{ "patch_size": [1, 2, 2], "model_type": "ti2v", "dim": 3072, "ffn_dim": 14336, "freq_dim": 256, "num_heads": 24, "num_layers": 30, "num_double_layers": 10, "num_single_layers": 20, "vid_in_dim": 48, "vid_out_dim": 48, "audio_in_dim": 128, "audio_out_dim": 128, "text_len": 512, "window_size": [-1, -1], "qk_norm": true, "cross_attn_norm": true, "eps": 1e-6, "temporal_rope_scaling_factor": 0.24 }