MATLOWAI
/

MiniMax-H3-ORB360-CardSpin

@@ -144,7 +144,8 @@ For the card spin, use `minimax_h3_orb360_cardspin_v2_step50.safetensors` and `p
     1500 was trained on, taken from two of our held-out tests (an orbit from 30 degrees below with no floor, and front
     and back pictures taken from 45 degrees with the orbit at 5 degrees). The numbers in them (camera distance, how
     much of the frame the subject fills) describe those test objects; edit the heights, times and angles for your own
-    use.
   - `cardspin_caption.txt`: the card spin.
 - Resolutions we have seen work: 800 x 800, 1024 x 768 and 1152 x 768 landscapes, 672 x 832 and 832 x 1024
   portraits.
@@ -157,6 +158,9 @@ For the card spin, use `minimax_h3_orb360_cardspin_v2_step50.safetensors` and `p
 Tested on three objects that were never rendered for training (a vintage film camera, a stylised character and a
 stone cat statue), one seed per test, against Blender renders of the true orbit:
 | Test | Result |
 |---|---|
 | Orbit at 20 degrees, 3 views (0/120/240) | Within one frame (2.9 degrees) of the true orbit on every frame |
@@ -206,6 +210,10 @@ reference pictures at 0, 120 and 240 degrees, 750 steps. Stage 2: the same 50 ca
   some subjects or seeds commit to the card less than others. Both CardSpin files soften the image (table above).
 - Step 1500 was trained on grey-backdrop renders of 23 objects. It orbits real photos well in our tests, but was
   evaluated on a handful of images and three held-out objects, one seed each, not a large benchmark.
 - Big jumps between the pictures' height and the orbit's height, and orbits whose height changes during the clip,
   do not work yet.
 - From one picture, the sides the camera has not seen are invented. More pictures (front and back, a turnaround, or

     1500 was trained on, taken from two of our held-out tests (an orbit from 30 degrees below with no floor, and front
     and back pictures taken from 45 degrees with the orbit at 5 degrees). The numbers in them (camera distance, how
     much of the frame the subject fills) describe those test objects; edit the heights, times and angles for your own
+    use. **Height prompts are still being trained:** they work on grey-backdrop renders like the ones step 1500 was
+    trained on, but on real photos the orbit currently stays at the photo's own height (see Limitations).
   - `cardspin_caption.txt`: the card spin.
 - Resolutions we have seen work: 800 x 800, 1024 x 768 and 1152 x 768 landscapes, 672 x 832 and 832 x 1024
   portraits.
 Tested on three objects that were never rendered for training (a vintage film camera, a stylised character and a
 stone cat statue), one seed per test, against Blender renders of the true orbit:
+All of these tests use grey-backdrop Blender renders. Height control has not yet carried over to real photos (see
+Limitations).
 | Test | Result |
 |---|---|
 | Orbit at 20 degrees, 3 views (0/120/240) | Within one frame (2.9 degrees) of the true orbit on every frame |
   some subjects or seeds commit to the card less than others. Both CardSpin files soften the image (table above).
 - Step 1500 was trained on grey-backdrop renders of 23 objects. It orbits real photos well in our tests, but was
   evaluated on a handful of images and three held-out objects, one seed each, not a large benchmark.
+- **Camera-height (angle offset) prompts do not hold on real photos yet.** They work on grey-backdrop renders, but in
+  our test an eye-level real photo prompted for an orbit 35 degrees above or 20 degrees below orbited at the photo's
+  own height both times. A follow-up run training straight-on references, tilted orbits and vertical loops is in
+  progress.
 - Big jumps between the pictures' height and the orbit's height, and orbits whose height changes during the clip,
   do not work yet.
 - From one picture, the sides the camera has not seen are invented. More pictures (front and back, a turnaround, or