Hugging Face's logo

tatjr13
/
damascus-t2-base-l1

GGUF
imatrix
conversational
Model card Files
xet
 Community

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

damascus-t2-base-l1

GGUF artifact for DAMASCUS test T2 (test plan v2.1): the untouched tpn-004 base exported with the L1 recipe.

  • Source: Magistral-Small-2509-ultra-uncensored-heretic-v2-BF16.gguf (sha256 5d425b36eeff7ce5...), never trained by us.
  • Export: llama.cpp b10020 llama-quantize with the evaluator-shaped imatrix — Q4_K_M 32 with Q8_0 for attn_q/k/v/output, ffn_down, token embeddings and output tensor (the exact L1 recipe; per-tensor conversion log identical to the L1-im export).
  • Template variant T1a: tokenizer.chat_template set to the T1a variant — the shipped 305-char Mistral [INST] loop with the "Answer directly and briefly..." system instruction prepended as its own [INST] ... [/INST] block after bos_token. Template sha256 d338a0af9865562ac73b6f67e82fc80a7edd4e86d384a0c0b10caba784f5ae11.

The single file model.gguf is the complete artifact.

Downloads last month
146
GGUF
Model size
24B params
Architecture
mistral3
Hardware compatibility
Log In to add your hardware

We're not able to determine the quantization variants.

View all variants
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support