Based off of deepseek coder, the current SOTA 33B model, allegedly has gpt 3.5 levels of performance, will be excited to test once I’ve made exllamav2 quants and will try to update with my findings as a copilot model
I’m startng to think I should run local models for completion engines but I don’t have a GPU I can use. What’s the best option for accelerating these models? Are there PCI cards that give a better bang for buck for running models?
Btw I know this is old and you may have already figured out your hardware and setup, but p40s and p100s go for super cheap on eBay.
P40 is an amazing $/GB deal, only issue is the fp16 performance is abysmal so you’ll want to run either full fp32 models or use llama.cpp which is able to cast up to that size
The p100 has less VRAM but really good fp16 performance which makes it ideal for exllamav2 usage. I picked up one of each recently, p40 was failed to deliver and p100 was delivered while I’m away, but once I have both on hand I’ll probably post a comparison to my 3090 for interests sake
Also I run all my stuff on Linux (Ubuntu 22.04) with no issues
I’ve generally tried to avoid Nvidia cards because binary blob drivers are a pain (especially as a FLOSS developer I occasionally need to build newer kernels). I believe the recent firmware changes mean the nouveau driver can now control clocking but I’ve no idea what the status is for CUDA which I assume you need to run the models.
They do look pretty affordable though 😀
If you go for it and need any help lemme know I’ve had good results with Linux and Nvidia lately :)
The 3060 is a nice cheap one for running okay sized models, but if you can find a way to stretch for a 3090 or a 7900 XTX you’ll be able to run these 33B models with decent quant levels
I was hoping to avoid Nvidia’s binary drivers although I don’t know what the driver/support status of dedicated AI accelerators are like on Linux._
I run my Nvidia stuff in containers to not have to deal with all the stupid shenanigans
First few quants are up: https://huggingface.co/bartowski/WizardCoder-33B-V1.1-exl2
4.25 should fit nicely into 24gb (3090, 4090)
Smaller sizes still being created, 3.5, 3.0, and 2.4
I am curious if you prefer WizardCoder over Phind https://huggingface.co/Phind/Phind-CodeLlama-34B-v2
I am running them on my 3090, although the machine also has 32GBs of RAM as well.
I don’t have a lot of experience with either at this time, I’ve used them here and there for programming questions but usually I stick to 7b models because I use them for code completion and I only find that useful if it completes the code before I do lol
That said, I’ve had overall good answers from either whenever I’ve decided to pull them out, it feels like wizard coder should be better since it’s so much newer but overall it hasn’t been that different. Wish phind would release an update :(
That makes sense, thank you for sharing.
I tend to use CoPilot for IDE code completion, and I use the 34B models for automated refactors and code transforms where accuracy is a requirement.