caption_channels value changed

Author: prishajain1

src/maxdiffusion/pipelines/ltx2/ltx2_pipeline.py

@@ -129,7 +129,7 @@ def create_model(rngs: nnx.Rngs, ltx2_config: dict):
         "audio_attention_head_dim": 64,
         "audio_cross_attention_dim": 2048,
         "num_layers": 48,
-        "caption_channels": 3840,
+        "caption_channels": 4096,
         "audio_caption_channels": 2048,
         "use_prompt_embeddings": True,
         "gated_attn": False,