YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

Q4_nocfg_49f_320x192

V8 Diffusion-DRF LoRA probe on Wan 2.2 I2V 14B (transformer_2).

  • Geometry: 49f @ 320x192, K=2, no-CFG-in-grad-window
  • rank16 lr5e-6 5 steps; train-geom held-out 0.8778 -> best 0.8743
  • Targets: to_q, to_k, to_v, to_out.0, ffn.net.0.proj, ffn.net.2
  • Base model: subnet king (hippius-hub snapshot)

Train-geometry held-out LPIPS improvements were NOT reproducible at full 832x480x81 eval geometry (all 4 probes within +-0.0008 of king baseline 0.4310).

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support