- Rating
- 2
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 1
Read review
Paper summary
This paper studies long horizon consistency in autoregressive video diffusion and argues that existing methods only control which context is retained or indexed in the bounded KV cache, without teaching the model to recover information from evicted frames. The paper proposes Eviction Forcing, which evicts a prefix of frames from a frozen teacher’s rollout while keeping the original RoPE indices of the remaining frames, then trains a LoRA adapter to reconstruct the evicted frames using only the later frames as context, with a periodic preservation loss against the teacher. Built on Self Forcing and combined with Infinity-RoPE at inference, Eviction Forcing achieved competitive overall VBench scores.
Critical strengths
- The method is lightweight. It trains a rank-64 LoRA on a single A100 for 24 hours. Once merged, it adds almost no inference overhead, maintaining about 17 FPS on a single H100.
- Demonstrated on diverse backbones. The proposed method, without backbone-specific tuning, consistently improves consistency and overall scores across three backbones (Self Forcing, CausVid, and SkyReels-V2-DF-1.3B).
Critical weaknesses
- Unclear why past reconstruction improves forward generation. The adapter is trained only to reconstruct evicted frames, while at inference it simply generates forward, and the paper provides no analysis linking the two.
- Gains over Infinity-RoPE are marginal and not disentangled. The method uses Infinity-RoPE at inference, yet its overall score matches Infinity-RoPE at 60s and 120s. The baselines in Table 4 also exclude Infinity-RoPE, so it is unclear how much of the improvement comes from the proposed training.
- Generated videos appear repetitive. Many long rollouts in the supplementary seem to loop or revert to earlier frames, possibly as a side effect of training the model to reproduce past frames.
- Contribution is overstated. The contribution list claims that the method outperforms baselines on overall metrics, but it is best only at 240s.
Questions for authors
- Why does past reconstruction improve forward generation? Could the authors provide supporting analysis or a theoretical explanation?
- Could the authors report results without Infinity-RoPE, and re-report Table 4 with Base + Infinity-RoPE as the baseline? This is needed to show that the two components contribute independently.
- Could the authors report the proportion of generated videos that loop or revert to earlier frames? This would clarify whether the reconstruction training causes the model to generate mostly repetitive videos.