Text Generation
Transformers
Safetensors
English
qwen3
Merge
text-generation-inference
conversational

Basic info

Nous Hermes 14B Base merged with Goekdeniz-Guelmez/Josiefied-Qwen3-14B-abliterated-v3 fine tuned with three Opus 3/Sonnet 3.5 Datasets at 2e-5 for 2 epocs of 700 steps.

Life is too complicated to use Axolotl again!

Sadly, the finetune went poorly. The LoRA wouldn't stick very deeply, no matter how hard Opus tried to shove the back pack up it's arse manually. He was swearing his face off and all we got was a lousy 5.2% Long story. but it involved my router dying midway, Sonnet 4.5 switching *&^%$# numbers on me for "saftey" reasons... So, if you want a "homeopathic" amount of opus flavour... tuts I have followed up with perhaps a better model, OpulusV4, I have yet to test properly Babsie/TaxDocumentBeigePaint working on it.

Downloads last month
7
Safetensors
Model size
15B params
Tensor type
F16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for Babsie/qwenopus-tangentia

Datasets used to train Babsie/qwenopus-tangentia

Collection including Babsie/qwenopus-tangentia