TULLUS/Qwen1.5-MoE-A2.7B-GGUF

HF-> F16.gguf converted with a fresh kaggle cookbook :]

• I am moving the other quants into this repo as well..

---

Qwen1.5-MoE-A2.7B

Introduction

Qwen1.5-MoE is a transformer-based MoE decoder-only language model pretrained on a large amount of data.

For more details, please refer to our blog post and GitHub repo.

Model Details

Qwen1.5-MoE employs Mixture of Experts (MoE) architecture, where the models are upcycled from dense language models. For instance, Qwen1.5-MoE-A2.7B is upcycled from Qwen-1.8B. It has 14.3B parameters in total and 2.7B activated parameters during runtime, while achieving comparable performance to Qwen1.5-7B, it only requires 25% of the training resources. We also observed that the inference speed is 1.74 times that of Qwen1.5-7B.

Requirements

The code of Qwen1.5-MoE has been in the latest Hugging face transformers and we advise you to build from source with command pip install git+https://github.com/huggingface/transformers, or you might encounter the following error:

KeyError: 'qwen2_moe'.

Usage

markdown 📊 Quantization Comparison

Quantization	File Size	Recommendation
F16	~28.6 GB	Best precision (for high-end GPUs)
Q5_K_M	soon	Best balance of logic and size
Q4_K_M	~9.5 GB	Recommended for most users
IQ4_XS	soon	Best for low-RAM devices

---

• Adding The FNG's

TULLUS/Qwen1.5-MoE-A2.7B-Q4_K_M-GGUF

This model was converted to GGUF format from TULLUS/Qwen1.5-MoE-A2.7B using llama.cpp via the ggml.ai's GGUF-my-repo space. Refer to the original model card for more details on the model.

Use with llama.cpp

Install llama.cpp through brew (works on Mac and Linux)

brew install llama.cpp

Invoke the llama.cpp server or the CLI.

CLI:

llama-cli --hf-repo TULLUS/Qwen1.5-MoE-A2.7B-Q4_K_M-GGUF --hf-file qwen1.5-moe-a2.7b-q4_k_m.gguf -p "The meaning to life and the universe is"

Server: