Make Step Audio runtime bundle self-contained

README.md

@@ -56,8 +56,11 @@ This bundle is self-contained — all weights are packaged in one repository.
 | `config.json` | Step1 LM config + quantization | JSON | — |
 | `flow-model-config.json` | Flow model config | JSON | — |
 | `vq02-config.json`, `vq06-config.json` | Tokenizer configs | JSON | — |
+| `step-audio-tokenizer-assets.safetensors` | VQ02 codebook + CMVN | FP32 | ~2 MB |
+| `step-audio-tokenizer-config.json` | Tokenizer runtime config | JSON | — |
+| `frontend-config.json` | Prompt mel frontend config | JSON | — |
 | `hift-config.json`, `campplus-config.json`, `flow-conditioner-config.json` | Component configs | JSON | — |
-| `tokenizer.json`, `tokenizer.model`, `tokenizer_config.json` | Step1 tokenizer | JSON | — |
+| `tokenizer.json`, `tokenizer_config.json` | Step1 tokenizer | JSON | — |
 
 ## How to Get Started
 

frontend-config.json

@@ -0,0 +1,9 @@
+{
+  "num_mels": 80,
+  "n_fft": 1920,
+  "hop_size": 480,
+  "win_size": 1920,
+  "sampling_rate": 24000,
+  "fmin": 0.0,
+  "fmax": 8000.0
+}

tokenizer.model → step-audio-tokenizer-assets.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:25e122d9205d035033a9994c4d46a6a1b467a938654e4178fc0e5f4f5d610674
-size 1264044
+oid sha256:9eb651ec816b90046dc8c9eaadc1ef0f19f36e29f5e904e616c67b3a9bdcd052
+size 2101864

step-audio-tokenizer-config.json

@@ -0,0 +1,24 @@
+{
+  "model_type": "step_audio_tokenizer",
+  "vq02_sample_rate": 16000,
+  "vq06_sample_rate": 16000,
+  "vq02_codebook_size": 1024,
+  "vq06_token_rate_hz": 25.0,
+  "vq06_n_fft": 400,
+  "vq06_hop_length": 160,
+  "vq06_num_mels": 128,
+  "vq06_max_chunk_seconds": 30.0,
+  "vq06_min_chunk_samples": 480,
+  "trim_top_db": 20.0,
+  "trim_frame_length": 512,
+  "trim_hop_length": 128,
+  "trim_keep_left_seconds": 0.05,
+  "trim_keep_right_seconds": 0.22,
+  "trim_output_hop_samples": 240,
+  "vq02_chunk_size": [
+    0,
+    4,
+    5
+  ],
+  "encoder_chunk_look_back": 4
+}