Indexnusrefather/Super-Slop-Machina-Micro-Roleplay-Instruct-750M-v1-0812

Super-Slop-Machina-Micro-Roleplay-Instruct-750M-v1-0812: Microwave Powered Local Roleplay. ---

"No limits, no sanity, no hardware constraints. Local SLM that was finetuned specifically for the purpose of roleplay, now not only on your toaster, but also your microwave." -

Improvements:

• —Better at roleplay than the base model
• —Writes more consistently
• —Somewhat better at keeping track of the plot
• —Handles long term conversations slightly better now

Quants:

Are lying in the repo, I can't say anything for sure about them but wouldn't recommend going any lower than Q6_K, since its my first experiment with under 1b SLM.

Note:

This model is an experiment, methodology I used with it can't make the model physically larger, meaning that logic is capped at 752M parameters either way.

This is also a text only version without vision encoder, just because I found it more convenient.

Also, used my own dataset to finetune it, I hope my data filtering helped a little here. LFM 2.5 350M and LFM 2.5 230M finetunes coming soon, I had to be more aggressive with them so it promises to become very interesting, of course, I fully removed thinking from both.

Big announcement:

I experimented with Qwen 3.5 9b and I think I am somewhat succeeding, I will soon release the finetuned weights, result that I was able to achieve with it was VERY interesting...