409,900,256 trainable world parameters. Twenty selected blocks from the pinned MiniWorld 0.5B DROID checkpoint, matched frozen Wan2.2 codec, 512×288 at 15 Hz, full three-second history, 30 Hz proxy angle/boost controls. A bounded frozen-teacher recovery phase precedes Slither adaptation and gated low-noise history self-resampling. No EMA. Training is action-conditioned; policy fitting is separate. See upstream model and data licenses; this repository does not relicense them. Resumable latest and best rollout-scored checkpoints: https://huggingface.co/buckets/mundoamundo/slither-miniworld-v5-fast-20261002. Read latest.json or best.json for verified file pointers, sizes, and SHA-256. Two…
Independent publisher
Seth Lupo
mundoamundo
Models in Library3
Datasets in Library1
Models on Hugging Face—
Followers—
Compare the published evidence
This profile connects the 4 model and dataset records shown below to their upstream cards and sources. The library counts describe SAVRN's coverage, rather than this publisher's entire output.
Published tasks: Video classification, Video to video.
Reported licenses: apache-2.0. Open a record to review its specific license and conditions.
Compare model specifications, source dates, and hardware estimates on the individual records. Publisher identity verification is separate from a model's quality, deployment readiness, or suitability for your work.
Models
The six-H200 full-corpus run is in progress. Its rolling latest checkpoint is in a mutable Storage Bucket, with the verified step recorded in latest.json. The final weights will be published in this model repository. Live training and held-out validation metrics are updated throughout the run. The model initializes from OpenWAM-α at revision 52df4e66c82c5c8b480adcc8d01f4db7415dfb56. Its Slither adapter replaces the 80-dimensional robot action head with three outputs: cosine and sine of estimated mouse angle, and binary boost. It removes robot proprioception and uses fixed empty-prompt conditioning. Each example uses 33 frames at 30 Hz and 32 transition actions; the nine video frames are…
1,427,969,472 trainable parameters, initialized from pinned Wan2.1-T2V-1.3B. Frozen native video codec, 512×288 at 15 Hz, three seconds observed history, 30 Hz proxy angle/boost controls. Joint stochastic future blocks with native sigma flow matching. Policy fitting is a later separate stage. Resumable latest and best checkpoints: https://huggingface.co/buckets/mundoamundo/slither-wan-v3-20261001. Read latest.json or best.json for verified file pointers, sizes, and SHA-256. Two rotating slots per kind; no repeated weight commits to this Git repository. Checkpoints include model, optimizer, data cursor, and per-rank RNG. Dataset: mundoamundo/slither-wam-video-actions at…
Datasets
This public repository is being built for video and action-conditioned world-model research. It is currently a staging dataset, not a released training set. Prepared gameplay videos, source provenance, and candidate actions will be added incrementally. Each source metadata file records its YouTube URL, creator, reuse basis, edit recipe, checksums, and prepared files. A prepared video is 30 FPS, silent, and one continuous reviewed interval. Actions are separate transition-aligned estimates [cos(mouse angle), sin(mouse angle), boost]; they are not ground truth until reviewed. Reports mark them needsreview. Source-specific redistribution terms are recorded in metadata/.json. No blanket license…