Cygnis Alpha 2 i1

Imatrix Optimized v0.3 Reasoning

Enhanced Importance Matrix (imatrix) quantizations for Cygnis Alpha 2.
Higher intelligence floor for low-bit inference, preserving complex CoT patterns.

Open in Ollama


        ollama run Simonc-44/Cygnis-Alpha-2-8B-v0.3-i1-GGUF

Available Imatrix Quants

Quantization	Size	Intelligence Level	Download
Q6_K (i1)	6.7 GB	REFERENCE GRADE	Link
Q4_K_M (i1)	5.0 GB	RECOMMENDED	Link
IQ4_XS (i1)	4.5 GB	HIGH LOGIC	Link
IQ3_M (i1)	3.9 GB	EFFICIENT	Link
IQ2_M (i1)	3.0 GB	EXPERIMENTAL	Link

Why Imatrix?

Standard quantization treats all weights equally. Imatrix (Importance Matrix) uses a calibration dataset to identify which neurons are vital for the model's logic. By protecting these neurons, we achieve 3-bit or 4-bit quants that often match the performance of standard 6-bit files.

Perplexity vs Bitrate (Efficiency)

Visual representation of why Imatrix (i1) quants outperform standard static quants at lower bitrates.

Ollama Config

FROM ./Cygnis-Alpha-2-7B-v0.3.i1-Q4_K_M.gguf

TEMPLATE """<|im_start|>system
{{ .System }}<|im_end|>
<|im_start|>user
{{ .Prompt }}<|im_end|>
<|im_start|>assistant
<|im_thought|>
"""
PARAMETER stop "<|im_end|>"

Quantized with ❤️ by Simonc-44 | Powered by Llama.cpp Imatrix