Astronex-Lab/Astronex-World
LoRA-style vs Wan-AI/Wan2.2-TI2V-5B. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of 0, 2, k, o, q, v changed, in 30 of 30 layers; norms, biases and embeddings are untouched.
Tensors changed: Attention 240/480, Norms 0/180, MLP 60/120, other 0/35, Embeddings 0/10