voxcpm2 bf16 export

Browse files

Files changed (8) hide show

README.md +84 -0
config.json +216 -0
model-00001.safetensors +3 -0
model-00002.safetensors +3 -0
model.safetensors.index.json +827 -0
special_tokens_map.json +81 -0
tokenizer.json +0 -0
tokenizer_config.json +212 -0

README.md ADDED Viewed

	@@ -0,0 +1,84 @@

+---
+license: apache-2.0
+base_model: openbmb/VoxCPM2
+tags:
+  - mlx
+  - tts
+  - text-to-speech
+  - voice-cloning
+  - voice-design
+  - multilingual
+library_name: mlx-audio
+pipeline_tag: text-to-speech
+language:
+  - en
+  - zh
+  - id
+  - ja
+  - ko
+  - multilingual
+---
+# VoxCPM2 - BFloat16 (full precision)
+MLX port of [openbmb/VoxCPM2](https://huggingface.co/openbmb/VoxCPM2) — a 2B-parameter multilingual TTS model with 48kHz studio-quality output, voice cloning, and voice design.
+Full precision BFloat16 weights. Best quality, largest size.
+## Features
+- **30 languages** — including English, Chinese, Indonesian, Japanese, Korean, and more
+- **48kHz output** — studio-quality audio
+- **Voice Design** — create voices from text descriptions (no reference audio needed)
+- **Voice Cloning** — clone any voice from a short audio reference
+- **4 generation modes** — zero-shot, continuation, reference cloning, combined
+## Usage
+```bash
+pip install mlx-audio
+# Zero-shot
+python -m mlx_audio.tts.generate --model mlx-community/VoxCPM2-bf16 --text "Hello world" --verbose
+# Voice design
+python -m mlx_audio.tts.generate --model mlx-community/VoxCPM2-bf16 \
+  --text "Hello world" \
+  --instruct "A young woman, gentle and sweet voice"
+# Voice cloning
+python -m mlx_audio.tts.generate --model mlx-community/VoxCPM2-bf16 \
+  --text "Hello world" \
+  --ref_audio speaker.wav --ref_text "reference text"
+```
+### Python API
+```python
+from mlx_audio.tts import load_model
+model = load_model("mlx-community/VoxCPM2-bf16")
+# Generate
+for result in model.generate(
+    text="Hello, this is VoxCPM2 on Apple Silicon.",
+    inference_timesteps=7,
+    cfg_value=2.0,
+):
+    print(f"Duration: {result.audio_duration}")
+```
+## Performance (Apple Silicon)
+| Variant | Size | RTF (7 timesteps) |
+|---------|------|--------------------|
+| bf16 | 4.96 GB | 0.48x |
+| **8-bit** | **3.23 GB** | **0.85x** |
+| **4-bit** | **2.30 GB** | **0.90x** |
+*RTF = Real-Time Factor (>1.0 = faster than realtime)*
+## Original Model
+- [openbmb/VoxCPM2](https://huggingface.co/openbmb/VoxCPM2)
+- Apache 2.0 License
+Converted with [mlx-audio](https://github.com/Blaizzy/mlx-audio).

config.json ADDED Viewed

	@@ -0,0 +1,216 @@

+{
+  "architecture": "voxcpm2",
+  "lm_config": {
+    "bos_token_id": 1,
+    "eos_token_id": 2,
+    "hidden_size": 2048,
+    "intermediate_size": 6144,
+    "max_position_embeddings": 32768,
+    "num_attention_heads": 16,
+    "num_hidden_layers": 28,
+    "num_key_value_heads": 2,
+    "rms_norm_eps": 1e-05,
+    "rope_theta": 10000,
+    "kv_channels": 128,
+    "rope_scaling": {
+      "type": "longrope",
+      "long_factor": [
+        0.9977997200264581,
+        1.014658295992452,
+        1.0349680404997148,
+        1.059429246056193,
+        1.0888815016813513,
+        1.1243301355211495,
+        1.166977103606075,
+        1.2182568066927284,
+        1.2798772354275727,
+        1.3538666751582975,
+        1.4426259039919596,
+        1.5489853358570191,
+        1.6762658237220625,
+        1.8283407612492941,
+        2.0096956085876183,
+        2.225478927469756,
+        2.481536379650452,
+        2.784415934557119,
+        3.1413289096347365,
+        3.560047844772632,
+        4.048719380066383,
+        4.615569542115128,
+        5.2684819496549835,
+        6.014438591970396,
+        6.858830049237097,
+        7.804668263503327,
+        8.851768731513417,
+        9.99600492938444,
+        11.228766118181639,
+        12.536757560834843,
+        13.902257701387796,
+        15.303885189125953,
+        16.717837610115794,
+        18.119465097853947,
+        19.484965238406907,
+        20.792956681060105,
+        22.02571786985731,
+        23.16995406772833,
+        24.217054535738416,
+        25.16289275000465,
+        26.007284207271347,
+        26.753240849586767,
+        27.40615325712662,
+        27.973003419175363,
+        28.461674954469114,
+        28.880393889607006,
+        29.237306864684626,
+        29.540186419591297,
+        29.79624387177199,
+        30.01202719065413,
+        30.193382037992453,
+        30.34545697551969,
+        30.47273746338473,
+        30.579096895249787,
+        30.66785612408345,
+        30.741845563814174,
+        30.80346599254902,
+        30.85474569563567,
+        30.897392663720595,
+        30.932841297560394,
+        30.962293553185553,
+        30.986754758742034,
+        31.007064503249293,
+        31.02392307921529
+      ],
+      "short_factor": [
+        0.9977997200264581,
+        1.014658295992452,
+        1.0349680404997148,
+        1.059429246056193,
+        1.0888815016813513,
+        1.1243301355211495,
+        1.166977103606075,
+        1.2182568066927284,
+        1.2798772354275727,
+        1.3538666751582975,
+        1.4426259039919596,
+        1.5489853358570191,
+        1.6762658237220625,
+        1.8283407612492941,
+        2.0096956085876183,
+        2.225478927469756,
+        2.481536379650452,
+        2.784415934557119,
+        3.1413289096347365,
+        3.560047844772632,
+        4.048719380066383,
+        4.615569542115128,
+        5.2684819496549835,
+        6.014438591970396,
+        6.858830049237097,
+        7.804668263503327,
+        8.851768731513417,
+        9.99600492938444,
+        11.228766118181639,
+        12.536757560834843,
+        13.902257701387796,
+        15.303885189125953,
+        16.717837610115794,
+        18.119465097853947,
+        19.484965238406907,
+        20.792956681060105,
+        22.02571786985731,
+        23.16995406772833,
+        24.217054535738416,
+        25.16289275000465,
+        26.007284207271347,
+        26.753240849586767,
+        27.40615325712662,
+        27.973003419175363,
+        28.461674954469114,
+        28.880393889607006,
+        29.237306864684626,
+        29.540186419591297,
+        29.79624387177199,
+        30.01202719065413,
+        30.193382037992453,
+        30.34545697551969,
+        30.47273746338473,
+        30.579096895249787,
+        30.66785612408345,
+        30.741845563814174,
+        30.80346599254902,
+        30.85474569563567,
+        30.897392663720595,
+        30.932841297560394,
+        30.962293553185553,
+        30.986754758742034,
+        31.007064503249293,
+        31.02392307921529
+      ],
+      "original_max_position_embeddings": 32768
+    },
+    "vocab_size": 73448,
+    "use_mup": false,
+    "scale_emb": 12,
+    "dim_model_base": 256,
+    "scale_depth": 1.4
+  },
+  "patch_size": 4,
+  "feat_dim": 64,
+  "scalar_quantization_latent_dim": 512,
+  "scalar_quantization_scale": 9,
+  "residual_lm_num_layers": 8,
+  "residual_lm_no_rope": true,
+  "encoder_config": {
+    "hidden_dim": 1024,
+    "ffn_dim": 4096,
+    "num_heads": 16,
+    "num_layers": 12,
+    "kv_channels": 128
+  },
+  "dit_config": {
+    "hidden_dim": 1024,
+    "ffn_dim": 4096,
+    "num_heads": 16,
+    "num_layers": 12,
+    "kv_channels": 128,
+    "mean_mode": false,
+    "cfm_config": {
+      "sigma_min": 1e-06,
+      "solver": "euler",
+      "t_scheduler": "log-norm",
+      "inference_cfg_rate": 2.0
+    }
+  },
+  "audio_vae_config": {
+    "encoder_dim": 128,
+    "encoder_rates": [
+      2,
+      5,
+      8,
+      8
+    ],
+    "latent_dim": 64,
+    "decoder_dim": 2048,
+    "decoder_rates": [
+      8,
+      6,
+      5,
+      2,
+      2,
+      2
+    ],
+    "sr_bin_boundaries": [
+      20000,
+      30000,
+      40000
+    ],
+    "sample_rate": 16000,
+    "out_sample_rate": 48000
+  },
+  "max_length": 8192,
+  "model_type": "voxcpm2",
+  "quantization": {
+    "bits": 16,
+    "variant": "bf16"
+  }
+}

model-00001.safetensors ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:676b05649d915e63f2291a0b091bd72250785b2448fe14ec81f3566cf5f47950
+size 4286190791

model-00002.safetensors ADDED Viewed

	@@ -0,0 +1,3 @@

+version https://git-lfs.github.com/spec/v1
+oid sha256:adddb9b85026763751beb20ba050aa46ef19b6b55cf6cb6a8c44ac8b06ec5cff
+size 669068875

model.safetensors.index.json ADDED Viewed

	@@ -0,0 +1,827 @@

+{
+  "metadata": {
+    "total_size": 4955158672
+  },
+  "weight_map": {
+    "audio_vae.decoder._sr_boundaries": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.0.conv_t.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.0.conv_t.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.0.res1.conv1.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.0.res1.conv1.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.0.res1.conv2.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.0.res1.conv2.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.0.res1.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.0.res1.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.0.res2.conv1.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.0.res2.conv1.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.0.res2.conv2.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.0.res2.conv2.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.0.res2.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.0.res2.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.0.res3.conv1.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.0.res3.conv1.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.0.res3.conv2.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.0.res3.conv2.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.0.res3.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.0.res3.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.0.snake.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.1.conv_t.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.1.conv_t.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.1.res1.conv1.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.1.res1.conv1.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.1.res1.conv2.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.1.res1.conv2.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.1.res1.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.1.res1.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.1.res2.conv1.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.1.res2.conv1.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.1.res2.conv2.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.1.res2.conv2.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.1.res2.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.1.res2.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.1.res3.conv1.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.1.res3.conv1.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.1.res3.conv2.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.1.res3.conv2.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.1.res3.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.1.res3.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.1.snake.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.2.conv_t.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.2.conv_t.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.2.res1.conv1.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.2.res1.conv1.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.2.res1.conv2.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.2.res1.conv2.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.2.res1.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.2.res1.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.2.res2.conv1.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.2.res2.conv1.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.2.res2.conv2.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.2.res2.conv2.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.2.res2.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.2.res2.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.2.res3.conv1.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.2.res3.conv1.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.2.res3.conv2.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.2.res3.conv2.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.2.res3.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.2.res3.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.2.snake.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.3.conv_t.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.3.conv_t.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.3.res1.conv1.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.3.res1.conv1.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.3.res1.conv2.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.3.res1.conv2.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.3.res1.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.3.res1.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.3.res2.conv1.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.3.res2.conv1.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.3.res2.conv2.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.3.res2.conv2.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.3.res2.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.3.res2.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.3.res3.conv1.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.3.res3.conv1.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.3.res3.conv2.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.3.res3.conv2.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.3.res3.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.3.res3.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.3.snake.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.4.conv_t.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.4.conv_t.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.4.res1.conv1.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.4.res1.conv1.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.4.res1.conv2.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.4.res1.conv2.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.4.res1.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.4.res1.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.4.res2.conv1.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.4.res2.conv1.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.4.res2.conv2.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.4.res2.conv2.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.4.res2.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.4.res2.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.4.res3.conv1.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.4.res3.conv1.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.4.res3.conv2.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.4.res3.conv2.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.4.res3.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.4.res3.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.4.snake.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.5.conv_t.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.5.conv_t.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.5.res1.conv1.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.5.res1.conv1.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.5.res1.conv2.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.5.res1.conv2.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.5.res1.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.5.res1.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.5.res2.conv1.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.5.res2.conv1.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.5.res2.conv2.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.5.res2.conv2.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.5.res2.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.5.res2.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.5.res3.conv1.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.5.res3.conv1.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.5.res3.conv2.bias": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.5.res3.conv2.weight": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.5.res3.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.5.res3.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.blocks.layers.5.snake.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.conv_in.layers.0.bias": "model-00001.safetensors",
+    "audio_vae.decoder.conv_in.layers.0.weight": "model-00001.safetensors",
+    "audio_vae.decoder.conv_in.layers.1.bias": "model-00001.safetensors",
+    "audio_vae.decoder.conv_in.layers.1.weight": "model-00001.safetensors",
+    "audio_vae.decoder.conv_out.bias": "model-00001.safetensors",
+    "audio_vae.decoder.conv_out.weight": "model-00001.safetensors",
+    "audio_vae.decoder.snake_out.alpha": "model-00001.safetensors",
+    "audio_vae.decoder.sr_cond_layers.0.bias_embed.weight": "model-00001.safetensors",
+    "audio_vae.decoder.sr_cond_layers.0.scale_embed.weight": "model-00001.safetensors",
+    "audio_vae.decoder.sr_cond_layers.1.bias_embed.weight": "model-00001.safetensors",
+    "audio_vae.decoder.sr_cond_layers.1.scale_embed.weight": "model-00001.safetensors",
+    "audio_vae.decoder.sr_cond_layers.2.bias_embed.weight": "model-00001.safetensors",
+    "audio_vae.decoder.sr_cond_layers.2.scale_embed.weight": "model-00001.safetensors",
+    "audio_vae.decoder.sr_cond_layers.3.bias_embed.weight": "model-00001.safetensors",
+    "audio_vae.decoder.sr_cond_layers.3.scale_embed.weight": "model-00001.safetensors",
+    "audio_vae.decoder.sr_cond_layers.4.bias_embed.weight": "model-00001.safetensors",
+    "audio_vae.decoder.sr_cond_layers.4.scale_embed.weight": "model-00001.safetensors",
+    "audio_vae.decoder.sr_cond_layers.5.bias_embed.weight": "model-00001.safetensors",
+    "audio_vae.decoder.sr_cond_layers.5.scale_embed.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.0.conv.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.0.conv.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.0.res1.conv1.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.0.res1.conv1.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.0.res1.conv2.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.0.res1.conv2.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.0.res1.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.0.res1.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.0.res2.conv1.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.0.res2.conv1.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.0.res2.conv2.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.0.res2.conv2.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.0.res2.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.0.res2.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.0.res3.conv1.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.0.res3.conv1.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.0.res3.conv2.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.0.res3.conv2.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.0.res3.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.0.res3.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.0.snake.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.1.conv.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.1.conv.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.1.res1.conv1.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.1.res1.conv1.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.1.res1.conv2.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.1.res1.conv2.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.1.res1.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.1.res1.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.1.res2.conv1.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.1.res2.conv1.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.1.res2.conv2.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.1.res2.conv2.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.1.res2.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.1.res2.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.1.res3.conv1.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.1.res3.conv1.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.1.res3.conv2.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.1.res3.conv2.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.1.res3.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.1.res3.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.1.snake.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.2.conv.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.2.conv.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.2.res1.conv1.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.2.res1.conv1.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.2.res1.conv2.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.2.res1.conv2.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.2.res1.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.2.res1.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.2.res2.conv1.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.2.res2.conv1.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.2.res2.conv2.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.2.res2.conv2.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.2.res2.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.2.res2.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.2.res3.conv1.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.2.res3.conv1.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.2.res3.conv2.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.2.res3.conv2.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.2.res3.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.2.res3.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.2.snake.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.3.conv.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.3.conv.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.3.res1.conv1.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.3.res1.conv1.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.3.res1.conv2.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.3.res1.conv2.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.3.res1.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.3.res1.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.3.res2.conv1.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.3.res2.conv1.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.3.res2.conv2.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.3.res2.conv2.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.3.res2.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.3.res2.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.3.res3.conv1.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.3.res3.conv1.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.3.res3.conv2.bias": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.3.res3.conv2.weight": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.3.res3.snake1.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.3.res3.snake2.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.blocks.layers.3.snake.alpha": "model-00001.safetensors",
+    "audio_vae.encoder.conv_in.bias": "model-00001.safetensors",
+    "audio_vae.encoder.conv_in.weight": "model-00001.safetensors",
+    "audio_vae.encoder.fc_mu.bias": "model-00001.safetensors",
+    "audio_vae.encoder.fc_mu.weight": "model-00001.safetensors",
+    "base_lm.embed_tokens.weight": "model-00001.safetensors",
+    "base_lm.layers.0.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.0.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.0.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.0.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.0.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.0.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.0.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.0.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.0.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.1.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.1.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.1.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.1.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.1.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.1.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.1.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.1.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.1.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.10.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.10.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.10.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.10.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.10.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.10.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.10.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.10.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.10.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.11.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.11.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.11.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.11.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.11.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.11.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.11.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.11.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.11.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.12.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.12.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.12.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.12.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.12.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.12.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.12.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.12.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.12.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.13.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.13.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.13.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.13.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.13.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.13.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.13.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.13.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.13.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.14.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.14.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.14.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.14.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.14.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.14.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.14.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.14.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.14.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.15.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.15.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.15.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.15.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.15.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.15.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.15.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.15.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.15.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.16.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.16.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.16.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.16.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.16.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.16.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.16.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.16.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.16.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.17.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.17.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.17.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.17.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.17.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.17.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.17.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.17.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.17.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.18.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.18.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.18.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.18.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.18.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.18.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.18.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.18.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.18.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.19.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.19.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.19.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.19.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.19.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.19.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.19.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.19.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.19.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.2.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.2.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.2.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.2.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.2.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.2.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.2.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.2.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.2.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.20.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.20.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.20.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.20.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.20.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.20.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.20.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.20.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.20.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.21.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.21.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.21.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.21.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.21.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.21.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.21.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.21.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.21.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.22.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.22.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.22.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.22.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.22.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.22.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.22.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.22.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.22.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.23.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.23.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.23.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.23.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.23.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.23.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.23.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.23.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.23.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.24.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.24.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.24.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.24.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.24.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.24.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.24.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.24.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.24.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.25.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.25.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.25.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.25.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.25.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.25.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.25.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.25.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.25.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.26.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.26.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.26.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.26.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.26.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.26.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.26.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.26.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.26.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.27.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.27.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.27.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.27.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.27.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.27.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.27.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.27.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.27.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.3.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.3.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.3.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.3.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.3.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.3.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.3.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.3.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.3.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.4.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.4.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.4.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.4.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.4.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.4.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.4.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.4.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.4.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.5.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.5.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.5.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.5.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.5.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.5.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.5.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.5.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.5.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.6.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.6.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.6.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.6.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.6.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.6.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.6.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.6.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.6.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.7.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.7.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.7.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.7.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.7.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.7.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.7.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.7.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.7.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.8.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.8.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.8.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.8.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.8.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.8.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.8.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.8.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.8.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.9.input_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.9.mlp.down_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.9.mlp.gate_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.9.mlp.up_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.9.post_attention_layernorm.weight": "model-00001.safetensors",
+    "base_lm.layers.9.self_attn.k_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.9.self_attn.o_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.9.self_attn.q_proj.weight": "model-00001.safetensors",
+    "base_lm.layers.9.self_attn.v_proj.weight": "model-00001.safetensors",
+    "base_lm.norm.weight": "model-00001.safetensors",
+    "base_lm.rope.inv_freq": "model-00001.safetensors",
+    "base_lm.rope.long_factor": "model-00001.safetensors",
+    "base_lm.rope.short_factor": "model-00001.safetensors",
+    "enc_to_lm_proj.bias": "model-00001.safetensors",
+    "enc_to_lm_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.cond_proj.bias": "model-00001.safetensors",
+    "feat_decoder.estimator.cond_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.0.input_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.0.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.0.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.0.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.0.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.0.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.0.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.0.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.0.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.1.input_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.1.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.1.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.1.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.1.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.1.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.1.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.1.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.1.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.10.input_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.10.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.10.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.10.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.10.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.10.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.10.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.10.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.10.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.11.input_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.11.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.11.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.11.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.11.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.11.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.11.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.11.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.11.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.2.input_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.2.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.2.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.2.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.2.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.2.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.2.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.2.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.2.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.3.input_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.3.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.3.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.3.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.3.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.3.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.3.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.3.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.3.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.4.input_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.4.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.4.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.4.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.4.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.4.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.4.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.4.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.4.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.5.input_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.5.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.5.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.5.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.5.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.5.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.5.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.5.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.5.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.6.input_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.6.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.6.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.6.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.6.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.6.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.6.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.6.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.6.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.7.input_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.7.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.7.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.7.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.7.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.7.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.7.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.7.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.7.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.8.input_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.8.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.8.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.8.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.8.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.8.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.8.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.8.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.8.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.9.input_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.9.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.9.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.9.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.9.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.9.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.9.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.9.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.layers.9.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.norm.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.rope.inv_freq": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.rope.long_factor": "model-00001.safetensors",
+    "feat_decoder.estimator.decoder.rope.short_factor": "model-00001.safetensors",
+    "feat_decoder.estimator.delta_time_mlp.linear_1.bias": "model-00001.safetensors",
+    "feat_decoder.estimator.delta_time_mlp.linear_1.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.delta_time_mlp.linear_2.bias": "model-00001.safetensors",
+    "feat_decoder.estimator.delta_time_mlp.linear_2.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.in_proj.bias": "model-00001.safetensors",
+    "feat_decoder.estimator.in_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.out_proj.bias": "model-00001.safetensors",
+    "feat_decoder.estimator.out_proj.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.time_mlp.linear_1.bias": "model-00001.safetensors",
+    "feat_decoder.estimator.time_mlp.linear_1.weight": "model-00001.safetensors",
+    "feat_decoder.estimator.time_mlp.linear_2.bias": "model-00001.safetensors",
+    "feat_decoder.estimator.time_mlp.linear_2.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.0.input_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.0.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.0.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.0.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.0.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.0.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.0.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.0.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.0.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.1.input_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.1.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.1.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.1.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.1.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.1.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.1.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.1.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.1.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.10.input_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.10.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.10.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.10.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.10.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.10.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.10.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.10.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.10.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.11.input_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.11.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.11.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.11.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.11.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.11.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.11.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.11.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.11.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.2.input_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.2.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.2.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.2.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.2.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.2.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.2.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.2.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.2.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.3.input_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.3.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.3.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.3.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.3.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.3.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.3.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.3.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.3.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.4.input_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.4.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.4.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.4.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.4.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.4.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.4.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.4.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.4.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.5.input_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.5.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.5.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.5.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.5.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.5.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.5.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.5.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.5.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.6.input_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.6.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.6.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.6.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.6.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.6.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.6.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.6.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.6.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.7.input_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.7.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.7.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.7.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.7.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.7.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.7.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.7.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.7.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.8.input_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.8.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.8.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.8.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.8.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.8.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.8.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.8.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.8.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.9.input_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.9.mlp.down_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.9.mlp.gate_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.9.mlp.up_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.9.post_attention_layernorm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.9.self_attn.k_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.9.self_attn.o_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.9.self_attn.q_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.layers.9.self_attn.v_proj.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.norm.weight": "model-00001.safetensors",
+    "feat_encoder.encoder.rope.inv_freq": "model-00001.safetensors",
+    "feat_encoder.encoder.rope.long_factor": "model-00001.safetensors",
+    "feat_encoder.encoder.rope.short_factor": "model-00001.safetensors",
+    "feat_encoder.in_proj.bias": "model-00001.safetensors",
+    "feat_encoder.in_proj.weight": "model-00001.safetensors",
+    "feat_encoder.special_token": "model-00001.safetensors",
+    "fsq_layer.in_proj.bias": "model-00001.safetensors",
+    "fsq_layer.in_proj.weight": "model-00001.safetensors",
+    "fsq_layer.out_proj.bias": "model-00001.safetensors",
+    "fsq_layer.out_proj.weight": "model-00001.safetensors",
+    "fusion_concat_proj.bias": "model-00001.safetensors",
+    "fusion_concat_proj.weight": "model-00001.safetensors",
+    "lm_to_dit_proj.bias": "model-00001.safetensors",
+    "lm_to_dit_proj.weight": "model-00001.safetensors",
+    "res_to_dit_proj.bias": "model-00001.safetensors",
+    "res_to_dit_proj.weight": "model-00001.safetensors",
+    "residual_lm.layers.0.input_layernorm.weight": "model-00001.safetensors",
+    "residual_lm.layers.0.mlp.down_proj.weight": "model-00001.safetensors",
+    "residual_lm.layers.0.mlp.gate_proj.weight": "model-00001.safetensors",
+    "residual_lm.layers.0.mlp.up_proj.weight": "model-00001.safetensors",
+    "residual_lm.layers.0.post_attention_layernorm.weight": "model-00001.safetensors",
+    "residual_lm.layers.0.self_attn.k_proj.weight": "model-00001.safetensors",
+    "residual_lm.layers.0.self_attn.o_proj.weight": "model-00001.safetensors",
+    "residual_lm.layers.0.self_attn.q_proj.weight": "model-00001.safetensors",
+    "residual_lm.layers.0.self_attn.v_proj.weight": "model-00001.safetensors",
+    "residual_lm.layers.1.input_layernorm.weight": "model-00001.safetensors",
+    "residual_lm.layers.1.mlp.down_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.1.mlp.gate_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.1.mlp.up_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.1.post_attention_layernorm.weight": "model-00002.safetensors",
+    "residual_lm.layers.1.self_attn.k_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.1.self_attn.o_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.1.self_attn.q_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.1.self_attn.v_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.2.input_layernorm.weight": "model-00002.safetensors",
+    "residual_lm.layers.2.mlp.down_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.2.mlp.gate_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.2.mlp.up_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.2.post_attention_layernorm.weight": "model-00002.safetensors",
+    "residual_lm.layers.2.self_attn.k_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.2.self_attn.o_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.2.self_attn.q_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.2.self_attn.v_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.3.input_layernorm.weight": "model-00002.safetensors",
+    "residual_lm.layers.3.mlp.down_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.3.mlp.gate_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.3.mlp.up_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.3.post_attention_layernorm.weight": "model-00002.safetensors",
+    "residual_lm.layers.3.self_attn.k_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.3.self_attn.o_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.3.self_attn.q_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.3.self_attn.v_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.4.input_layernorm.weight": "model-00002.safetensors",
+    "residual_lm.layers.4.mlp.down_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.4.mlp.gate_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.4.mlp.up_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.4.post_attention_layernorm.weight": "model-00002.safetensors",
+    "residual_lm.layers.4.self_attn.k_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.4.self_attn.o_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.4.self_attn.q_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.4.self_attn.v_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.5.input_layernorm.weight": "model-00002.safetensors",
+    "residual_lm.layers.5.mlp.down_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.5.mlp.gate_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.5.mlp.up_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.5.post_attention_layernorm.weight": "model-00002.safetensors",
+    "residual_lm.layers.5.self_attn.k_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.5.self_attn.o_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.5.self_attn.q_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.5.self_attn.v_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.6.input_layernorm.weight": "model-00002.safetensors",
+    "residual_lm.layers.6.mlp.down_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.6.mlp.gate_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.6.mlp.up_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.6.post_attention_layernorm.weight": "model-00002.safetensors",
+    "residual_lm.layers.6.self_attn.k_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.6.self_attn.o_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.6.self_attn.q_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.6.self_attn.v_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.7.input_layernorm.weight": "model-00002.safetensors",
+    "residual_lm.layers.7.mlp.down_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.7.mlp.gate_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.7.mlp.up_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.7.post_attention_layernorm.weight": "model-00002.safetensors",
+    "residual_lm.layers.7.self_attn.k_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.7.self_attn.o_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.7.self_attn.q_proj.weight": "model-00002.safetensors",
+    "residual_lm.layers.7.self_attn.v_proj.weight": "model-00002.safetensors",
+    "residual_lm.norm.weight": "model-00002.safetensors",
+    "stop_head.weight": "model-00002.safetensors",
+    "stop_proj.bias": "model-00002.safetensors",
+    "stop_proj.weight": "model-00002.safetensors"
+  }
+}

special_tokens_map.json ADDED Viewed

	@@ -0,0 +1,81 @@

+{
+  "additional_special_tokens": [
+    {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<|tool_call|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<|execute_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<|execute_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    },
+    {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false
+    }
+  ],
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.json ADDED Viewed

The diff for this file is too large to render. See raw diff

tokenizer_config.json ADDED Viewed

	@@ -0,0 +1,212 @@

+{
+  "add_bos_token": true,
+  "add_eos_token": false,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "101": {
+      "content": "<|audio_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "102": {
+      "content": "<|audio_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "103": {
+      "content": "<|audio_prompt_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "104": {
+      "content": "<|audio_prompt_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "105": {
+      "content": "<|background|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "106": {
+      "content": "<|/background|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "107": {
+      "content": "<|characters|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "108": {
+      "content": "<|/characters|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "109": {
+      "content": "<|speaker_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "110": {
+      "content": "<|/speaker_id|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "111": {
+      "content": "<|span|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "112": {
+      "content": "<|/span|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "73440": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "73441": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "73442": {
+      "content": "<|tool_call|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "73443": {
+      "content": "<|execute_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "73444": {
+      "content": "<|execute_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "73445": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "73446": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "73447": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_end|>",
+    "<|im_start|>",
+    "<|tool_call|>",
+    "<|execute_start|>",
+    "<|execute_end|>",
+    "<|fim_prefix|>",
+    "<|fim_middle|>",
+    "<|fim_suffix|>"
+  ],
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "legacy": true,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": null,
+  "sp_model_kwargs": {},
+  "spaces_between_special_tokens": false,
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": false,
+  "chat_template": "{% for message in messages %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}"
+}