Mobius Labs GmbH

company

https://www.mobiuslabs.com/

AI & ML interests

Computer Vision, LLMs, Multimodal Models, Model Compression

updated a model 9 months ago

mobiuslabsgmbh/CLIP-ViT-H-14-laion2B-2bit_g16_s128-HQQ

Image Classification • Updated Aug 22, 2025 • 37 • 5

updated a collection 10 months ago

Aana Episodic Finetunes

4 items • Updated Aug 2, 2025

in dropbox-dash/Llama-3.1-8B-Instruct_gemlite-ao_a16w4_gs_128_pack_32bit 10 months ago

🙋🏻‍♂️ gemlite hqq quant version "unlearns" tool calling (?)

#2 opened 10 months ago by

updated a model 10 months ago

mobiuslabsgmbh/Qwen2.5-VL-7B-Instruct-leftpad

Updated Jul 22, 2025

published a model 10 months ago

mobiuslabsgmbh/Qwen2.5-VL-7B-Instruct-leftpad

Updated Jul 22, 2025

updated a collection 11 months ago

FP4

MXFP4/NVFP4 models • 2 items • Updated Jun 27, 2025

published 2 models 11 months ago

dropbox-dash/Llama-3.1-8B-Instruct_nvfp4_weights_calib_demo

Text Generation • Updated Jun 26, 2025 • 6 • 1

dropbox-dash/Llama-3.1-8B-Instruct_mxfp4_weights_calib_demo

Text Generation • Updated Jun 26, 2025 • 6 • 1

updated a model 11 months ago

dropbox-dash/Llama-3.1-8B-Instruct_mxfp4_weights_calib_demo

Text Generation • Updated Jun 26, 2025 • 6 • 1

posted an update almost 2 years ago

Post

2143

Releasing HQQ Llama-3.1-70b 4-bit quantized version! Check it out at https://huggingface.co/mobiuslabsgmbh/Llama-3.1-70b-instruct_4bitgs64_hqq.

Achieves 99% of the base model performance across various benchmarks! Details in the model card.

posted an update almost 2 years ago

Post

1811

Excited to announce the release of our high-quality Llama-3.1 8B 4-bit HQQ/calibrated quantized model! Achieving an impressive 99.3% relative performance to FP16, it also delivers the fastest inference speed for transformers.

https://huggingface.co/mobiuslabsgmbh/Llama-3.1-8b-instruct_4bitgs64_hqq_calib

1 reply

·

Jilt

updated a dataset about 2 years ago

mobiuslabsgmbh/youtube-commons-asr-eval

Viewer • Updated May 2, 2024 • 94 • 12 • 5

updated a Space about 2 years ago

README

Multimodal AI for the world's scale

updated 2 models about 2 years ago

mobiuslabsgmbh/Llama-2-7b-chat-hf_2bitgs8_hqq

Text Generation • Updated Feb 5, 2025 • 11 • 36

mobiuslabsgmbh/aanaphi2-v0.1

Text Generation • 3B • Updated Mar 4, 2024 • 68 • • 28

updated a model over 2 years ago

mobiuslabsgmbh/Mixtral-8x7B-Instruct-v0.1-hf-attn-4bit-moe-3bit-metaoffload-HQQ

Text Generation • Updated Feb 5, 2025 • 11 • 13