{ "": { "type": "GuidedScriptModule" }, "model": { "type": "GestureDecoder", "nfeats": 42, "chunk_size": 5, "past_size": 75, "future_size": 5, "latent_dim": 512, "num_layers": 18, "num_heads": 8, "cond_feature_dim": 1024 }, "model.cond_dropout": { "type": "MulticonditionDropout" }, "model.rope": { "type": "RoPE" }, "model.time_mlp": { "type": "Sequential" }, "model.time_mlp.0": { "type": "SinusoidalPosEmb", "dim": 512 }, "model.time_mlp.1": { "type": "Linear" }, "model.time_mlp.2": { "type": "Mish" }, "model.time_mlp.3": { "type": "Linear" }, "model.quality_embed": { "type": "Sequential" }, "model.quality_embed.0": { "type": "SinusoidalPosEmb", "dim": 512 }, "model.quality_embed.1": { "type": "Linear" }, "model.quality_embed.2": { "type": "GELU" }, "model.quality_embed.3": { "type": "Linear" }, "model.norm_base_kp_cond": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.norm_past_cond": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.input_projection": { "type": "Linear" }, "model.cond_encoder_self": { "type": "Sequential" }, "model.cond_encoder_self.0": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.cond_encoder_self.1": { "type": "TransformerEncoderLayer" }, "model.cond_encoder_self.1.self_attn": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": true }, "model.cond_encoder_self.1.self_attn.to_q": { "type": "Linear" }, "model.cond_encoder_self.1.self_attn.to_k": { "type": "Linear" }, "model.cond_encoder_self.1.self_attn.to_v": { "type": "Linear" }, "model.cond_encoder_self.1.self_attn.to_gate": { "type": "Linear" }, "model.cond_encoder_self.1.self_attn.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.cond_encoder_self.1.self_attn.proj": { "type": "Linear" }, "model.cond_encoder_self.1.self_attn.rotary": { "type": "RoPE" }, "model.cond_encoder_self.1.linear1": { "type": "Linear" }, "model.cond_encoder_self.1.linear2": { "type": "Linear" }, "model.cond_encoder_self.1.norm1": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.cond_encoder_self.1.norm2": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.cond_encoder_self.2": { "type": "TransformerEncoderLayer" }, "model.cond_encoder_self.2.self_attn": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": true }, "model.cond_encoder_self.2.self_attn.to_q": { "type": "Linear" }, "model.cond_encoder_self.2.self_attn.to_k": { "type": "Linear" }, "model.cond_encoder_self.2.self_attn.to_v": { "type": "Linear" }, "model.cond_encoder_self.2.self_attn.to_gate": { "type": "Linear" }, "model.cond_encoder_self.2.self_attn.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.cond_encoder_self.2.self_attn.proj": { "type": "Linear" }, "model.cond_encoder_self.2.self_attn.rotary": { "type": "RoPE" }, "model.cond_encoder_self.2.linear1": { "type": "Linear" }, "model.cond_encoder_self.2.linear2": { "type": "Linear" }, "model.cond_encoder_self.2.norm1": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.cond_encoder_self.2.norm2": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.cond_encoder_self.3": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.cond_encoder_other": { "type": "Sequential" }, "model.cond_encoder_other.0": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.cond_encoder_other.1": { "type": "TransformerEncoderLayer" }, "model.cond_encoder_other.1.self_attn": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": true }, "model.cond_encoder_other.1.self_attn.to_q": { "type": "Linear" }, "model.cond_encoder_other.1.self_attn.to_k": { "type": "Linear" }, "model.cond_encoder_other.1.self_attn.to_v": { "type": "Linear" }, "model.cond_encoder_other.1.self_attn.to_gate": { "type": "Linear" }, "model.cond_encoder_other.1.self_attn.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.cond_encoder_other.1.self_attn.proj": { "type": "Linear" }, "model.cond_encoder_other.1.self_attn.rotary": { "type": "RoPE" }, "model.cond_encoder_other.1.linear1": { "type": "Linear" }, "model.cond_encoder_other.1.linear2": { "type": "Linear" }, "model.cond_encoder_other.1.norm1": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.cond_encoder_other.1.norm2": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.cond_encoder_other.2": { "type": "TransformerEncoderLayer" }, "model.cond_encoder_other.2.self_attn": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": true }, "model.cond_encoder_other.2.self_attn.to_q": { "type": "Linear" }, "model.cond_encoder_other.2.self_attn.to_k": { "type": "Linear" }, "model.cond_encoder_other.2.self_attn.to_v": { "type": "Linear" }, "model.cond_encoder_other.2.self_attn.to_gate": { "type": "Linear" }, "model.cond_encoder_other.2.self_attn.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.cond_encoder_other.2.self_attn.proj": { "type": "Linear" }, "model.cond_encoder_other.2.self_attn.rotary": { "type": "RoPE" }, "model.cond_encoder_other.2.linear1": { "type": "Linear" }, "model.cond_encoder_other.2.linear2": { "type": "Linear" }, "model.cond_encoder_other.2.norm1": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.cond_encoder_other.2.norm2": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.cond_encoder_other.3": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.audio_fusion_gate": { "type": "AudioFusionGate" }, "model.audio_fusion_gate.ln_x": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.audio_fusion_gate.ln_y": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.audio_fusion_gate.proj_x": { "type": "Linear" }, "model.audio_fusion_gate.proj_y": { "type": "Linear" }, "model.audio_fusion_gate.gate": { "type": "Sequential" }, "model.audio_fusion_gate.gate.0": { "type": "Linear" }, "model.audio_fusion_gate.gate.1": { "type": "GELU" }, "model.audio_fusion_gate.gate.2": { "type": "Linear" }, "model.canon_projection": { "type": "Sequential" }, "model.canon_projection.0": { "type": "Linear" }, "model.canon_projection.1": { "type": "GELU" }, "model.canon_projection.2": { "type": "Linear" }, "model.seqTransDecoder": { "type": "DecoderLayerStack" }, "model.seqTransDecoder.stack": { "type": "ModuleList" }, "model.seqTransDecoder.stack.0": { "type": "TransformerDecoderLayer" }, "model.seqTransDecoder.stack.0.self_attn": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": true }, "model.seqTransDecoder.stack.0.self_attn.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.0.self_attn.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.0.self_attn.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.0.self_attn.to_gate": { "type": "Linear" }, "model.seqTransDecoder.stack.0.self_attn.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.0.self_attn.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.0.self_attn.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.0.kp_cross": { "type": "Sequential" }, "model.seqTransDecoder.stack.0.kp_cross.0": { "type": "GELU" }, "model.seqTransDecoder.stack.0.kp_cross.1": { "type": "Linear" }, "model.seqTransDecoder.stack.0.multihead_attn_last_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.0.multihead_attn_last_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.0.multihead_attn_last_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.0.multihead_attn_last_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.0.multihead_attn_last_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.0.multihead_attn_last_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.0.multihead_attn_last_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.0.multihead_attn_last_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.0.multihead_attn_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.0.multihead_attn_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.0.multihead_attn_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.0.multihead_attn_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.0.multihead_attn_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.0.multihead_attn_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.0.multihead_attn_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.0.multihead_attn_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.0.multihead_attn_audio_self": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.0.multihead_attn_audio_self.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.0.multihead_attn_audio_self.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.0.multihead_attn_audio_self.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.0.multihead_attn_audio_self.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.0.multihead_attn_audio_self.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.0.multihead_attn_audio_self.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.0.multihead_attn_audio_self.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.0.multihead_attn_audio_other": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.0.multihead_attn_audio_other.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.0.multihead_attn_audio_other.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.0.multihead_attn_audio_other.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.0.multihead_attn_audio_other.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.0.multihead_attn_audio_other.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.0.multihead_attn_audio_other.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.0.multihead_attn_audio_other.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.0.linear1": { "type": "Linear" }, "model.seqTransDecoder.stack.0.linear2": { "type": "Linear" }, "model.seqTransDecoder.stack.0.norm1": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.0.norm2": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.0.norm3": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.0.adaln1": { "type": "AdaLN" }, "model.seqTransDecoder.stack.0.adaln1.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.0.adaln1.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.0.adaln1.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.0.adaln2": { "type": "AdaLN" }, "model.seqTransDecoder.stack.0.adaln2.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.0.adaln2.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.0.adaln2.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.0.adaln3": { "type": "AdaLN" }, "model.seqTransDecoder.stack.0.adaln3.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.0.adaln3.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.0.adaln3.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.1": { "type": "TransformerDecoderLayer" }, "model.seqTransDecoder.stack.1.self_attn": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": true }, "model.seqTransDecoder.stack.1.self_attn.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.1.self_attn.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.1.self_attn.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.1.self_attn.to_gate": { "type": "Linear" }, "model.seqTransDecoder.stack.1.self_attn.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.1.self_attn.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.1.self_attn.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.1.kp_cross": { "type": "Sequential" }, "model.seqTransDecoder.stack.1.kp_cross.0": { "type": "GELU" }, "model.seqTransDecoder.stack.1.kp_cross.1": { "type": "Linear" }, "model.seqTransDecoder.stack.1.multihead_attn_last_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.1.multihead_attn_last_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.1.multihead_attn_last_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.1.multihead_attn_last_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.1.multihead_attn_last_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.1.multihead_attn_last_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.1.multihead_attn_last_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.1.multihead_attn_last_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.1.multihead_attn_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.1.multihead_attn_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.1.multihead_attn_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.1.multihead_attn_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.1.multihead_attn_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.1.multihead_attn_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.1.multihead_attn_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.1.multihead_attn_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.1.multihead_attn_audio_self": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.1.multihead_attn_audio_self.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.1.multihead_attn_audio_self.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.1.multihead_attn_audio_self.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.1.multihead_attn_audio_self.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.1.multihead_attn_audio_self.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.1.multihead_attn_audio_self.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.1.multihead_attn_audio_self.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.1.multihead_attn_audio_other": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.1.multihead_attn_audio_other.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.1.multihead_attn_audio_other.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.1.multihead_attn_audio_other.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.1.multihead_attn_audio_other.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.1.multihead_attn_audio_other.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.1.multihead_attn_audio_other.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.1.multihead_attn_audio_other.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.1.linear1": { "type": "Linear" }, "model.seqTransDecoder.stack.1.linear2": { "type": "Linear" }, "model.seqTransDecoder.stack.1.norm1": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.1.norm2": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.1.norm3": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.1.adaln1": { "type": "AdaLN" }, "model.seqTransDecoder.stack.1.adaln1.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.1.adaln1.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.1.adaln1.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.1.adaln2": { "type": "AdaLN" }, "model.seqTransDecoder.stack.1.adaln2.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.1.adaln2.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.1.adaln2.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.1.adaln3": { "type": "AdaLN" }, "model.seqTransDecoder.stack.1.adaln3.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.1.adaln3.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.1.adaln3.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.2": { "type": "TransformerDecoderLayer" }, "model.seqTransDecoder.stack.2.self_attn": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": true }, "model.seqTransDecoder.stack.2.self_attn.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.2.self_attn.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.2.self_attn.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.2.self_attn.to_gate": { "type": "Linear" }, "model.seqTransDecoder.stack.2.self_attn.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.2.self_attn.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.2.self_attn.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.2.kp_cross": { "type": "Sequential" }, "model.seqTransDecoder.stack.2.kp_cross.0": { "type": "GELU" }, "model.seqTransDecoder.stack.2.kp_cross.1": { "type": "Linear" }, "model.seqTransDecoder.stack.2.multihead_attn_last_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.2.multihead_attn_last_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.2.multihead_attn_last_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.2.multihead_attn_last_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.2.multihead_attn_last_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.2.multihead_attn_last_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.2.multihead_attn_last_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.2.multihead_attn_last_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.2.multihead_attn_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.2.multihead_attn_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.2.multihead_attn_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.2.multihead_attn_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.2.multihead_attn_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.2.multihead_attn_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.2.multihead_attn_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.2.multihead_attn_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.2.multihead_attn_audio_self": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.2.multihead_attn_audio_self.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.2.multihead_attn_audio_self.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.2.multihead_attn_audio_self.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.2.multihead_attn_audio_self.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.2.multihead_attn_audio_self.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.2.multihead_attn_audio_self.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.2.multihead_attn_audio_self.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.2.multihead_attn_audio_other": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.2.multihead_attn_audio_other.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.2.multihead_attn_audio_other.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.2.multihead_attn_audio_other.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.2.multihead_attn_audio_other.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.2.multihead_attn_audio_other.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.2.multihead_attn_audio_other.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.2.multihead_attn_audio_other.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.2.linear1": { "type": "Linear" }, "model.seqTransDecoder.stack.2.linear2": { "type": "Linear" }, "model.seqTransDecoder.stack.2.norm1": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.2.norm2": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.2.norm3": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.2.adaln1": { "type": "AdaLN" }, "model.seqTransDecoder.stack.2.adaln1.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.2.adaln1.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.2.adaln1.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.2.adaln2": { "type": "AdaLN" }, "model.seqTransDecoder.stack.2.adaln2.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.2.adaln2.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.2.adaln2.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.2.adaln3": { "type": "AdaLN" }, "model.seqTransDecoder.stack.2.adaln3.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.2.adaln3.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.2.adaln3.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.3": { "type": "TransformerDecoderLayer" }, "model.seqTransDecoder.stack.3.self_attn": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": true }, "model.seqTransDecoder.stack.3.self_attn.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.3.self_attn.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.3.self_attn.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.3.self_attn.to_gate": { "type": "Linear" }, "model.seqTransDecoder.stack.3.self_attn.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.3.self_attn.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.3.self_attn.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.3.kp_cross": { "type": "Sequential" }, "model.seqTransDecoder.stack.3.kp_cross.0": { "type": "GELU" }, "model.seqTransDecoder.stack.3.kp_cross.1": { "type": "Linear" }, "model.seqTransDecoder.stack.3.multihead_attn_last_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.3.multihead_attn_last_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.3.multihead_attn_last_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.3.multihead_attn_last_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.3.multihead_attn_last_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.3.multihead_attn_last_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.3.multihead_attn_last_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.3.multihead_attn_last_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.3.multihead_attn_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.3.multihead_attn_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.3.multihead_attn_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.3.multihead_attn_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.3.multihead_attn_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.3.multihead_attn_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.3.multihead_attn_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.3.multihead_attn_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.3.multihead_attn_audio_self": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.3.multihead_attn_audio_self.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.3.multihead_attn_audio_self.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.3.multihead_attn_audio_self.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.3.multihead_attn_audio_self.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.3.multihead_attn_audio_self.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.3.multihead_attn_audio_self.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.3.multihead_attn_audio_self.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.3.multihead_attn_audio_other": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.3.multihead_attn_audio_other.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.3.multihead_attn_audio_other.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.3.multihead_attn_audio_other.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.3.multihead_attn_audio_other.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.3.multihead_attn_audio_other.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.3.multihead_attn_audio_other.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.3.multihead_attn_audio_other.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.3.linear1": { "type": "Linear" }, "model.seqTransDecoder.stack.3.linear2": { "type": "Linear" }, "model.seqTransDecoder.stack.3.norm1": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.3.norm2": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.3.norm3": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.3.adaln1": { "type": "AdaLN" }, "model.seqTransDecoder.stack.3.adaln1.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.3.adaln1.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.3.adaln1.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.3.adaln2": { "type": "AdaLN" }, "model.seqTransDecoder.stack.3.adaln2.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.3.adaln2.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.3.adaln2.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.3.adaln3": { "type": "AdaLN" }, "model.seqTransDecoder.stack.3.adaln3.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.3.adaln3.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.3.adaln3.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.4": { "type": "TransformerDecoderLayer" }, "model.seqTransDecoder.stack.4.self_attn": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": true }, "model.seqTransDecoder.stack.4.self_attn.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.4.self_attn.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.4.self_attn.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.4.self_attn.to_gate": { "type": "Linear" }, "model.seqTransDecoder.stack.4.self_attn.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.4.self_attn.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.4.self_attn.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.4.kp_cross": { "type": "Sequential" }, "model.seqTransDecoder.stack.4.kp_cross.0": { "type": "GELU" }, "model.seqTransDecoder.stack.4.kp_cross.1": { "type": "Linear" }, "model.seqTransDecoder.stack.4.multihead_attn_last_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.4.multihead_attn_last_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.4.multihead_attn_last_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.4.multihead_attn_last_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.4.multihead_attn_last_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.4.multihead_attn_last_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.4.multihead_attn_last_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.4.multihead_attn_last_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.4.multihead_attn_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.4.multihead_attn_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.4.multihead_attn_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.4.multihead_attn_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.4.multihead_attn_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.4.multihead_attn_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.4.multihead_attn_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.4.multihead_attn_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.4.multihead_attn_audio_self": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.4.multihead_attn_audio_self.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.4.multihead_attn_audio_self.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.4.multihead_attn_audio_self.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.4.multihead_attn_audio_self.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.4.multihead_attn_audio_self.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.4.multihead_attn_audio_self.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.4.multihead_attn_audio_self.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.4.multihead_attn_audio_other": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.4.multihead_attn_audio_other.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.4.multihead_attn_audio_other.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.4.multihead_attn_audio_other.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.4.multihead_attn_audio_other.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.4.multihead_attn_audio_other.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.4.multihead_attn_audio_other.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.4.multihead_attn_audio_other.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.4.linear1": { "type": "Linear" }, "model.seqTransDecoder.stack.4.linear2": { "type": "Linear" }, "model.seqTransDecoder.stack.4.norm1": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.4.norm2": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.4.norm3": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.4.adaln1": { "type": "AdaLN" }, "model.seqTransDecoder.stack.4.adaln1.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.4.adaln1.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.4.adaln1.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.4.adaln2": { "type": "AdaLN" }, "model.seqTransDecoder.stack.4.adaln2.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.4.adaln2.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.4.adaln2.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.4.adaln3": { "type": "AdaLN" }, "model.seqTransDecoder.stack.4.adaln3.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.4.adaln3.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.4.adaln3.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.5": { "type": "TransformerDecoderLayer" }, "model.seqTransDecoder.stack.5.self_attn": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": true }, "model.seqTransDecoder.stack.5.self_attn.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.5.self_attn.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.5.self_attn.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.5.self_attn.to_gate": { "type": "Linear" }, "model.seqTransDecoder.stack.5.self_attn.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.5.self_attn.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.5.self_attn.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.5.kp_cross": { "type": "Sequential" }, "model.seqTransDecoder.stack.5.kp_cross.0": { "type": "GELU" }, "model.seqTransDecoder.stack.5.kp_cross.1": { "type": "Linear" }, "model.seqTransDecoder.stack.5.multihead_attn_last_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.5.multihead_attn_last_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.5.multihead_attn_last_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.5.multihead_attn_last_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.5.multihead_attn_last_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.5.multihead_attn_last_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.5.multihead_attn_last_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.5.multihead_attn_last_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.5.multihead_attn_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.5.multihead_attn_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.5.multihead_attn_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.5.multihead_attn_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.5.multihead_attn_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.5.multihead_attn_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.5.multihead_attn_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.5.multihead_attn_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.5.multihead_attn_audio_self": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.5.multihead_attn_audio_self.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.5.multihead_attn_audio_self.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.5.multihead_attn_audio_self.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.5.multihead_attn_audio_self.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.5.multihead_attn_audio_self.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.5.multihead_attn_audio_self.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.5.multihead_attn_audio_self.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.5.multihead_attn_audio_other": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.5.multihead_attn_audio_other.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.5.multihead_attn_audio_other.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.5.multihead_attn_audio_other.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.5.multihead_attn_audio_other.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.5.multihead_attn_audio_other.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.5.multihead_attn_audio_other.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.5.multihead_attn_audio_other.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.5.linear1": { "type": "Linear" }, "model.seqTransDecoder.stack.5.linear2": { "type": "Linear" }, "model.seqTransDecoder.stack.5.norm1": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.5.norm2": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.5.norm3": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.5.adaln1": { "type": "AdaLN" }, "model.seqTransDecoder.stack.5.adaln1.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.5.adaln1.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.5.adaln1.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.5.adaln2": { "type": "AdaLN" }, "model.seqTransDecoder.stack.5.adaln2.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.5.adaln2.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.5.adaln2.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.5.adaln3": { "type": "AdaLN" }, "model.seqTransDecoder.stack.5.adaln3.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.5.adaln3.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.5.adaln3.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.6": { "type": "TransformerDecoderLayer" }, "model.seqTransDecoder.stack.6.self_attn": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": true }, "model.seqTransDecoder.stack.6.self_attn.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.6.self_attn.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.6.self_attn.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.6.self_attn.to_gate": { "type": "Linear" }, "model.seqTransDecoder.stack.6.self_attn.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.6.self_attn.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.6.self_attn.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.6.kp_cross": { "type": "Sequential" }, "model.seqTransDecoder.stack.6.kp_cross.0": { "type": "GELU" }, "model.seqTransDecoder.stack.6.kp_cross.1": { "type": "Linear" }, "model.seqTransDecoder.stack.6.multihead_attn_last_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.6.multihead_attn_last_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.6.multihead_attn_last_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.6.multihead_attn_last_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.6.multihead_attn_last_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.6.multihead_attn_last_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.6.multihead_attn_last_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.6.multihead_attn_last_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.6.multihead_attn_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.6.multihead_attn_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.6.multihead_attn_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.6.multihead_attn_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.6.multihead_attn_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.6.multihead_attn_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.6.multihead_attn_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.6.multihead_attn_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.6.multihead_attn_audio_self": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.6.multihead_attn_audio_self.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.6.multihead_attn_audio_self.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.6.multihead_attn_audio_self.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.6.multihead_attn_audio_self.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.6.multihead_attn_audio_self.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.6.multihead_attn_audio_self.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.6.multihead_attn_audio_self.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.6.multihead_attn_audio_other": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.6.multihead_attn_audio_other.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.6.multihead_attn_audio_other.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.6.multihead_attn_audio_other.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.6.multihead_attn_audio_other.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.6.multihead_attn_audio_other.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.6.multihead_attn_audio_other.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.6.multihead_attn_audio_other.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.6.linear1": { "type": "Linear" }, "model.seqTransDecoder.stack.6.linear2": { "type": "Linear" }, "model.seqTransDecoder.stack.6.norm1": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.6.norm2": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.6.norm3": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.6.adaln1": { "type": "AdaLN" }, "model.seqTransDecoder.stack.6.adaln1.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.6.adaln1.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.6.adaln1.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.6.adaln2": { "type": "AdaLN" }, "model.seqTransDecoder.stack.6.adaln2.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.6.adaln2.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.6.adaln2.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.6.adaln3": { "type": "AdaLN" }, "model.seqTransDecoder.stack.6.adaln3.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.6.adaln3.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.6.adaln3.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.7": { "type": "TransformerDecoderLayer" }, "model.seqTransDecoder.stack.7.self_attn": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": true }, "model.seqTransDecoder.stack.7.self_attn.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.7.self_attn.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.7.self_attn.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.7.self_attn.to_gate": { "type": "Linear" }, "model.seqTransDecoder.stack.7.self_attn.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.7.self_attn.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.7.self_attn.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.7.kp_cross": { "type": "Sequential" }, "model.seqTransDecoder.stack.7.kp_cross.0": { "type": "GELU" }, "model.seqTransDecoder.stack.7.kp_cross.1": { "type": "Linear" }, "model.seqTransDecoder.stack.7.multihead_attn_last_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.7.multihead_attn_last_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.7.multihead_attn_last_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.7.multihead_attn_last_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.7.multihead_attn_last_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.7.multihead_attn_last_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.7.multihead_attn_last_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.7.multihead_attn_last_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.7.multihead_attn_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.7.multihead_attn_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.7.multihead_attn_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.7.multihead_attn_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.7.multihead_attn_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.7.multihead_attn_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.7.multihead_attn_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.7.multihead_attn_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.7.multihead_attn_audio_self": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.7.multihead_attn_audio_self.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.7.multihead_attn_audio_self.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.7.multihead_attn_audio_self.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.7.multihead_attn_audio_self.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.7.multihead_attn_audio_self.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.7.multihead_attn_audio_self.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.7.multihead_attn_audio_self.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.7.multihead_attn_audio_other": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.7.multihead_attn_audio_other.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.7.multihead_attn_audio_other.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.7.multihead_attn_audio_other.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.7.multihead_attn_audio_other.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.7.multihead_attn_audio_other.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.7.multihead_attn_audio_other.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.7.multihead_attn_audio_other.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.7.linear1": { "type": "Linear" }, "model.seqTransDecoder.stack.7.linear2": { "type": "Linear" }, "model.seqTransDecoder.stack.7.norm1": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.7.norm2": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.7.norm3": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.7.adaln1": { "type": "AdaLN" }, "model.seqTransDecoder.stack.7.adaln1.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.7.adaln1.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.7.adaln1.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.7.adaln2": { "type": "AdaLN" }, "model.seqTransDecoder.stack.7.adaln2.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.7.adaln2.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.7.adaln2.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.7.adaln3": { "type": "AdaLN" }, "model.seqTransDecoder.stack.7.adaln3.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.7.adaln3.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.7.adaln3.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.8": { "type": "TransformerDecoderLayer" }, "model.seqTransDecoder.stack.8.self_attn": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": true }, "model.seqTransDecoder.stack.8.self_attn.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.8.self_attn.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.8.self_attn.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.8.self_attn.to_gate": { "type": "Linear" }, "model.seqTransDecoder.stack.8.self_attn.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.8.self_attn.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.8.self_attn.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.8.kp_cross": { "type": "Sequential" }, "model.seqTransDecoder.stack.8.kp_cross.0": { "type": "GELU" }, "model.seqTransDecoder.stack.8.kp_cross.1": { "type": "Linear" }, "model.seqTransDecoder.stack.8.multihead_attn_last_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.8.multihead_attn_last_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.8.multihead_attn_last_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.8.multihead_attn_last_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.8.multihead_attn_last_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.8.multihead_attn_last_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.8.multihead_attn_last_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.8.multihead_attn_last_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.8.multihead_attn_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.8.multihead_attn_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.8.multihead_attn_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.8.multihead_attn_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.8.multihead_attn_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.8.multihead_attn_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.8.multihead_attn_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.8.multihead_attn_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.8.multihead_attn_audio_self": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.8.multihead_attn_audio_self.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.8.multihead_attn_audio_self.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.8.multihead_attn_audio_self.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.8.multihead_attn_audio_self.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.8.multihead_attn_audio_self.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.8.multihead_attn_audio_self.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.8.multihead_attn_audio_self.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.8.multihead_attn_audio_other": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.8.multihead_attn_audio_other.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.8.multihead_attn_audio_other.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.8.multihead_attn_audio_other.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.8.multihead_attn_audio_other.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.8.multihead_attn_audio_other.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.8.multihead_attn_audio_other.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.8.multihead_attn_audio_other.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.8.linear1": { "type": "Linear" }, "model.seqTransDecoder.stack.8.linear2": { "type": "Linear" }, "model.seqTransDecoder.stack.8.norm1": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.8.norm2": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.8.norm3": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.8.adaln1": { "type": "AdaLN" }, "model.seqTransDecoder.stack.8.adaln1.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.8.adaln1.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.8.adaln1.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.8.adaln2": { "type": "AdaLN" }, "model.seqTransDecoder.stack.8.adaln2.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.8.adaln2.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.8.adaln2.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.8.adaln3": { "type": "AdaLN" }, "model.seqTransDecoder.stack.8.adaln3.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.8.adaln3.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.8.adaln3.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.9": { "type": "TransformerDecoderLayer" }, "model.seqTransDecoder.stack.9.self_attn": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": true }, "model.seqTransDecoder.stack.9.self_attn.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.9.self_attn.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.9.self_attn.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.9.self_attn.to_gate": { "type": "Linear" }, "model.seqTransDecoder.stack.9.self_attn.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.9.self_attn.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.9.self_attn.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.9.kp_cross": { "type": "Sequential" }, "model.seqTransDecoder.stack.9.kp_cross.0": { "type": "GELU" }, "model.seqTransDecoder.stack.9.kp_cross.1": { "type": "Linear" }, "model.seqTransDecoder.stack.9.multihead_attn_last_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.9.multihead_attn_last_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.9.multihead_attn_last_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.9.multihead_attn_last_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.9.multihead_attn_last_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.9.multihead_attn_last_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.9.multihead_attn_last_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.9.multihead_attn_last_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.9.multihead_attn_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.9.multihead_attn_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.9.multihead_attn_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.9.multihead_attn_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.9.multihead_attn_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.9.multihead_attn_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.9.multihead_attn_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.9.multihead_attn_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.9.multihead_attn_audio_self": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.9.multihead_attn_audio_self.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.9.multihead_attn_audio_self.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.9.multihead_attn_audio_self.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.9.multihead_attn_audio_self.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.9.multihead_attn_audio_self.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.9.multihead_attn_audio_self.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.9.multihead_attn_audio_self.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.9.multihead_attn_audio_other": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.9.multihead_attn_audio_other.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.9.multihead_attn_audio_other.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.9.multihead_attn_audio_other.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.9.multihead_attn_audio_other.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.9.multihead_attn_audio_other.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.9.multihead_attn_audio_other.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.9.multihead_attn_audio_other.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.9.linear1": { "type": "Linear" }, "model.seqTransDecoder.stack.9.linear2": { "type": "Linear" }, "model.seqTransDecoder.stack.9.norm1": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.9.norm2": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.9.norm3": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.9.adaln1": { "type": "AdaLN" }, "model.seqTransDecoder.stack.9.adaln1.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.9.adaln1.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.9.adaln1.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.9.adaln2": { "type": "AdaLN" }, "model.seqTransDecoder.stack.9.adaln2.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.9.adaln2.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.9.adaln2.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.9.adaln3": { "type": "AdaLN" }, "model.seqTransDecoder.stack.9.adaln3.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.9.adaln3.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.9.adaln3.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.10": { "type": "TransformerDecoderLayer" }, "model.seqTransDecoder.stack.10.self_attn": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": true }, "model.seqTransDecoder.stack.10.self_attn.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.10.self_attn.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.10.self_attn.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.10.self_attn.to_gate": { "type": "Linear" }, "model.seqTransDecoder.stack.10.self_attn.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.10.self_attn.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.10.self_attn.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.10.kp_cross": { "type": "Sequential" }, "model.seqTransDecoder.stack.10.kp_cross.0": { "type": "GELU" }, "model.seqTransDecoder.stack.10.kp_cross.1": { "type": "Linear" }, "model.seqTransDecoder.stack.10.multihead_attn_last_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.10.multihead_attn_last_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.10.multihead_attn_last_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.10.multihead_attn_last_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.10.multihead_attn_last_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.10.multihead_attn_last_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.10.multihead_attn_last_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.10.multihead_attn_last_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.10.multihead_attn_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.10.multihead_attn_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.10.multihead_attn_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.10.multihead_attn_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.10.multihead_attn_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.10.multihead_attn_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.10.multihead_attn_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.10.multihead_attn_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.10.multihead_attn_audio_self": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.10.multihead_attn_audio_self.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.10.multihead_attn_audio_self.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.10.multihead_attn_audio_self.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.10.multihead_attn_audio_self.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.10.multihead_attn_audio_self.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.10.multihead_attn_audio_self.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.10.multihead_attn_audio_self.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.10.multihead_attn_audio_other": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.10.multihead_attn_audio_other.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.10.multihead_attn_audio_other.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.10.multihead_attn_audio_other.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.10.multihead_attn_audio_other.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.10.multihead_attn_audio_other.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.10.multihead_attn_audio_other.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.10.multihead_attn_audio_other.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.10.linear1": { "type": "Linear" }, "model.seqTransDecoder.stack.10.linear2": { "type": "Linear" }, "model.seqTransDecoder.stack.10.norm1": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.10.norm2": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.10.norm3": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.10.adaln1": { "type": "AdaLN" }, "model.seqTransDecoder.stack.10.adaln1.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.10.adaln1.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.10.adaln1.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.10.adaln2": { "type": "AdaLN" }, "model.seqTransDecoder.stack.10.adaln2.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.10.adaln2.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.10.adaln2.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.10.adaln3": { "type": "AdaLN" }, "model.seqTransDecoder.stack.10.adaln3.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.10.adaln3.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.10.adaln3.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.11": { "type": "TransformerDecoderLayer" }, "model.seqTransDecoder.stack.11.self_attn": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": true }, "model.seqTransDecoder.stack.11.self_attn.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.11.self_attn.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.11.self_attn.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.11.self_attn.to_gate": { "type": "Linear" }, "model.seqTransDecoder.stack.11.self_attn.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.11.self_attn.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.11.self_attn.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.11.kp_cross": { "type": "Sequential" }, "model.seqTransDecoder.stack.11.kp_cross.0": { "type": "GELU" }, "model.seqTransDecoder.stack.11.kp_cross.1": { "type": "Linear" }, "model.seqTransDecoder.stack.11.multihead_attn_last_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.11.multihead_attn_last_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.11.multihead_attn_last_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.11.multihead_attn_last_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.11.multihead_attn_last_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.11.multihead_attn_last_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.11.multihead_attn_last_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.11.multihead_attn_last_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.11.multihead_attn_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.11.multihead_attn_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.11.multihead_attn_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.11.multihead_attn_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.11.multihead_attn_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.11.multihead_attn_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.11.multihead_attn_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.11.multihead_attn_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.11.multihead_attn_audio_self": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.11.multihead_attn_audio_self.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.11.multihead_attn_audio_self.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.11.multihead_attn_audio_self.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.11.multihead_attn_audio_self.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.11.multihead_attn_audio_self.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.11.multihead_attn_audio_self.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.11.multihead_attn_audio_self.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.11.multihead_attn_audio_other": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.11.multihead_attn_audio_other.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.11.multihead_attn_audio_other.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.11.multihead_attn_audio_other.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.11.multihead_attn_audio_other.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.11.multihead_attn_audio_other.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.11.multihead_attn_audio_other.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.11.multihead_attn_audio_other.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.11.linear1": { "type": "Linear" }, "model.seqTransDecoder.stack.11.linear2": { "type": "Linear" }, "model.seqTransDecoder.stack.11.norm1": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.11.norm2": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.11.norm3": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.11.adaln1": { "type": "AdaLN" }, "model.seqTransDecoder.stack.11.adaln1.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.11.adaln1.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.11.adaln1.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.11.adaln2": { "type": "AdaLN" }, "model.seqTransDecoder.stack.11.adaln2.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.11.adaln2.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.11.adaln2.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.11.adaln3": { "type": "AdaLN" }, "model.seqTransDecoder.stack.11.adaln3.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.11.adaln3.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.11.adaln3.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.12": { "type": "TransformerDecoderLayer" }, "model.seqTransDecoder.stack.12.self_attn": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": true }, "model.seqTransDecoder.stack.12.self_attn.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.12.self_attn.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.12.self_attn.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.12.self_attn.to_gate": { "type": "Linear" }, "model.seqTransDecoder.stack.12.self_attn.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.12.self_attn.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.12.self_attn.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.12.kp_cross": { "type": "Sequential" }, "model.seqTransDecoder.stack.12.kp_cross.0": { "type": "GELU" }, "model.seqTransDecoder.stack.12.kp_cross.1": { "type": "Linear" }, "model.seqTransDecoder.stack.12.multihead_attn_last_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.12.multihead_attn_last_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.12.multihead_attn_last_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.12.multihead_attn_last_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.12.multihead_attn_last_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.12.multihead_attn_last_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.12.multihead_attn_last_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.12.multihead_attn_last_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.12.multihead_attn_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.12.multihead_attn_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.12.multihead_attn_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.12.multihead_attn_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.12.multihead_attn_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.12.multihead_attn_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.12.multihead_attn_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.12.multihead_attn_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.12.multihead_attn_audio_self": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.12.multihead_attn_audio_self.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.12.multihead_attn_audio_self.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.12.multihead_attn_audio_self.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.12.multihead_attn_audio_self.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.12.multihead_attn_audio_self.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.12.multihead_attn_audio_self.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.12.multihead_attn_audio_self.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.12.multihead_attn_audio_other": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.12.multihead_attn_audio_other.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.12.multihead_attn_audio_other.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.12.multihead_attn_audio_other.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.12.multihead_attn_audio_other.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.12.multihead_attn_audio_other.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.12.multihead_attn_audio_other.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.12.multihead_attn_audio_other.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.12.linear1": { "type": "Linear" }, "model.seqTransDecoder.stack.12.linear2": { "type": "Linear" }, "model.seqTransDecoder.stack.12.norm1": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.12.norm2": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.12.norm3": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.12.adaln1": { "type": "AdaLN" }, "model.seqTransDecoder.stack.12.adaln1.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.12.adaln1.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.12.adaln1.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.12.adaln2": { "type": "AdaLN" }, "model.seqTransDecoder.stack.12.adaln2.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.12.adaln2.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.12.adaln2.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.12.adaln3": { "type": "AdaLN" }, "model.seqTransDecoder.stack.12.adaln3.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.12.adaln3.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.12.adaln3.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.13": { "type": "TransformerDecoderLayer" }, "model.seqTransDecoder.stack.13.self_attn": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": true }, "model.seqTransDecoder.stack.13.self_attn.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.13.self_attn.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.13.self_attn.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.13.self_attn.to_gate": { "type": "Linear" }, "model.seqTransDecoder.stack.13.self_attn.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.13.self_attn.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.13.self_attn.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.13.kp_cross": { "type": "Sequential" }, "model.seqTransDecoder.stack.13.kp_cross.0": { "type": "GELU" }, "model.seqTransDecoder.stack.13.kp_cross.1": { "type": "Linear" }, "model.seqTransDecoder.stack.13.multihead_attn_last_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.13.multihead_attn_last_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.13.multihead_attn_last_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.13.multihead_attn_last_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.13.multihead_attn_last_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.13.multihead_attn_last_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.13.multihead_attn_last_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.13.multihead_attn_last_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.13.multihead_attn_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.13.multihead_attn_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.13.multihead_attn_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.13.multihead_attn_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.13.multihead_attn_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.13.multihead_attn_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.13.multihead_attn_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.13.multihead_attn_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.13.multihead_attn_audio_self": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.13.multihead_attn_audio_self.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.13.multihead_attn_audio_self.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.13.multihead_attn_audio_self.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.13.multihead_attn_audio_self.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.13.multihead_attn_audio_self.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.13.multihead_attn_audio_self.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.13.multihead_attn_audio_self.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.13.multihead_attn_audio_other": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.13.multihead_attn_audio_other.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.13.multihead_attn_audio_other.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.13.multihead_attn_audio_other.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.13.multihead_attn_audio_other.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.13.multihead_attn_audio_other.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.13.multihead_attn_audio_other.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.13.multihead_attn_audio_other.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.13.linear1": { "type": "Linear" }, "model.seqTransDecoder.stack.13.linear2": { "type": "Linear" }, "model.seqTransDecoder.stack.13.norm1": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.13.norm2": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.13.norm3": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.13.adaln1": { "type": "AdaLN" }, "model.seqTransDecoder.stack.13.adaln1.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.13.adaln1.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.13.adaln1.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.13.adaln2": { "type": "AdaLN" }, "model.seqTransDecoder.stack.13.adaln2.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.13.adaln2.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.13.adaln2.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.13.adaln3": { "type": "AdaLN" }, "model.seqTransDecoder.stack.13.adaln3.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.13.adaln3.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.13.adaln3.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.14": { "type": "TransformerDecoderLayer" }, "model.seqTransDecoder.stack.14.self_attn": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": true }, "model.seqTransDecoder.stack.14.self_attn.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.14.self_attn.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.14.self_attn.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.14.self_attn.to_gate": { "type": "Linear" }, "model.seqTransDecoder.stack.14.self_attn.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.14.self_attn.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.14.self_attn.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.14.kp_cross": { "type": "Sequential" }, "model.seqTransDecoder.stack.14.kp_cross.0": { "type": "GELU" }, "model.seqTransDecoder.stack.14.kp_cross.1": { "type": "Linear" }, "model.seqTransDecoder.stack.14.multihead_attn_last_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.14.multihead_attn_last_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.14.multihead_attn_last_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.14.multihead_attn_last_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.14.multihead_attn_last_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.14.multihead_attn_last_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.14.multihead_attn_last_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.14.multihead_attn_last_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.14.multihead_attn_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.14.multihead_attn_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.14.multihead_attn_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.14.multihead_attn_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.14.multihead_attn_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.14.multihead_attn_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.14.multihead_attn_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.14.multihead_attn_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.14.multihead_attn_audio_self": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.14.multihead_attn_audio_self.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.14.multihead_attn_audio_self.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.14.multihead_attn_audio_self.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.14.multihead_attn_audio_self.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.14.multihead_attn_audio_self.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.14.multihead_attn_audio_self.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.14.multihead_attn_audio_self.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.14.multihead_attn_audio_other": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.14.multihead_attn_audio_other.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.14.multihead_attn_audio_other.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.14.multihead_attn_audio_other.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.14.multihead_attn_audio_other.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.14.multihead_attn_audio_other.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.14.multihead_attn_audio_other.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.14.multihead_attn_audio_other.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.14.linear1": { "type": "Linear" }, "model.seqTransDecoder.stack.14.linear2": { "type": "Linear" }, "model.seqTransDecoder.stack.14.norm1": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.14.norm2": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.14.norm3": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.14.adaln1": { "type": "AdaLN" }, "model.seqTransDecoder.stack.14.adaln1.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.14.adaln1.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.14.adaln1.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.14.adaln2": { "type": "AdaLN" }, "model.seqTransDecoder.stack.14.adaln2.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.14.adaln2.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.14.adaln2.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.14.adaln3": { "type": "AdaLN" }, "model.seqTransDecoder.stack.14.adaln3.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.14.adaln3.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.14.adaln3.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.15": { "type": "TransformerDecoderLayer" }, "model.seqTransDecoder.stack.15.self_attn": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": true }, "model.seqTransDecoder.stack.15.self_attn.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.15.self_attn.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.15.self_attn.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.15.self_attn.to_gate": { "type": "Linear" }, "model.seqTransDecoder.stack.15.self_attn.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.15.self_attn.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.15.self_attn.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.15.kp_cross": { "type": "Sequential" }, "model.seqTransDecoder.stack.15.kp_cross.0": { "type": "GELU" }, "model.seqTransDecoder.stack.15.kp_cross.1": { "type": "Linear" }, "model.seqTransDecoder.stack.15.multihead_attn_last_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.15.multihead_attn_last_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.15.multihead_attn_last_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.15.multihead_attn_last_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.15.multihead_attn_last_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.15.multihead_attn_last_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.15.multihead_attn_last_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.15.multihead_attn_last_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.15.multihead_attn_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.15.multihead_attn_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.15.multihead_attn_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.15.multihead_attn_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.15.multihead_attn_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.15.multihead_attn_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.15.multihead_attn_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.15.multihead_attn_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.15.multihead_attn_audio_self": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.15.multihead_attn_audio_self.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.15.multihead_attn_audio_self.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.15.multihead_attn_audio_self.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.15.multihead_attn_audio_self.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.15.multihead_attn_audio_self.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.15.multihead_attn_audio_self.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.15.multihead_attn_audio_self.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.15.multihead_attn_audio_other": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.15.multihead_attn_audio_other.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.15.multihead_attn_audio_other.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.15.multihead_attn_audio_other.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.15.multihead_attn_audio_other.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.15.multihead_attn_audio_other.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.15.multihead_attn_audio_other.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.15.multihead_attn_audio_other.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.15.linear1": { "type": "Linear" }, "model.seqTransDecoder.stack.15.linear2": { "type": "Linear" }, "model.seqTransDecoder.stack.15.norm1": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.15.norm2": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.15.norm3": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.15.adaln1": { "type": "AdaLN" }, "model.seqTransDecoder.stack.15.adaln1.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.15.adaln1.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.15.adaln1.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.15.adaln2": { "type": "AdaLN" }, "model.seqTransDecoder.stack.15.adaln2.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.15.adaln2.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.15.adaln2.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.15.adaln3": { "type": "AdaLN" }, "model.seqTransDecoder.stack.15.adaln3.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.15.adaln3.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.15.adaln3.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.16": { "type": "TransformerDecoderLayer" }, "model.seqTransDecoder.stack.16.self_attn": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": true }, "model.seqTransDecoder.stack.16.self_attn.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.16.self_attn.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.16.self_attn.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.16.self_attn.to_gate": { "type": "Linear" }, "model.seqTransDecoder.stack.16.self_attn.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.16.self_attn.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.16.self_attn.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.16.kp_cross": { "type": "Sequential" }, "model.seqTransDecoder.stack.16.kp_cross.0": { "type": "GELU" }, "model.seqTransDecoder.stack.16.kp_cross.1": { "type": "Linear" }, "model.seqTransDecoder.stack.16.multihead_attn_last_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.16.multihead_attn_last_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.16.multihead_attn_last_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.16.multihead_attn_last_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.16.multihead_attn_last_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.16.multihead_attn_last_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.16.multihead_attn_last_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.16.multihead_attn_last_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.16.multihead_attn_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.16.multihead_attn_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.16.multihead_attn_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.16.multihead_attn_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.16.multihead_attn_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.16.multihead_attn_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.16.multihead_attn_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.16.multihead_attn_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.16.multihead_attn_audio_self": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.16.multihead_attn_audio_self.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.16.multihead_attn_audio_self.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.16.multihead_attn_audio_self.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.16.multihead_attn_audio_self.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.16.multihead_attn_audio_self.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.16.multihead_attn_audio_self.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.16.multihead_attn_audio_self.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.16.multihead_attn_audio_other": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.16.multihead_attn_audio_other.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.16.multihead_attn_audio_other.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.16.multihead_attn_audio_other.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.16.multihead_attn_audio_other.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.16.multihead_attn_audio_other.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.16.multihead_attn_audio_other.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.16.multihead_attn_audio_other.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.16.linear1": { "type": "Linear" }, "model.seqTransDecoder.stack.16.linear2": { "type": "Linear" }, "model.seqTransDecoder.stack.16.norm1": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.16.norm2": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.16.norm3": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.16.adaln1": { "type": "AdaLN" }, "model.seqTransDecoder.stack.16.adaln1.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.16.adaln1.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.16.adaln1.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.16.adaln2": { "type": "AdaLN" }, "model.seqTransDecoder.stack.16.adaln2.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.16.adaln2.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.16.adaln2.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.16.adaln3": { "type": "AdaLN" }, "model.seqTransDecoder.stack.16.adaln3.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.16.adaln3.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.16.adaln3.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.17": { "type": "TransformerDecoderLayer" }, "model.seqTransDecoder.stack.17.self_attn": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": true }, "model.seqTransDecoder.stack.17.self_attn.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.17.self_attn.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.17.self_attn.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.17.self_attn.to_gate": { "type": "Linear" }, "model.seqTransDecoder.stack.17.self_attn.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.17.self_attn.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.17.self_attn.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.17.kp_cross": { "type": "Sequential" }, "model.seqTransDecoder.stack.17.kp_cross.0": { "type": "GELU" }, "model.seqTransDecoder.stack.17.kp_cross.1": { "type": "Linear" }, "model.seqTransDecoder.stack.17.multihead_attn_last_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.17.multihead_attn_last_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.17.multihead_attn_last_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.17.multihead_attn_last_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.17.multihead_attn_last_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.17.multihead_attn_last_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.17.multihead_attn_last_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.17.multihead_attn_last_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.17.multihead_attn_past": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.17.multihead_attn_past.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.17.multihead_attn_past.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.17.multihead_attn_past.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.17.multihead_attn_past.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.17.multihead_attn_past.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.17.multihead_attn_past.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.17.multihead_attn_past.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.17.multihead_attn_audio_self": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.17.multihead_attn_audio_self.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.17.multihead_attn_audio_self.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.17.multihead_attn_audio_self.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.17.multihead_attn_audio_self.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.17.multihead_attn_audio_self.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.17.multihead_attn_audio_self.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.17.multihead_attn_audio_self.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.17.multihead_attn_audio_other": { "type": "Attention", "d_model": 512, "nhead": 8, "inner_dim": 64, "use_gate": false }, "model.seqTransDecoder.stack.17.multihead_attn_audio_other.to_q": { "type": "Linear" }, "model.seqTransDecoder.stack.17.multihead_attn_audio_other.to_k": { "type": "Linear" }, "model.seqTransDecoder.stack.17.multihead_attn_audio_other.to_v": { "type": "Linear" }, "model.seqTransDecoder.stack.17.multihead_attn_audio_other.to_gate": { "type": "Identity" }, "model.seqTransDecoder.stack.17.multihead_attn_audio_other.norm": { "type": "QKNorm", "eps": 1e-06 }, "model.seqTransDecoder.stack.17.multihead_attn_audio_other.proj": { "type": "Linear" }, "model.seqTransDecoder.stack.17.multihead_attn_audio_other.rotary": { "type": "RoPE" }, "model.seqTransDecoder.stack.17.linear1": { "type": "Linear" }, "model.seqTransDecoder.stack.17.linear2": { "type": "Linear" }, "model.seqTransDecoder.stack.17.norm1": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.17.norm2": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.17.norm3": { "type": "LayerNorm", "eps": 1e-05, "normalized_shape": [ 512 ] }, "model.seqTransDecoder.stack.17.adaln1": { "type": "AdaLN" }, "model.seqTransDecoder.stack.17.adaln1.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.17.adaln1.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.17.adaln1.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.17.adaln2": { "type": "AdaLN" }, "model.seqTransDecoder.stack.17.adaln2.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.17.adaln2.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.17.adaln2.adaln_proj.1": { "type": "Linear" }, "model.seqTransDecoder.stack.17.adaln3": { "type": "AdaLN" }, "model.seqTransDecoder.stack.17.adaln3.adaln_proj": { "type": "Sequential" }, "model.seqTransDecoder.stack.17.adaln3.adaln_proj.0": { "type": "SiLU" }, "model.seqTransDecoder.stack.17.adaln3.adaln_proj.1": { "type": "Linear" }, "model.split_layer": { "type": "VectorizedSplitLayer", "split_sizes": [ 36, 109, 219, 148 ] }, "model.last_layer": { "type": "VectorizedFinalLayer", "split_sizes": [ 36, 109, 219, 148 ] } }