1.5 models, my process involves a lot of upscaling since things come out around 512 base size; I don’t really fuck with SDXL because generating at 1024 halves and halves again the number of images I can generate in any pass (and I have a lot of 1.5-based LORA models). I do really like SDXL’s general capabilities but I really rarely dip into that world (I feel like I locked in my process like 1.5 years ago and it works for me, don’t know what you kids are doing with your fancy pony diffusions 😃)
Oh I meant for image generation on a 4080, with LLM work I have the 64gb of the Mac available.
Ok but it’s not done by a bunch of Robin Hoods, they rip off (read: steal and then destroy the economic capacity of) small/independent designers all the time too.
It fails whenever it exceeds the vram capacity, I’ve not been able to get it to spillover to the system.
Oh I didn’t mean “should cost $4000” just “would cost $4000”. I wish that the vram on video cards was modular, there’s so much ewaste generated by these bottlenecks.
Apple price gouges for memory, yes, but a 64gb theoretical 4090 would have cost as much in this market as the whole computer did. If you’re using it to its full capabilities then I think it’s one of the best values on the market. I just run the 20b models because they meet my needs (and in open webui I can combine a couple at that size), as I use the Mac for personal use also.
I’ll look into the Amd Strix though.
I know it’s a downvote earner on Lemmy but my 64gb M1 Max with its unified memory runs these large scale LLMs like a champ. My 4080 (which is ACHING for more VRAM) wishes it could. But when it comes to image generation, the 4080 smokes the Mac. The issue with image generation and VRAM size is you can think of the VRAM like an aperture, and the lesser VRAM closes off how much you can do in a single pass.
I’m in management now but I say go get ‘em please.
Yeah this would still hit 100% of ads that have happened before now. And all linear tv ads on streams. Seriously willing to build it if anyone wants to work on it with me. Pm me