πŸ“‹ Model Description


license: apache-2.0 base_model:
  • theprint/theprint-10B-MoE-A3B-0126
tags:
  • moe

!Theprint MoE 10B 0126

theprint-10B-MoE-A3B (GGUF)

A Mixture of Experts model built on Llama 3.2 3B, combining four specialized fine-tunes with a general-purpose model.

Architecture

  • Base model: theprint/GeneralChat-Llama3.2-3B
  • Gate mode: Hidden
  • Dtype: bfloat16
  • Experts: 4

Experts

ExpertSpecialization
LLM-Data-Science-Llama3.2-3BMachine learning, neural networks, fine-tuning, pre-training
CreativeWriter-Llama3.2-3BFiction writing, story structure, scene development, plot analysis
Llama-3.2-3B-VanRossumPython programming, debugging, algorithm implementation
CogBeTh-Llama3.2-3BMental health support, anxiety, stress management, self-care

How It Works

The model uses a hidden gate mechanism to route inputs to the most relevant expert(s) based on the content of the prompt. Each expert was fine-tuned for its domain before being merged into this MoE architecture using mergekit.

Usage

Compatible with any Llama 3.2 inference setup. No special configuration required β€” the routing happens automatically.

πŸ“‚ GGUF File List

πŸ“ Filename πŸ“¦ Size ⚑ Download
theprint-10b-moe-a3b-0126-f16.gguf
LFS FP16
17.81 GB Download
theprint-10b-moe-a3b-0126-iq4_xs.gguf
LFS Q4
4.89 GB Download
theprint-10b-moe-a3b-0126-q2_k.gguf
LFS Q2
3.41 GB Download
theprint-10b-moe-a3b-0126-q3_k_l.gguf
LFS Q3
4.74 GB Download
theprint-10b-moe-a3b-0126-q3_k_m.gguf
LFS Q3
4.38 GB Download
theprint-10b-moe-a3b-0126-q3_k_s.gguf
LFS Q3
3.98 GB Download
theprint-10b-moe-a3b-0126-q4_0.gguf
Recommended LFS Q4
5.11 GB Download
theprint-10b-moe-a3b-0126-q4_1.gguf
LFS Q4
5.64 GB Download
theprint-10b-moe-a3b-0126-q4_k_m.gguf
LFS Q4
5.46 GB Download
theprint-10b-moe-a3b-0126-q5_k_m.gguf
LFS Q5
6.36 GB Download
theprint-10b-moe-a3b-0126-q6_k.gguf
LFS Q6
7.31 GB Download
theprint-10b-moe-a3b-0126-q8_0.gguf
LFS Q8
9.46 GB Download