k-l-lambda commited on
Commit
9c671dd
·
verified ·
1 Parent(s): 3bebdca

Add five Chinese and English poetry speech recordings

Browse files
.gitattributes CHANGED
@@ -33,3 +33,8 @@ saved_model/**/* 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
 
 
 
 
 
 
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
+ examples/poetry/quiet-night.wav filter=lfs diff=lfs merge=lfs -text
37
+ examples/poetry/sonnet-18.wav filter=lfs diff=lfs merge=lfs -text
38
+ examples/poetry/spring-dawn.wav filter=lfs diff=lfs merge=lfs -text
39
+ examples/poetry/stork-tower.wav filter=lfs diff=lfs merge=lfs -text
40
+ examples/poetry/the-tyger.wav filter=lfs diff=lfs merge=lfs -text
examples/poetry/README.md ADDED
@@ -0,0 +1,51 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Poetry voice examples
2
+
3
+ Five public-domain poetry readings synthesized with Qwen3-TTS 1.7B VoiceDesign. Chinese poems are complete; English readings are excerpts of the original English poems. No modern translations are used.
4
+
5
+ Select a poem in the application to load its recording into **Audio → MIDI**. You can listen immediately and then generate piano MIDI. These assets contain speech, not precomputed piano reconstructions.
6
+
7
+ | Poem | Author | Language | Duration | Audio |
8
+ |---|---|---|---:|---|
9
+ | 静夜思 | 李白 | Chinese | 8.64 s | [Listen](quiet-night.wav) |
10
+ | 春晓 | 孟浩然 | Chinese | 8.16 s | [Listen](spring-dawn.wav) |
11
+ | 登鹳雀楼 | 王之涣 | Chinese | 7.60 s | [Listen](stork-tower.wav) |
12
+ | Sonnet 18 (opening couplet) | William Shakespeare | English | 4.64 s | [Listen](sonnet-18.wav) |
13
+ | The Tyger (first stanza) | William Blake | English | 8.24 s | [Listen](the-tyger.wav) |
14
+
15
+ All recordings use the requested voice description:
16
+
17
+ > 一位温柔自信的年轻女性,声音清晰,语气亲切,表达轻快而富有感染力。
18
+
19
+ Model revision, seed, generation settings, timings, and file hashes are recorded in [manifest.json](manifest.json); [prompts.json](prompts.json) contains exact texts. Audio is 24 kHz mono. These fixtures were generated on CUDA with bfloat16 and seed 42 reset for every sample; the application also retains its CPU inference path.
20
+
21
+ ## Poems
22
+
23
+ ### 静夜思 — 李白
24
+
25
+ 床前明月光,疑是地上霜。举头望明月,低头思故乡。
26
+
27
+ ### 春晓 — 孟浩然
28
+
29
+ 春眠不觉晓,处处闻啼鸟。夜来风雨声,花落知多少。
30
+
31
+ ### 登鹳雀楼 — 王之涣
32
+
33
+ 白日依山尽,黄河入海流。欲穷千里目,更上一层楼。
34
+
35
+ ### Sonnet 18 (opening couplet) — William Shakespeare
36
+
37
+ Shall I compare thee to a summer’s day? Thou art more lovely and more temperate.
38
+
39
+ ### The Tyger (first stanza) — William Blake
40
+
41
+ Tyger Tyger, burning bright, In the forests of the night; What immortal hand or eye, Could frame thy fearful symmetry?
42
+
43
+ ## Regenerate
44
+
45
+ ```bash
46
+ python scripts/generate_speech.py --device cpu \
47
+ --model Qwen/Qwen3-TTS-12Hz-1.7B-VoiceDesign \
48
+ --samples examples/poetry/prompts.json --out outputs/poetry
49
+ ```
50
+
51
+ Different hardware and library versions can change the generated voice. Use the committed recordings for direct fitting comparisons.
examples/poetry/manifest.json ADDED
@@ -0,0 +1,91 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model": "Qwen/Qwen3-TTS-12Hz-1.7B-VoiceDesign",
3
+ "source": {
4
+ "repo": "Qwen/Qwen3-TTS-12Hz-1.7B-VoiceDesign",
5
+ "revision": "5ecdb67327fd37bb2e042aab12ff7391903235d3"
6
+ },
7
+ "mode": "voice_design",
8
+ "seed": 42,
9
+ "seed_policy": "Reset RNG before each sample",
10
+ "device": "cuda",
11
+ "dtype": "bfloat16",
12
+ "attention": "sdpa",
13
+ "max_new_tokens": 512,
14
+ "samples": [
15
+ {
16
+ "name": "quiet-night",
17
+ "title": "静夜思",
18
+ "author": "李白",
19
+ "text": "床前明月光,疑是地上霜。举头望明月,低头思故乡。",
20
+ "language": "Chinese",
21
+ "instruct": "一位温柔自信的年轻女性,声音清晰,语气亲切,表达轻快而富有感染力。",
22
+ "seed": 42,
23
+ "sample_rate": 24000,
24
+ "seconds": 8.64,
25
+ "wall_seconds": 24.568893432617188,
26
+ "file": "quiet-night.wav",
27
+ "sha256": "7bdbca39bd9c406b639b9fda231219484711ca319bd4da9c0c9d991874707981",
28
+ "bytes": 829520
29
+ },
30
+ {
31
+ "name": "spring-dawn",
32
+ "title": "春晓",
33
+ "author": "孟浩然",
34
+ "text": "春眠不觉晓,处处闻啼鸟。夜来风雨声,花落知多少。",
35
+ "language": "Chinese",
36
+ "instruct": "一位温柔自信的年轻女性,声音清晰,语气亲切,表达轻快而富有感染力。",
37
+ "seed": 42,
38
+ "sample_rate": 24000,
39
+ "seconds": 8.16,
40
+ "wall_seconds": 22.90131402015686,
41
+ "file": "spring-dawn.wav",
42
+ "sha256": "316f58a7865d1f31822471e2e84f507af683ece80dab7b1f41414310796b6b67",
43
+ "bytes": 783440
44
+ },
45
+ {
46
+ "name": "stork-tower",
47
+ "title": "登鹳雀楼",
48
+ "author": "王之涣",
49
+ "text": "白日依山尽,黄河入海流。欲穷千里目,更上一层楼。",
50
+ "language": "Chinese",
51
+ "instruct": "一位温柔自信的年轻女性,声音清晰,语气亲切,表达轻快而富有感染力。",
52
+ "seed": 42,
53
+ "sample_rate": 24000,
54
+ "seconds": 7.6,
55
+ "wall_seconds": 21.23459029197693,
56
+ "file": "stork-tower.wav",
57
+ "sha256": "0aae65399547fef6d999e4b929b252fc9a428e82c7f159779d3a9bd1ca061c08",
58
+ "bytes": 729680
59
+ },
60
+ {
61
+ "name": "sonnet-18",
62
+ "title": "Sonnet 18 (opening couplet)",
63
+ "author": "William Shakespeare",
64
+ "text": "Shall I compare thee to a summer’s day? Thou art more lovely and more temperate.",
65
+ "language": "English",
66
+ "instruct": "一位温柔自信的年轻女性,声音清晰,语气亲切,表达轻快而富有感染力。",
67
+ "seed": 42,
68
+ "sample_rate": 24000,
69
+ "seconds": 4.64,
70
+ "wall_seconds": 14.138381242752075,
71
+ "file": "sonnet-18.wav",
72
+ "sha256": "55f718b8da437c778f35c9a22c8b0b8c7ab8c176e8a91ad460f036afa53cc5e5",
73
+ "bytes": 445520
74
+ },
75
+ {
76
+ "name": "the-tyger",
77
+ "title": "The Tyger (first stanza)",
78
+ "author": "William Blake",
79
+ "text": "Tyger Tyger, burning bright, In the forests of the night; What immortal hand or eye, Could frame thy fearful symmetry?",
80
+ "language": "English",
81
+ "instruct": "一位温柔自信的年轻女性,声音清晰,语气亲切,表达轻快而富有感染力。",
82
+ "seed": 42,
83
+ "sample_rate": 24000,
84
+ "seconds": 8.24,
85
+ "wall_seconds": 24.128756523132324,
86
+ "file": "the-tyger.wav",
87
+ "sha256": "bd807433cebced5629a63b5cf0ecaf213c301ccf3fc110b9b20ab74dc6f94d04",
88
+ "bytes": 791120
89
+ }
90
+ ]
91
+ }
examples/poetry/prompts.json ADDED
@@ -0,0 +1,42 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "name": "quiet-night",
4
+ "title": "静夜思",
5
+ "author": "李白",
6
+ "text": "床前明月光,疑是地上霜。举头望明月,低头思故乡。",
7
+ "language": "Chinese",
8
+ "instruct": "一位温柔自信的年轻女性,声音清晰,语气亲切,表达轻快而富有感染力。"
9
+ },
10
+ {
11
+ "name": "spring-dawn",
12
+ "title": "春晓",
13
+ "author": "孟浩然",
14
+ "text": "春眠不觉晓,处处闻啼鸟。夜来风雨声,花落知多少。",
15
+ "language": "Chinese",
16
+ "instruct": "一位温柔自信的年轻女性,声音清晰,语气亲切,表达轻快而富有感染力。"
17
+ },
18
+ {
19
+ "name": "stork-tower",
20
+ "title": "登鹳雀楼",
21
+ "author": "王之涣",
22
+ "text": "白日依山尽,黄河入海流。欲穷千里目,更上一层楼。",
23
+ "language": "Chinese",
24
+ "instruct": "一位温柔自信的年轻女性,声音清晰,语气亲切,表达轻快而富有感染力。"
25
+ },
26
+ {
27
+ "name": "sonnet-18",
28
+ "title": "Sonnet 18 (opening couplet)",
29
+ "author": "William Shakespeare",
30
+ "text": "Shall I compare thee to a summer’s day? Thou art more lovely and more temperate.",
31
+ "language": "English",
32
+ "instruct": "一位温柔自信的年轻女性,声音清晰,语气亲切,表达轻快而富有感染力。"
33
+ },
34
+ {
35
+ "name": "the-tyger",
36
+ "title": "The Tyger (first stanza)",
37
+ "author": "William Blake",
38
+ "text": "Tyger Tyger, burning bright, In the forests of the night; What immortal hand or eye, Could frame thy fearful symmetry?",
39
+ "language": "English",
40
+ "instruct": "一位温柔自信的年轻女性,声音清晰,语气亲切,表达轻快而富有感染力。"
41
+ }
42
+ ]
examples/poetry/quiet-night.wav ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7bdbca39bd9c406b639b9fda231219484711ca319bd4da9c0c9d991874707981
3
+ size 829520
examples/poetry/sonnet-18.wav ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:55f718b8da437c778f35c9a22c8b0b8c7ab8c176e8a91ad460f036afa53cc5e5
3
+ size 445520
examples/poetry/spring-dawn.wav ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:316f58a7865d1f31822471e2e84f507af683ece80dab7b1f41414310796b6b67
3
+ size 783440
examples/poetry/stork-tower.wav ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0aae65399547fef6d999e4b929b252fc9a428e82c7f159779d3a9bd1ca061c08
3
+ size 729680
examples/poetry/the-tyger.wav ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bd807433cebced5629a63b5cf0ecaf213c301ccf3fc110b9b20ab74dc6f94d04
3
+ size 791120