_target_: gem.network.gem_diffusion.GEMDiffusion
args: ${pipeline.args}
latent_dim: ${.model_cfg.denoiser.latent_dim}
cond_merge_strategy: "add"
music_mask_prob: ${.model_cfg.denoiser.music_mask_prob}
speech_mask_prob: ${.model_cfg.denoiser.speech_mask_prob}
encoded_music_dim: ${pipeline.args.encoded_music_dim}
model_cfg:
  diffusion: ${model_cfg.diffusion}
  denoiser:
    _target_: gem.network.gem_denoiser.NetworkEncoderRoPE
    output_dim: ${endecoder.feat_dim}
    xt_dim: ${.output_dim}
    njoints: ${.xt_dim}
    text_mask_prob: 0.1
    music_mask_prob: 0.1
    speech_mask_prob: 0.1
    use_text_pos_enc: true
    text_encoder_cfg:
      mode: all
      cross_attn_type: mha
    latent_dim: 512
    num_layers: 12
    num_heads: 8
    mlp_ratio: 4