ViAble-1.7-9B
ViAble-1.7-9B — это ориентированная на рассуждения тонкая настройка Qwen 3.5 9B (архитектура qwen35): контролируемая тонкая настройка на ~98...
Метка: qwen3_5_text
Qwen3.8-27B-Obliterated-E03
manitcor/Qwen3.8-27B-Obliterated-E03 is a BF16, text-only derivative of Qwen/Qwen3.8-27B. It was produced with the experimental qwen38e03` refusal-direction intervention in...
Qwen3.8-27B-PrismaAQUA-gridbook-13GB-5080-vllm
A 13 GB download: 12,982,426,694 bytes = 12.98 GB = 12.09 GiB. Weights and codebooks are 12,958,885,485 B...
Nyx-RP-9B-Instruct-2608-v1
Nyx-RP-9B-Instruct-2608-v1: Final release version of a roleplay finetune for Qwen 3.5 9B. Way better roleplay than the base...
turkish-pii-guard-0.8b-legacy
Модель Türkçe metindeki kişisel verileri tespit edip yerinde maskeleyen 0.8B parametreli. Girdiyi alır, kişisel veriyi köşeli parantezli etiketle...
Ornith-1.0-9B-Hardware-Expert-fp16
Тонко настроенный эксперт по аппаратной инженерии — низкоуровневый аппаратный код, разработка драйверов, встроенные системы, RTL-дизайн и электронная инженерия....
Ornith-1.0-9B-Uncensored
Since this model is original base (no fine-tune) just with safety vectors steered — it will likely give...
Holo-3.1-9B-Coder
LiveCodeBench evaluated with official codegenmetrics, greedy decoding, 6s timeout. Proof: eval/lcbv2official.json` Quantization was performed with llama.cpp after merging...
Dillionv2-1.3M
Dillionv2 is our second generation model of the Dillion SLM family. It is a significant improvement over v1...
Qwen3.6-27B-Opus-CoT-S1-Hermes-S2-SFT
> Calling for independent benchmarks. I’ve only run internal smoke tests against this model — full eval numbers...