Valen: A multimodal decision model inspired by Jev

Valen-Team released Valen-Preview-0923, a multimodal decision model built on a Qwen3.5-0.8B or 2B backbone that scores supplied candidates and returns decision probabilities instead of generating answer tokens. On a Sokoban puzzle, Valen-Preview-0923 solved the level in 9 decisions with 1.13 seconds of cumulative decision latency, while Qwen3.8-27B-FP8 took 198.05 seconds in thinking mode and failed in no-thinking mode. The checkpoint is available on Hugging Face alongside General 100k training and General 5k evaluation datasets, with decision confidence dropping from 91.6% on a clear image to 19.2% at the strongest Gaussian blur.

A multimodal decision model inspired by Jev https://typesafe.ai/blog/introducing-system-one-models-and-jev — text, images and video in; decision probabilities out. English · 简体中文 https://github.com/Liuziyu77/Valen/blob/main/README zh.md Intro introduction · Demos demos · Model downloads model-downloads · Eval results results · Quick start quick-start Valen 万澜 brings visual perception to System One decision-making. Inspired by Jev https://typesafe.ai/blog/introducing-system-one-models-and-jev , it evaluates text, images and video against task instructions and returns probabilities over supplied candidates, giving software a structured decision interface. A Qwen3.5-0.8B or 2B backbone and a shared decision head score candidates without generating answer tokens. The repository includes the model implementation, data processing, SFT and experimental RLCD training, and inference and evaluation commands, with support for training on your own data. Try Valen-Preview-0923 in the online demo https://huggingface.co/spaces/yuhangzang/Valen-Preview-0923 . Valen-Preview-0923 reads the board image and selects a movement direction at each step. On the same level, Valen-Preview-0923 solves the puzzle in 9 decisions with 1.13 seconds of cumulative decision latency. Qwen3.8-27B-FP8 takes 198.05 seconds in thinking mode and fails to solve it in no-thinking mode. Gaussian blur reveals how Valen-Preview-0923 adapts its decisions and confidence as visual detail decreases. Decision confidence is 91.6% on the clear image and 19.2% at the strongest blur. Four successful Valen-Preview-0923 trajectories run side by side at 1× speed with no playback acceleration. Each game takes 7–10 decisions, averaging 122–128 ms per step, and all four finish within 1.24 seconds. Running Valen requires both the Valen checkpoint and the Qwen3.5-2B base model . | Model | Role | Download | |---|---|---| | Valen-Preview-0923 | Valen checkpoint | 🤗 Hugging Face https://huggingface.co/Valen-Team/Valen-Preview-0923 | Datasets: General 100k training https://huggingface.co/datasets/Valen-Team/Valen-Training-General-100k · General 5k evaluation https://huggingface.co/datasets/Valen-Team/Valen-Eval-General-5k · Sokoban training and evaluation https://huggingface.co/datasets/Valen-Team/Valen-Eval-Game . Lower latency, higher accuracy. Four panels compare General accuracy, General latency, Sokoban accuracy and Sokoban latency, from left to right. Each panel shows Qwen3.5-0.8B, Qwen3.5-2B and a Valen 2B RL checkpoint. General uses 5,000 questions from multiple VQA datasets; Sokoban uses 500 single-step questions from 100 levels. See the technical notes https://github.com/Liuziyu77/Valen/blob/main/docs/technical.md %E5%AE%8C%E6%95%B4%E5%AE%9E%E9%AA%8C for training data, the Model Card and loss curves. Install the dependencies listed in requirements https://github.com/Liuziyu77/Valen/blob/main/docs/technical.md %E7%8E%AF%E5%A2%83%E8%A6%81%E6%B1%82 , then download both the Preview checkpoint and its Qwen3.5-2B base model. Download the Valen-Preview-0923 checkpoint. hf download Valen-Team/Valen-Preview-0923 --local-dir models/Valen-Preview-0923 Download the Qwen3.5-2B base model. hf download Qwen/Qwen3.5-2B --local-dir models/Qwen3.5-2B Train the model with your own configuration. python -m valen.train \ --config configs/train/sft warmup.json Run inference with the downloaded Valen-Preview-0923 checkpoint. python -m valen.inference \ --checkpoint models/Valen-Preview-0923 \ --data data/smoke/train.jsonl \ --output output/sft warmup/predictions.jsonl Check the evaluation pipeline on the same synthetic examples. python -m valen.evaluate \ --checkpoint models/Valen-Preview-0923 \ --data data/smoke/train.jsonl \ --output output/sft warmup/smoke eval data/smoke contains a small set of simple questions for checking that the pipeline runs correctly. A labeled record with an image input Each JSONL line contains one record. The example below uses the evaluation overview figure https://github.com/Liuziyu77/Valen/blob/main/assets/figures/evaluation-results.png from the repository, assuming the file is saved as example.jsonl in the repository root. { "group id": "evaluation-general-2b", "request": { "state": { "messages": { "role": "user", "content": {"type": "text", "text": "Compare the accuracy and the average latency per question of the 2B models on General in the figure."}, {"type": "image url", "image url": {"url": "assets/figures/evaluation-results.png"}} } }, "questions": { "best 2b": { "type": "choice", "instructions": "On General, among the 2B models with average latency below 200 ms per question, which has the highest accuracy?", "criteria": { "qwen": "Qwen3.5-2B", "valen": "Valen-Preview-0923" } } } }, "targets": { "best 2b": {"probabilities": {"qwen": 0.0, "valen": 1.0}} } } Contributions to Valen are welcome. Open an issue https://github.com/Liuziyu77/Valen/issues to report a problem, share a use case or discuss experimental results. Submit a pull request https://github.com/Liuziyu77/Valen/pulls to improve the code or documentation, contribute training data or add evaluation tasks. Scan the QR code below to join the Valen WeChat group, discuss the project and share your experiments. The code is released under Apache 2.0 https://github.com/Liuziyu77/Valen/blob/main/LICENSE . Base models and source datasets retain their respective licenses. Built on Qwen3.5 https://huggingface.co/Qwen/Qwen3.5-2B , with the decision interface inspired by TypeSafe's Jev https://typesafe.ai/blog/introducing-system-one-models-and-jev .