Update README.md
Browse files
README.md
CHANGED
|
@@ -6,4 +6,6 @@ base_model:
|
|
| 6 |
- inflatebot/MN-12B-Mag-Mell-R1
|
| 7 |
- Nitral-AI/Captain-Eris_Violet-GRPO-v0.420
|
| 8 |
---
|
| 9 |
-
# Test Q5_K_M Quant: [Here](https://huggingface.co/Nitral-AI/Violet_Magcap-12B-Q5_K_M-GGUF)
|
|
|
|
|
|
|
|
|
| 6 |
- inflatebot/MN-12B-Mag-Mell-R1
|
| 7 |
- Nitral-AI/Captain-Eris_Violet-GRPO-v0.420
|
| 8 |
---
|
| 9 |
+
# Test Q5_K_M Quant: [Here](https://huggingface.co/Nitral-AI/Violet_Magcap-12B-Q5_K_M-GGUF)
|
| 10 |
+
|
| 11 |
+
## "So here’s the lore: I might have taken [mag-mell-12B-r1](https://huggingface.co/inflatebot/MN-12B-Mag-Mell-R1), jacked it up on SFT reasoning data like it was pre-workout for logic bros. Then for 'chaos' slapped it together with [Captain_Eris_Violet-GRPO](Nitral-AI/Captain-Eris_Violet-GRPO-v0.420) like it was some kind of twisted AI Voltron. Did i stop there? ...'no'. Imediately after merging, double-tapped the training with not one, but two iterations of completely fresh reasoning data—stuff neither of its parent models had ever seen before. This thing ate that data like a Reddit mod eats power trips. Now it’s out here reasoning through problems like Bill Nye with KCTS, except it also thinks in memes (probably) and might have a cursed folder of philosophy shitposts."
|