LeWAM-Simulation-Checkpoint
Collection
LeWAM v1 checkpoints for simulation tasks: OGBench scene, puzzle (3x3), cube, DexMimicGen drawer, and PushT. • 10 items • Updated
LeWAM DexMimicGen TwoArmDrawerCleanup
Goal-reaching checkpoint in the LeWAM v1 format. Files are stored at the repository root:
lewam_best.pt
lewam_config.json
Download both files into $STABLEWM_HOME/checkpoints/<run_name>/, then evaluate with the v1 code:
python scripts/eval_lewam.py --config-name drawer policy=<run_name>
The checkpoint was trained with lewam-jointflow@2d52888 and converted to the v1 schema.
Every retained tensor is exactly equal to the original after renaming; strict model loading,
attention masks, encoder output, supplied-goal action prediction and dynamics prediction passed.
Only the obsolete, all-zero null_goal tensor was removed; the original goal-dropout setting was zero.
No optimizer state is included.
Original checkpoint and historical performance: source repository. Drawer uses goal offset 50 and budget 100; the other tasks use offset 25 and budget 50.
SHA256 of lewam_best.pt: c91ddb9ec2cafa95e4166fdbec4093c7b4ff67bede8ee36bd12ee2598494f4f6.
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support