Von, an Open-Source Jev Alternative

Von, an Open-Source Jev Alternative wfzyx / von Public Notifications You must be signed in to change notification settings Fork 52 Star 698 Branches Tags Open more actions menu Latest commit History 136 Commits 136 Commits Folders and files Name Name Last commit message Last commit date .github/ workflows .github/ workflows assets assets benchmarks benchmarks examples examples hf hf js js scripts scripts src/ von src/ von tests tests training training .gitignore .gitignore .python-version .python-version LICENSE.md LICENSE.md README.md README.md pyproject.toml pyproject.toml uv.lock uv.lock Repository files navigation Von An Open-Source, Non-Autoregressive System One Decision Model. Calibrated discrete, probabilistic, and ordinal inference in sub-25ms.

This Partnership is relevant to the technology intelligence record because it involves GitHub, Intel, NVIDIA, AMD. The source article should remain the factual reference for follow-up coverage.

• wfzyx / von Public Notifications You must be signed in to change notification settings Fork 52 Star 698 Branches Tags Open more actions menu Latest commit History 136 Commits 136 Commits Folders and files Name Name Last commit message Last commit date .github/ workflows .github/ workflows assets assets benchmarks benchmarks examples examples hf hf js js scripts scripts src/ von src/ von tests tests training training .gitignore .gitignore .python-version .python-version LICENSE.md LICENSE.md README.md README.md pyproject.toml pyproject.toml uv.lock uv.lock Repository files navigation Von An Open-Source, Non-Autoregressive System One Decision Model.
• Calibrated discrete, probabilistic, and ordinal inference in sub-25ms.
• Every movement decision — advance, back off, sidestep a fireball — is one forward pass scoring six option descriptions against a text rendering of the depth buffer.
• No policy network, no fine-tuning, no reinforcement learning: the same shipped Von weights that answer routing questions, wired to a game loop.
• 33 kills, 36 dodges, 80 seconds, on a GPU.
• What's new in 1.2: order-invariant option scoring Von 1.1 had a defect shared with most option-packing models: the answer could depend on the order the options were listed in.