VideoWorldmodel commited on
Commit
c88a9c6
·
verified ·
1 Parent(s): a7b22c5

Add frozen HY-WorldPlay R50 evaluation results

Browse files

Promote the common-resolution frozen R50 SC comparison, add HY-WorldPlay R50 SC/GT and Endpoint FID/KID reports, and update the reviewer-response artifact.

README.md CHANGED
@@ -32,16 +32,16 @@ Reference videos and trajectory metadata are hosted separately in [VideoWorldmod
32
 
33
  ## Main results
34
 
35
- ### Common camera-trajectory SC (frozen R20)
36
 
37
- minWM and HY-WorldPlay receive the same frozen graph IDs and camera trajectories through their official pose-control pathways. Each model completed 80/80 rollouts with zero evaluation failures.
38
 
39
  | Model | Inverse LPIPS / PSNR | Loop LPIPS / PSNR | Equivalence LPIPS / PSNR |
40
  | --- | ---: | ---: | ---: |
41
- | HY-WorldPlay 1.5 AR Distill | **0.4038 / 16.12** | **0.4562 / 15.48** | **0.3542 / 17.37** |
42
- | minWM | 0.7229 / 10.46 | 0.7442 / 11.08 | 0.4227 / 14.05 |
43
 
44
- R20 contains 20 graphs per relation: 60 graph units and 80 videos because every Equivalence graph has A/B branches. The code archive includes graph-bootstrap intervals and the corresponding GT-anchor summaries.
45
 
46
  ### External-model Inverse SC versus revisit horizon
47
 
 
32
 
33
  ## Main results
34
 
35
+ ### Common camera-trajectory SC (frozen R50)
36
 
37
+ minWM and HY-WorldPlay receive the same frozen graph IDs and camera trajectories through their official pose-control pathways. Each model completed 200/200 rollouts with zero evaluation failures, using the same 512x288 evaluation preprocessing.
38
 
39
  | Model | Inverse LPIPS / PSNR | Loop LPIPS / PSNR | Equivalence LPIPS / PSNR |
40
  | --- | ---: | ---: | ---: |
41
+ | HY-WorldPlay 1.5 AR Distill | **0.4597 / 15.07** | **0.4941 / 15.50** | **0.3723 / 16.94** |
42
+ | minWM | 0.7118 / 10.46 | 0.7442 / 10.85 | 0.4437 / 14.05 |
43
 
44
+ R50 contains 50 graphs per relation: 150 graph units and 200 videos because every Equivalence graph has A/B branches. The code archive includes graph-bootstrap intervals, the corresponding GT-anchor summaries, and frozen R50 Endpoint FID/KID reports. Endpoint KID uses an unbiased finite-sample estimator, so valid negative estimates are retained in the detailed artifact rather than interpreted as distances below zero.
45
 
46
  ### External-model Inverse SC versus revisit horizon
47
 
Reasoning-Structured-Videos-Rebuttal-main.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:93a0d123b757b8689e089d4268bcbebcedf7e073cae6f2d6609c645600414dd4
3
- size 371752
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:00798e8d424c18e7e39c94680b53f2851db106473c277404ca25ca981c287b90
3
+ size 379758