Instructions to use VextLabsinc/gem-ruby with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use VextLabsinc/gem-ruby with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("VextLabsinc/juwel-beryl") model = PeftModel.from_pretrained(base_model, "VextLabsinc/gem-ruby") - Notebooks
- Google Colab
- Kaggle
| NOTICE — JUWEL GEM open-weight adapter (code) | |
| Copyright 2026 Vext Labs, Inc. | |
| This is a LoRA adapter for the JUWEL base line (JUWEL Beryl, 80-layer), | |
| which is derived from Qwen3-VL-32B-Instruct by Alibaba Cloud (Tongyi Qianwen Lab), | |
| released under the Apache License 2.0. Original base model: | |
| https://huggingface.co/Qwen/Qwen3-VL-32B-Instruct | |
| Modifications by Vext Labs, Inc.: refusal-direction abliteration, CIP organic | |
| upscale (additional transformer layers, trained), and this domain LoRA adapter. | |
| Attribution only — the "Qwen"/"Alibaba" names and marks are used solely to satisfy | |
| Apache-2.0 attribution and do NOT imply endorsement by Alibaba Cloud. | |
| Apache-2.0 requires retaining this NOTICE and the LICENSE file in redistributions, | |
| and stating that changes were made (they were — see above). | |