Upload 5 files

ace_step/ace_step_v1_dcae_config.json

@@ -0,0 +1,69 @@
+{
+  "_class_name": "AutoencoderDC",
+  "_diffusers_version": "0.32.2",
+  "_name_or_path": "checkpoints/music_dcae_f8c8",
+  "attention_head_dim": 32,
+  "decoder_act_fns": "silu",
+  "decoder_block_out_channels": [
+    128,
+    256,
+    512,
+    1024
+  ],
+  "decoder_block_types": [
+    "ResBlock",
+    "ResBlock",
+    "ResBlock",
+    "EfficientViTBlock"
+  ],
+  "decoder_layers_per_block": [
+    3,
+    3,
+    3,
+    3
+  ],
+  "decoder_norm_types": "rms_norm",
+  "decoder_qkv_multiscales": [
+    [],
+    [],
+    [
+      5
+    ],
+    [
+      5
+    ]
+  ],
+  "downsample_block_type": "Conv",
+  "encoder_block_out_channels": [
+    128,
+    256,
+    512,
+    1024
+  ],
+  "encoder_block_types": [
+    "ResBlock",
+    "ResBlock",
+    "ResBlock",
+    "EfficientViTBlock"
+  ],
+  "encoder_layers_per_block": [
+    2,
+    2,
+    3,
+    3
+  ],
+  "encoder_qkv_multiscales": [
+    [],
+    [],
+    [
+      5
+    ],
+    [
+      5
+    ]
+  ],
+  "in_channels": 2,
+  "latent_channels": 8,
+  "scaling_factor": 0.41407,
+  "upsample_block_type": "interpolate"
+}

ace_step/ace_step_v1_music_dcae_f8c8_bf16.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2b0cb469307ac50659d1880db2a99bae47d0df335cbb36853964662d4b80e8ee
+size 313646516

ace_step/ace_step_v1_music_vocoder_bf16.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c92c9b46e28ab7b37b777780cf4308ad7ddac869636bb77aa61599358c4bc1c0
+size 206350988

ace_step/ace_step_v1_transformer_config.json

@@ -0,0 +1,35 @@
+{
+  "_class_name": "ACEStepTransformer2DModel",
+  "_diffusers_version": "0.32.2",
+  "attention_head_dim": 128,
+  "in_channels": 8,
+  "inner_dim": 2560,
+  "lyric_encoder_vocab_size": 6693,
+  "lyric_hidden_size": 1024,
+  "max_height": 16,
+  "max_position": 32768,
+  "max_width": 32768,
+  "mlp_ratio": 2.5,
+  "num_attention_heads": 20,
+  "num_layers": 24,
+  "out_channels": 8,
+  "patch_size": [
+    16,
+    1
+  ],
+  "rope_theta": 1000000.0,
+  "speaker_embedding_dim": 512,
+  "ssl_encoder_depths": [
+    8,
+    8
+  ],
+  "ssl_latent_dims": [
+    1024,
+    768
+  ],
+  "ssl_names": [
+    "mert",
+    "m-hubert"
+  ],
+  "text_embedding_dim": 768
+}

ace_step/ace_step_v1_vocoder_config.json

@@ -0,0 +1,79 @@
+{
+  "_class_name": "ADaMoSHiFiGANV1",
+  "_diffusers_version": "0.32.2",
+  "depths": [
+    3,
+    3,
+    9,
+    3
+  ],
+  "dims": [
+    128,
+    256,
+    384,
+    512
+  ],
+  "drop_path_rate": 0.0,
+  "f_max": 16000,
+  "f_min": 40,
+  "hop_length": 512,
+  "input_channels": 128,
+  "kernel_sizes": [
+    7
+  ],
+  "n_fft": 2048,
+  "n_mels": 128,
+  "num_mels": 512,
+  "post_conv_kernel_size": 13,
+  "pre_conv_kernel_size": 13,
+  "resblock_dilation_sizes": [
+    [
+      1,
+      3,
+      5
+    ],
+    [
+      1,
+      3,
+      5
+    ],
+    [
+      1,
+      3,
+      5
+    ],
+    [
+      1,
+      3,
+      5
+    ]
+  ],
+  "resblock_kernel_sizes": [
+    3,
+    7,
+    11,
+    13
+  ],
+  "sampling_rate": 44100,
+  "upsample_initial_channel": 1024,
+  "upsample_kernel_sizes": [
+    8,
+    8,
+    4,
+    4,
+    4,
+    4,
+    4
+  ],
+  "upsample_rates": [
+    4,
+    4,
+    2,
+    2,
+    2,
+    2,
+    2
+  ],
+  "use_template": false,
+  "win_length": 2048
+}