Text Generation
PEFT
Safetensors
English
Korean
mistral3
decision-model
typed-decisions
jev
jevbench
calibration
decode-free
lora
Instructions to use StandardThinking/StandardOne-3B-LoRA with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use StandardThinking/StandardOne-3B-LoRA with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("mistralai/Ministral-3-3B-Instruct-2512-BF16") model = PeftModel.from_pretrained(base_model, "StandardThinking/StandardOne-3B-LoRA") - Notebooks
- Google Colab
- Kaggle
Download NOTICE from StandardThinking/StandardOne-3B-LoRA: direct link, hf CLI and curl.
- Browser
- Download file 1.17 kB
-
https://huggingface.co/StandardThinking/StandardOne-3B-LoRA/resolve/main/NOTICE
- Command line
-
hf download hf://StandardThinking/StandardOne-3B-LoRA/NOTICE
-
curl -L -o NOTICE https://huggingface.co/StandardThinking/StandardOne-3B-LoRA/resolve/main/NOTICE
1.17 kB
| Jev adapter | |
| The initial protocol helpers and their unit tests were extracted from the | |
| System One extension in the local sglang-jev experiment, commit | |
| 633f05e (Add prefill-only System One decisions and visual game evaluation). | |
| That repository was based on SGLang upstream commit | |
| 0024efa0de38794ee309ba10ab00ebb891a3d050. | |
| The accompanying Apache License 2.0 is retained from that source tree. | |
| SGLang project: https://github.com/sgl-project/sglang | |
| This repository is an independent API adapter. It does not include the | |
| proprietary Jev model and does not bundle the SGLang inference engine. | |
| Benchmark metric definitions and dataset schema mapping follow Jared Palmer's | |
| Apache-2.0 KEV project, pinned at 4f8110a3f8620cc3a182ae9a708e4398492c4b1a: | |
| https://github.com/jaredpalmer/kev | |
| The benchmark runner is an independent implementation. Downloaded benchmark | |
| data retains upstream provenance and any underlying dataset license terms. | |
| This adapter is trained on top of mistralai/Ministral-3-3B-Instruct-2512-BF16 (revision b6d637bef2393152b3da2b2fde72eecdee30557e), released by Mistral AI under the Apache License 2.0. The base weights are not included in this repository. | |