{
  "_class_name": "AutoencoderKL",
  "_diffusers_version": "0.36.0",
  "_name_or_path": "/root/.cache/huggingface/hub/models--cvssp--audioldm-s-full-v2/snapshots/feeb3d14203495a4b6ac0893cbdedb2159b4819c/vae",
  "act_fn": "silu",
  "block_out_channels": [
    128,
    256,
    512
  ],
  "down_block_types": [
    "DownEncoderBlock2D",
    "DownEncoderBlock2D",
    "DownEncoderBlock2D"
  ],
  "force_upcast": true,
  "in_channels": 1,
  "latent_channels": 8,
  "latents_mean": null,
  "latents_std": null,
  "layers_per_block": 2,
  "mid_block_add_attention": true,
  "norm_num_groups": 32,
  "out_channels": 1,
  "sample_size": 512,
  "scaling_factor": 0.9392935633659363,
  "shift_factor": null,
  "up_block_types": [
    "UpDecoderBlock2D",
    "UpDecoderBlock2D",
    "UpDecoderBlock2D"
  ],
  "use_post_quant_conv": true,
  "use_quant_conv": true
}