nvidia/Cosmos3-Super-Text2Image
1901.7k
1{2 "_class_name": "Cosmos3OmniTransformer",3 "_diffusers_version": "0.37.1",4 "action_dim": 32,5 "action_gen": false,6 "attention_bias": false,7 "attention_dropout": 0.0,8 "base_fps": 24,9 "dtype": "bfloat16",10 "enable_fps_modulation": true,11 "freeze_und": false,12 "head_dim": 128,13 "hidden_act": "silu",14 "hidden_size": 5120,15 "initializer_range": 0.02,16 "intermediate_size": 25600,17 "joint_attn_implementation": "two_way",18 "latent_channel": 48,19 "latent_patch_size": 2,20 "max_action_dim": 32,21 "max_position_embeddings": 262144,22 "model_type": "qwen3_vl_text",23 "num_attention_heads": 64,24 "num_embodiment_domains": 32,25 "num_hidden_layers": 64,26 "num_key_value_heads": 8,27 "patch_latent_dim": 192,28 "position_embedding_type": "unified_3d_mrope",29 "qk_norm": false,30 "qk_norm_for_diffusion": true,31 "qk_norm_for_text": true,32 "rms_norm_eps": 1e-06,33 "rope_scaling": {34 "mrope_interleaved": true,35 "mrope_section": [36 24,37 20,38 2039 ],40 "rope_type": "default"41 },42 "rope_theta": 5000000,43 "sound_dim": 64,44 "sound_gen": true,45 "sound_latent_fps": 25,46 "temporal_compression_factor_sound": 1,47 "timestep_scale": 0.001,48 "unified_3d_mrope_reset_spatial_ids": true,49 "unified_3d_mrope_temporal_modality_margin": 15000,50 "use_cache": true,51 "use_moe": true,52 "video_temporal_causal": false,53 "vocab_size": 15193654}55 