SqrtRyan

Dear Authors of Submission #45242,

Thanks to the efforts of all reviewers, emergency reviewers, ACs, and SACs, preliminary reviews are now available. Below you will find the preliminary reviews for your CVPR 2026 submission, MotionV2V: Editing Motion in a Video (#45242). Authors have the opportunity to submit a rebuttal by January 29, 2026 11:59 PM AoE. Please review the rest of this email, and the Author Guidelines for additional details on the rebuttal process.

We refer to the current reviews as preliminary because review revisions to improve clarity are possible for reviewers until January 23 11:59pm AoE.

Reviews

You can access the reviews for your submission by clicking on the corresponding submission in OpenReview: https://openreview.net/group?id=thecvf.com/CVPR/2026/Conference/Authors

SqrtRyan

  Full Research Report: LTX-2 Model Architecture Comparison

  Executive Summary

  Your key question answered: The IC (In-Context) models and I2V (Image-to-Video) models use fundamentally different conditioning approaches, but critically, neither uses channel
  concatenation to expand the input channels. LTX-2 maintains 128 latent channels for ALL model variants.

  ---
  1. Input Channel Dimensions (The Core Answer)

SqrtRyan

from rp import *
from rp.git.CommonSource.projects.mask_generator.random_video_mask_generator import get_random_video_mask
def megamask():
    S=10
    minimasks=[get_random_video_mask() for x in range(S*S)]
    video=tiled_videos(minimasks)
    video=slowmo_video_via_rife(video)
    video=slowmo_video_via_rife(video)
    video=slowmo_video_via_rife(video)
    video=slowmo_video_via_rife(video)

SqrtRyan

╭─── Claude Code v2.0.73 ──────────────────────────────────────────────────────────────────────────────╮
│                                  │ Tips for getting started                                          │
│           Welcome back!          │ Run /init to create a CLAUDE.md file with instructions for Claude │
│                                  │ ───────────────────────────────────────────────────────────────── │
│            * ▐▛███▜▌ *           │ Recent activity                                                   │
│           * ▝▜█████▛▘ *          │ No recent activity                                                │
│            *  ▘▘ ▝▝  *           │                                                                   │
│                                  │                                                                   │
│   Opus 4.5 · API Usage Billing   │                                                                   │

SqrtRyan

import rp
from google import genai
from google.genai.types import (
    GenerateContentConfig,
    Modality,
    ThinkingConfig,
    ImageConfig,
)
import os

SqrtRyan

url = "https://huggingface.co/datasets/OneOverZero/Calvin__task_ABC_D_h5__training_s224_h"
dataset_parent = "/"
clone_command = f"""
    cd {dataset_parent}
    git lfs install
    git clone {url}
"""  # sh
dataset_folder = path_join(dataset_parent, get_folder_name(url))
if not folder_exists(dataset_folder):
    r._run_sys_command(clone_command)

SqrtRyan

def film_strip(video, length=None, height=None, width=None, vertical=False):
    """
    Create a film strip effect from a video sequence.
    
    Args:
        video: List of video frames/images
        length: Optional number of frames to use (defaults to all frames)
        height: Height to resize frames to (default: 480)
        width: Width to resize frames to (default: 720)
    

SqrtRyan

#CYCLESORT!!!
def swap(a,i,j):
    a[i],a[j]=a[j],a[i]

class Counted:
    def __repr__(self):
        return 'X'
X=Counted()

a=shuffled([3,1,7,8,8,3,3,3,2,4,5,5,0])

SqrtRyan

from rp.git.Figures.film_strip.film_strip import film_strip

video = "/Users/ryan/CleanCode/Projects/Google2025_Paper/OverleafGit/figures_src/windmill_chase.mp4"
video = load_video(video, use_cache=True)
video = as_float_images(video)
# video=mean(video[:-3],video[1:-2],video[2:-1],video[3:])

before, after = split_tensor_into_regions(video, 1, 1, 2)

frames = [0, 5, 10]

SqrtRyan

from rp.git.Figures.film_strip.film_strip import film_strip
video='/Users/ryan/CleanCode/Projects/Google2025_Paper/OverleafGit/figures_src/CatFish.mp4'
video=load_video(video, use_cache=True)
for vert in [True,False]:
    before=video[:,80:,:720]
    after=video[:,80:,-720:]

    #display_video(after)
    frames=[0,25,46]