Endy-Qwen3.6-CyberSec-35B-A3B (abliterated, vision) โ€” fp16 weights

QLoRA fine-tune of huihui-ai/Huihui-Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated, specialized for coding, IT and cybersecurity, kept uncensored. Full-precision (fp16) merged weights.

  • Architecture: qwen3_5_moe โ€” MoE (~35B total, ~3B active) + linear-attention (DeltaNet) + native MTP + vision (multimodal). Merged directly so the MTP head + vision tower are intact.
  • GGUF quantizations (Q2_Kโ€ฆQ8_0 + mmproj for llama.cpp): endystrike/Endy-Qwen3.6-CyberSec-35B-A3B-GGUF
  • Use these fp16 weights to run with transformers / vLLM (incl. image input) or to produce your own quantizations.

Training

QLoRA (Unsloth, 4-bit NF4, r32 ฮฑ64 on q/k/v/o_proj), 2 epochs, train_on_responses_only, on 90,470 coding+cybersecurity chat examples (+914 val). Checkpoint step 2250 selected by validation loss. LoRA merged directly into the fp16 base (preserving MTP + vision), keeping the model multimodal.

Datasets (examples used, licenses)

Dataset Examples Domain License Teacher
AlicanKiraz0/Cybersecurity-Dataset-Fenrir-v2.1 39,286 cybersecurity Apache-2.0 โ€”
Trendyol/Trendyol-Cybersecurity-Instruction-Tuning-Dataset 22,677 cybersecurity Apache-2.0 โ€”
WithinUsAI/fable_5_distillation_merged_cleaned_25k 12,464 coding Apache-2.0 Claude Fable 5
Jackrong/DeepSeek-V4-Distill-8000x 6,292 coding MIT DeepSeek-V4
lordx64/agentic-distill-fable-5-sft 4,593 agentic coding AGPL-3.0 Claude Fable 5
CyberNative/Code_Vulnerability_Security_DPO 4,111 secure coding Apache-2.0 DeepSeek-Coder-33B
beyoru/Deepseek-v4-pro-max-distill-1500x 946 coding Apache-2.0 DeepSeek-V4
WithinUsAI/claude_mythos_distilled (stripped) 101 reasoning Apache-2.0 declared synthetic
Total 90,470

License

AGPL-3.0 (one training dataset, lordx64/agentic-distill-fable-5-sft, is AGPL-3.0 copyleft โ†’ the derived model inherits it).

Disclaimers

  • Parts of the base lineage were distilled from proprietary models (Claude Opus 4.7 / Fable 5, DeepSeek V4) by third-party authors; their usage policies may restrict training competing models on their outputs. Disclosed, not waived.
  • Uncensored/abliterated โ€” outputs are unfiltered. Intended for authorized security research, pentesting, secure-coding and education. You are responsible for lawful use.
  • Not affiliated with or endorsed by Qwen, Anthropic, DeepSeek, or the dataset authors.
Downloads last month
51
Safetensors
Model size
36B params
Tensor type
BF16
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for endystrike/Endy-Qwen3.6-CyberSec-35B-A3B