Hugging Face's logo

jasonzhang76
/
Qwen3-TTS-12Hz-0.6B-CustomVoice-ONNX

Text-to-Speech
ONNX
tts
qwen3-tts
Model card Files
xet
 Community
jasonzhang76 elbruno commited on
Commit
4b441a6
·
0 Parent(s):

Duplicate from elbruno/Qwen3-TTS-12Hz-0.6B-CustomVoice-ONNX

Browse files

Co-authored-by: Bruno Capuano <elbruno@users.noreply.huggingface.co>

Files changed (35) hide show
  1. .gitattributes +38 -0
  2. README.md +50 -0
  3. code_predictor.onnx +3 -0
  4. embeddings/codec_head_weight.npy +3 -0
  5. embeddings/config.json +61 -0
  6. embeddings/cp_codec_embedding_0.npy +3 -0
  7. embeddings/cp_codec_embedding_1.npy +3 -0
  8. embeddings/cp_codec_embedding_10.npy +3 -0
  9. embeddings/cp_codec_embedding_11.npy +3 -0
  10. embeddings/cp_codec_embedding_12.npy +3 -0
  11. embeddings/cp_codec_embedding_13.npy +3 -0
  12. embeddings/cp_codec_embedding_14.npy +3 -0
  13. embeddings/cp_codec_embedding_2.npy +3 -0
  14. embeddings/cp_codec_embedding_3.npy +3 -0
  15. embeddings/cp_codec_embedding_4.npy +3 -0
  16. embeddings/cp_codec_embedding_5.npy +3 -0
  17. embeddings/cp_codec_embedding_6.npy +3 -0
  18. embeddings/cp_codec_embedding_7.npy +3 -0
  19. embeddings/cp_codec_embedding_8.npy +3 -0
  20. embeddings/cp_codec_embedding_9.npy +3 -0
  21. embeddings/speaker_ids.json +11 -0
  22. embeddings/talker_codec_embedding.npy +3 -0
  23. embeddings/text_embedding.npy +3 -0
  24. embeddings/text_projection_fc1_bias.npy +3 -0
  25. embeddings/text_projection_fc1_weight.npy +3 -0
  26. embeddings/text_projection_fc2_bias.npy +3 -0
  27. embeddings/text_projection_fc2_weight.npy +3 -0
  28. talker_decode.onnx +3 -0
  29. talker_decode.onnx.data +3 -0
  30. talker_prefill.onnx +3 -0
  31. talker_prefill.onnx.data +3 -0
  32. tokenizer/merges.txt +0 -0
  33. tokenizer/vocab.json +0 -0
  34. vocoder.onnx +3 -0
  35. vocoder.onnx.data +3 -0
.gitattributes ADDED
@@ -0,0 +1,38 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ *.7z filter=lfs diff=lfs merge=lfs -text
2
+ *.arrow filter=lfs diff=lfs merge=lfs -text
3
+ *.bin filter=lfs diff=lfs merge=lfs -text
4
+ *.bz2 filter=lfs diff=lfs merge=lfs -text
5
+ *.ckpt filter=lfs diff=lfs merge=lfs -text
6
+ *.ftz filter=lfs diff=lfs merge=lfs -text
7
+ *.gz filter=lfs diff=lfs merge=lfs -text
8
+ *.h5 filter=lfs diff=lfs merge=lfs -text
9
+ *.joblib filter=lfs diff=lfs merge=lfs -text
10
+ *.lfs.* filter=lfs diff=lfs merge=lfs -text
11
+ *.mlmodel filter=lfs diff=lfs merge=lfs -text
12
+ *.model filter=lfs diff=lfs merge=lfs -text
13
+ *.msgpack filter=lfs diff=lfs merge=lfs -text
14
+ *.npy filter=lfs diff=lfs merge=lfs -text
15
+ *.npz filter=lfs diff=lfs merge=lfs -text
16
+ *.onnx filter=lfs diff=lfs merge=lfs -text
17
+ *.ot filter=lfs diff=lfs merge=lfs -text
18
+ *.parquet filter=lfs diff=lfs merge=lfs -text
19
+ *.pb filter=lfs diff=lfs merge=lfs -text
20
+ *.pickle filter=lfs diff=lfs merge=lfs -text
21
+ *.pkl filter=lfs diff=lfs merge=lfs -text
22
+ *.pt filter=lfs diff=lfs merge=lfs -text
23
+ *.pth filter=lfs diff=lfs merge=lfs -text
24
+ *.rar filter=lfs diff=lfs merge=lfs -text
25
+ *.safetensors filter=lfs diff=lfs merge=lfs -text
26
+ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
27
+ *.tar.* filter=lfs diff=lfs merge=lfs -text
28
+ *.tar filter=lfs diff=lfs merge=lfs -text
29
+ *.tflite filter=lfs diff=lfs merge=lfs -text
30
+ *.tgz filter=lfs diff=lfs merge=lfs -text
31
+ *.wasm filter=lfs diff=lfs merge=lfs -text
32
+ *.xz filter=lfs diff=lfs merge=lfs -text
33
+ *.zip filter=lfs diff=lfs merge=lfs -text
34
+ *.zst filter=lfs diff=lfs merge=lfs -text
35
+ *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ talker_decode.onnx.data filter=lfs diff=lfs merge=lfs -text
37
+ talker_prefill.onnx.data filter=lfs diff=lfs merge=lfs -text
38
+ vocoder.onnx.data filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,50 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ tags:
4
+ - onnx
5
+ - tts
6
+ - qwen3-tts
7
+ - text-to-speech
8
+ base_model: Qwen/Qwen3-TTS-12Hz-0.6B-CustomVoice
9
+ ---
10
+
11
+ # Qwen3-TTS 12Hz 0.6B CustomVoice — ONNX
12
+
13
+ ONNX export of [Qwen/Qwen3-TTS-12Hz-0.6B-CustomVoice](https://huggingface.co/Qwen/Qwen3-TTS-12Hz-0.6B-CustomVoice) for local inference with C# / ONNX Runtime.
14
+
15
+ ## Files
16
+
17
+ | File | Description | Size |
18
+ |------|-------------|------|
19
+ | `talker_prefill.onnx` + `.data` | Talker LM prefill (28 layers) | ~1.7 GB |
20
+ | `talker_decode.onnx` + `.data` | Talker LM single-step decode | ~1.7 GB |
21
+ | `code_predictor.onnx` | Code Predictor (5 layers, 15 groups) | ~420 MB |
22
+ | `vocoder.onnx` + `.data` | Vocoder decoder (24kHz output) | ~437 MB |
23
+ | `embeddings/` | Text/codec embeddings as .npy + config | ~1.4 GB |
24
+ | `tokenizer/` | BPE tokenizer (vocab.json, merges.txt) | ~4 MB |
25
+
26
+ ## Usage with C#
27
+
28
+ ```bash
29
+ # Clone the app repo
30
+ git clone https://github.com/elbruno/qwen-labs-cs.git
31
+ cd qwen-labs-cs
32
+
33
+ # Download models
34
+ python python/download_onnx_models.py --repo-id elbruno/Qwen3-TTS-12Hz-0.6B-CustomVoice-ONNX
35
+
36
+ # Run
37
+ dotnet run --project src/QwenTTS -- --model-dir python/onnx_runtime --text "Hello world" --speaker ryan --language english
38
+ ```
39
+
40
+ ## Architecture
41
+
42
+ - **Talker**: 28 transformer layers, 16 attn heads, 8 KV heads, hidden=1024
43
+ - **Code Predictor**: 5 layers, generates codebook groups 1-15
44
+ - **Vocoder**: RVQ dequantize → transformer → BigVGAN decoder, 12Hz → 24kHz (1920× upsample)
45
+ - **KV Cache**: Decode uses stacked format (num_layers, B, num_kv_heads, T, head_dim)
46
+ - **Speakers**: serena, vivian, uncle_fu, ryan, aiden, ono_anna, sohee, eric, dylan
47
+
48
+ ## License
49
+
50
+ Apache-2.0 (same as base model)
code_predictor.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4e741d1e16ba61ca446b060b16d2da9519b11d18aae5d7bbbfcd273745a38225
3
+ size 440686845
embeddings/codec_head_weight.npy ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b54db90bbad7148ddb49eb9c693d0956055e80bcd7c6f985293d7da050f0adc6
3
+ size 12583040
embeddings/config.json ADDED
@@ -0,0 +1,61 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "talker": {
3
+ "hidden_size": 1024,
4
+ "text_hidden_size": 2048,
5
+ "vocab_size": 3072,
6
+ "num_hidden_layers": 28,
7
+ "num_attention_heads": 16,
8
+ "num_key_value_heads": 8,
9
+ "head_dim": 128,
10
+ "num_code_groups": 16,
11
+ "codec_eos_token_id": 2150,
12
+ "codec_think_id": 2154,
13
+ "codec_nothink_id": 2155,
14
+ "codec_think_bos_id": 2156,
15
+ "codec_think_eos_id": 2157,
16
+ "codec_pad_id": 2148,
17
+ "codec_bos_id": 2149,
18
+ "rope_theta": 1000000
19
+ },
20
+ "code_predictor": {
21
+ "hidden_size": 1024,
22
+ "vocab_size": 2048,
23
+ "num_hidden_layers": 5,
24
+ "num_attention_heads": 16,
25
+ "num_key_value_heads": 8,
26
+ "head_dim": 128,
27
+ "rope_theta": 1000000
28
+ },
29
+ "tts": {
30
+ "tts_bos_token_id": 151672,
31
+ "tts_eos_token_id": 151673,
32
+ "tts_pad_token_id": 151671,
33
+ "im_start_token_id": 151644,
34
+ "im_end_token_id": 151645
35
+ },
36
+ "language_ids": {
37
+ "chinese": 2055,
38
+ "english": 2050,
39
+ "german": 2053,
40
+ "italian": 2070,
41
+ "portuguese": 2071,
42
+ "spanish": 2054,
43
+ "japanese": 2058,
44
+ "korean": 2064,
45
+ "french": 2061,
46
+ "russian": 2069,
47
+ "beijing_dialect": 2074,
48
+ "sichuan_dialect": 2062
49
+ },
50
+ "speaker_dialect": {
51
+ "serena": false,
52
+ "vivian": false,
53
+ "uncle_fu": false,
54
+ "ryan": false,
55
+ "aiden": false,
56
+ "ono_anna": false,
57
+ "sohee": false,
58
+ "eric": "sichuan_dialect",
59
+ "dylan": "beijing_dialect"
60
+ }
61
+ }
embeddings/cp_codec_embedding_0.npy ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9b1bd89a06be7fc6eb07e6b03d6d3fdc4762b5d0e0db124d74c608c0638082e3
3
+ size 8388736
embeddings/cp_codec_embedding_1.npy ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f0d6e187dec6f037980e38911b27016c0531bb34718bb96ccb954366fc495d21
3
+ size 8388736
embeddings/cp_codec_embedding_10.npy ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0e6d1bd968d39d53064bf430a0117cd83d452ee3ecf78dbe2648ad4b783d9091
3
+ size 8388736
embeddings/cp_codec_embedding_11.npy ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c9a1cb9ed2749c9f94fc29a939afdbb8ec341ba3f734c10b507f82a441a7c20e
3
+ size 8388736
embeddings/cp_codec_embedding_12.npy ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4cc774bda9a3b868461a36d08cd92370b71e162589598d1796b66fd1bb0520d1
3
+ size 8388736
embeddings/cp_codec_embedding_13.npy ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bf0454c373ff115501d1642675125146e92d91703442ffdf87a35a7b4fe94919
3
+ size 8388736
embeddings/cp_codec_embedding_14.npy ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:57cb0ce39de77611a7be8f3932debe3a2fa9eba9189e7b3fac890c482ce9a8ba
3
+ size 8388736
embeddings/cp_codec_embedding_2.npy ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5ae15feb83dc7f97e6f939fe98c9dda91c08749c2a69384d71a8414166792825
3
+ size 8388736
embeddings/cp_codec_embedding_3.npy ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2d740ecfaa6591dcf014570a322192dd022f169bcca80849c41801ff496a866a
3
+ size 8388736
embeddings/cp_codec_embedding_4.npy ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3b466f7d3213b7a8b29381ee47338a9ac4de30460ac3d2defdaceaa0bab657e2
3
+ size 8388736
embeddings/cp_codec_embedding_5.npy ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:494fde5327423a7462d5e3063dcd8f2ad87de411fe5093c587cf8e56b3ae53df
3
+ size 8388736
embeddings/cp_codec_embedding_6.npy ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e6d6780db3685caee084edbce3b2517e5fa52b6a077890dff9c14b6638bfb564
3
+ size 8388736
embeddings/cp_codec_embedding_7.npy ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4d10c51b112e7188062a712e0d664a22dd448cea52ed102e094943c871ec3577
3
+ size 8388736
embeddings/cp_codec_embedding_8.npy ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3bc9ba51d40d91be04f478d9330363af3d1ac1218144e7bf0c8b920bd4d1832a
3
+ size 8388736
embeddings/cp_codec_embedding_9.npy ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:41add7203fe62f322fa66ad353d5a91f265413a5c711681769d5979e526ced51
3
+ size 8388736
embeddings/speaker_ids.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "serena": 3066,
3
+ "vivian": 3065,
4
+ "uncle_fu": 3010,
5
+ "ryan": 3061,
6
+ "aiden": 2861,
7
+ "ono_anna": 2873,
8
+ "sohee": 2864,
9
+ "eric": 2875,
10
+ "dylan": 2878
11
+ }
embeddings/talker_codec_embedding.npy ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:65a94ed3c86cb17e40504a581c746b274a4e56ba200c6f99f383a1bb6274cfbd
3
+ size 12583040
embeddings/text_embedding.npy ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:210383e70be9dd7f5debcfcb24acb0de9c111836dd4920bc545b4cd8dd35956b
3
+ size 1244659840
embeddings/text_projection_fc1_bias.npy ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e9a50b7668d089b5fc250f15a61941b97e573e43430587be06c404cbe447c5dc
3
+ size 8320
embeddings/text_projection_fc1_weight.npy ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4b39093c0011eca2a955fcd272a7165271e3a37ff0d24ab51d5cfa60925b77af
3
+ size 16777344
embeddings/text_projection_fc2_bias.npy ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e66b17cef017c40c37970d27a626171d9a75ab4729fdab7843f8d1f9e9ce7963
3
+ size 4224
embeddings/text_projection_fc2_weight.npy ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:104adb52eabb786a4e26696b323e11a92e4b3568849c04960e907ac44d687b4d
3
+ size 8388736
talker_decode.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1a0ba437c0c39a011eddfbb43e2838e8eadfcc46021b2cdd5c7d562597935296
3
+ size 4718468
talker_decode.onnx.data ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:389276b5ef745a38a5d9a5a4e5a39746ed590643f9cd3571eb8d6125c93bfd6e
3
+ size 1774452736
talker_prefill.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cbd0bf1cd0ccc66a68d2238ae64622c6e033b7b50e1a0b7bc64d1758ec94c113
3
+ size 4617430
talker_prefill.onnx.data ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:389276b5ef745a38a5d9a5a4e5a39746ed590643f9cd3571eb8d6125c93bfd6e
3
+ size 1774452736
tokenizer/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
vocoder.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4ee20178c7ab322891ce412d92edcfbade2e5e94a8cffe054e17b760fc764e45
3
+ size 2712067
vocoder.onnx.data ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f4cd93d2b48b833a6aaca7d5a3c95dd99853baba565514cb91777e3ce3c4cc8d
3
+ size 456261632