Descriptions were fine for both. I do find it interesting though that it interpreted the stitching in the hat as the word 'Jar' though it couldn't identify the right character even when I tried to direct it to the character. su…
I've tested 3 models: gemma4-26B-A4B-it-UD-Q4_K_M gemma4-31B-it-Q4_K_M qwen3.6-35B-A3B-UD-IQ4_XS Asked following question: We developing a Godot 4 3D RPG game. First task would be to make a professional and smooth 3rd person camera control…
Hey all, we’re a mid-sized company (~70 people) and currently planning to bring a lot of our workloads on-prem instead of relying on cloud APIs. The goal for the moment is to run small to mid-sized models in the range of 30B like Qwen3.6 or Gemma4. Us…
Hey, I've recently updated my narrative battle simulator SLOP FIGHTER to run Gemma4 2B. I've also made it compatible with Windows machines and hugely stepped up the quality of responses across the board. I find Gemma4 2B to be incredibly …
So i currently have an rtx 4070 super, and it can easily run models like gemma3 12b and even gpt-oss 20b (although it takes up to a minute to generate a response). I want to get a second gpu so i can run larger models around 20b-30b params. What gpu do…
I'm using the https://github.com/PrismML-Eng/llama.cpp fork for Bonsai, regular llama.cpp for Gemma. Without embedding parameters: Gemma 4 has 2.3B at 4.8 bpw (Q4_K_M) = 1104 MB Bonsai-8B has 6.95B at 1.125 bpw (Q1_0) = 782 MB (-29% smaller) …
This is an unhelpful rant, but it's been getting to me. I don't code. I don't care about python. I don't know and don't care how agents work and what they do. I don't build websites and I couldn't care less about github inte…
mii-llm just released a detailed technical report on the development of the Zagreus and Nesso model families: a set of 0.4B parameter language models trained from scratch with a focus on edge deployment, multilingual capability, and European lang…
https://huggingface.co/mradermacher/Qwen3.5-35B-A3B-Base-GGUF Yes, Qwen 3.6 is out and it's a great model. However, who needs an even more "uncensored but official" model, can try out this one. With a small clever DAN-Sysprompt you …
This spring is really hot since the localLLM giant, both Qwen and Gemma released major models. I'm really excited with those release and happy with their capability. Both are real hero for local LLM, although I have feeling they have different stre…