You can deploy this model and start earning money today!

canvit/
pgpolicy-ade20k-c64-t5-floorband-2026-07-07-s6

Name: pgpolicy-ade20k-c64-t5-floorband-2026-07-07-s6
Availability: InStock

canvit-pytorch-rl

safetensors

reinforcement-learning

mit

Views

No views yet

Viewpoint policy-gradient actor — ADE20K, c64, T=5

Glimpse-placement policy for a frozen CanViT, over a discrete grid of candidate viewpoints (2 scales x 16x16 centres): scores per candidate are policy logits trained on-policy by score-function credit (loss -(z * log pi) - entropy_bonus * H, K=1 fractional-CE rewards); deploy = full-scene glimpse at t0, argmax afterwards — the same deploy rule as the Q-regression policies. Method + code: CanViT-PyTorch-RL.

Provenance: run 20260707-160155_pgfloor_s6, git d63026b8589a6fc035f724fd491b3049837b5681, seed 6, step 6000 (deploy ckpt = best mean(t1-t4) validation CE over the run's evals). Backbone canvit/canvitb16-add-vpe-pretrain-g128px-s512px-in21k-dv3b16-2026-02-02; probe resolved for canvas grid 64. Full-val metrics at this step:

glimpse step t	val CE	val mIoU %
t0	0.7649	39.60
t1	0.7155	42.49
t2	0.6893	43.66
t3	0.6770	44.29
t4	0.6679	44.65

python

from canvit_pytorch_rl.policy.net import ViewpointScorer
net = ViewpointScorer.from_pretrained("canvit/pgpolicy-ade20k-c64-t5-floorband-2026-07-07-s6").eval()

Drive it through canvit_pytorch_rl.policy.deploy.ArgmaxPolicy + canvit_eval.run_episode for protocol-exact rollouts; canvit_pytorch_rl.policy.eval is the standalone evaluator.

Seed 6 of the 8-seed pg entropy-floor band (floor 1.0 nat), band objective 0.6865±0.0012 vs qband 0.6853±0.0007; see docs/sessions/2026-07-05-e2e-diff-single-image-probe.md.