| dataloader_val: | |
| dataset: | |
| num_frames: 241 | |
| model: | |
| conditioner: | |
| tokenizer_path: text_tokenizer | |
| student: | |
| config: | |
| attention_bias: false | |
| base_fps: 24 | |
| enable_fps_modulation: true | |
| first_frame_reference_embedding: true | |
| head_dim: 128 | |
| hidden_size: 4096 | |
| intermediate_size: 12288 | |
| latent_channels: 48 | |
| mrope_section: | |
| - 24 | |
| - 20 | |
| - 20 | |
| num_attention_heads: 32 | |
| num_hidden_layers: 36 | |
| num_key_value_heads: 8 | |
| patch_size: 2 | |
| rms_norm_eps: 1.0e-06 | |
| rope_theta: 5000000.0 | |
| special_tokens: | |
| end_of_generation: 151653 | |
| eos_token_id: 151645 | |
| start_of_generation: 151652 | |
| temporal_modality_margin: 15000 | |
| text_bucket_tokens: 514 | |
| timestep_scale: 0.001 | |
| vocab_size: 151936 | |
| lora: null | |