Add iOS ONNX VoiceDesign model pack

600m-v3-vd/caption_encoder.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c429f84f5beb19e5bc78ed2fadd1d59f22a148525265fc8272858806a2cd9ced
+size 339140739

600m-v3-vd/config.json

@@ -0,0 +1,65 @@
+{
+  "model": {
+    "latent_dim": 32,
+    "latent_patch_size": 1,
+    "model_dim": 1280,
+    "num_layers": 12,
+    "num_heads": 20,
+    "mlp_ratio": 2.875,
+    "text_mlp_ratio": 2.6,
+    "speaker_mlp_ratio": 2.6,
+    "dropout": 0.0,
+    "text_vocab_size": 99574,
+    "text_tokenizer_repo": "llm-jp/llm-jp-3-150m",
+    "text_add_bos": true,
+    "text_dim": 512,
+    "text_layers": 10,
+    "text_heads": 8,
+    "use_caption_condition": true,
+    "use_speaker_condition": true,
+    "caption_vocab_size": 99574,
+    "caption_tokenizer_repo": "llm-jp/llm-jp-3-150m",
+    "caption_add_bos": true,
+    "caption_dim": 512,
+    "caption_layers": 10,
+    "caption_heads": 8,
+    "caption_mlp_ratio": 2.6,
+    "speaker_dim": 768,
+    "speaker_layers": 8,
+    "speaker_heads": 12,
+    "speaker_patch_size": 1,
+    "timestep_embed_dim": 512,
+    "adaln_rank": 192,
+    "norm_eps": 1e-05,
+    "use_duration_predictor": true,
+    "duration_aux_dim": 14,
+    "duration_hidden_dim": 1024,
+    "duration_layers": 3,
+    "duration_dropout": 0.1,
+    "duration_attention_heads": 8,
+    "duration_architecture": "token_sum_dual_adarn_zero_no_aux",
+    "duration_token_init_frames": 9.0,
+    "duration_speaker_fusion": "adarn_zero",
+    "duration_caption_fusion": "adarn_zero",
+    "duration_caption_pooling": "masked_mean"
+  },
+  "inference": {
+    "max_text_len": 256,
+    "max_caption_len": 512
+  },
+  "export": {
+    "seq_len": 128,
+    "text_len": 256,
+    "caption_len": 512,
+    "ref_len": 200,
+    "sample_rate": 48000,
+    "hop_length": 1920,
+    "latent_dim": 32,
+    "num_layers": 12,
+    "has_caption": true,
+    "has_speaker_kv_scale": true,
+    "lora_merged": false,
+    "dacvae_snake_coreml_safe": true,
+    "context_kv_cache": true
+  }
+}

600m-v3-vd/context_kv.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0a24fa7bac2afa2803fa6ff9003a9b8e9d10669c3d586266e87519d3f014d2d2
+size 220370335

600m-v3-vd/dacvae_decode.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8a48e87e347601cb5eb6ba7dfbdc3aafa586e588448980c5678670fd74b25715
+size 261377318

600m-v3-vd/dacvae_encode.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:43f5d220fd3126a39803d3c779953449a566edab555ed76f0ea16dc0c2a013a9
+size 109422044

600m-v3-vd/dit_step.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c784a61a7a0951c65f23c09e1abc38f74b059d0cfa6da4903322c708e36dddce
+size 1243780954

600m-v3-vd/duration.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:952377bbd10623eb7b5d353c2047bbbc4f2ae34c3362f54f02528230419bf375
+size 87156783

600m-v3-vd/speaker_encoder.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8b3412ec1ba05b3f3894e70ce1e670311fe97e16fb759e4443c8f0ec130ba859
+size 242582307

600m-v3-vd/text_encoder.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:67105de59d868cfffbde8c705d4a70c0afa47304584caa62efce9e4906f5a6bb
+size 339121171