de-thorsten-1p57m/manifest.json
42.1 KB · 1505 lines · json Raw
1 {
2 "format": "roota.raw-fp16.v1",
3 "package_name": "de_DE-flores-20260908",
4 "language": "de_DE",
5 "voice": "de_DE-thorsten-medium",
6 "created_from": "tools/export_roota_self_contained_package.py",
7 "sample_rate": 22050,
8 "hop_length": 256,
9 "inference": {
10 "duration_length_scale": 1.0,
11 "duration_policy": "round(exp(log_duration) * duration_length_scale)"
12 },
13 "weights_file": "weights.fp16.bin",
14 "weights_size_bytes": 3130648,
15 "weights_sha256": "ddecc2f8db2f15eaeac86582416d74b40a367a92d60f32dd6d005d08e224ea34",
16 "total_parameters": 1565324,
17 "components": {
18 "duration": {
19 "checkpoint": "artifacts/voices/de_DE/duration/duration-student.pt",
20 "checkpoint_sha256": "3d047670a54996fa58fb01681625347bccaaa79740f850b304c0a6c8a53faf43",
21 "config": {
22 "architecture": "duration_conv",
23 "vocab_size": 142,
24 "hidden": 64,
25 "depth": 3,
26 "kernel_size": 5,
27 "max_tokens": 759,
28 "max_duration": 80,
29 "target_duration_scale": 1.0,
30 "pause_preserve_scale": 0.0,
31 "pause_preserve_window": 0,
32 "pause_token_ids": [
33 4,
34 8,
35 10,
36 11,
37 12,
38 13
39 ],
40 "long_preserve_threshold": 0.0,
41 "long_preserve_scale": 0.0,
42 "duration_vocab_size_forced": 0,
43 "duration_oov_id": 59
44 },
45 "parameters": 136772,
46 "tensors": [
47 {
48 "name": "blocks.0.net.0.bias",
49 "shape": [
50 64
51 ],
52 "dtype": "float16",
53 "offset_bytes": 0,
54 "nbytes": 128,
55 "sha256": "d5963ca2fcaf383258ed9878e8e91092b376bc2a5d33f1b702de513070ae1cdb"
56 },
57 {
58 "name": "blocks.0.net.0.weight",
59 "shape": [
60 64,
61 64,
62 5
63 ],
64 "dtype": "float16",
65 "offset_bytes": 128,
66 "nbytes": 40960,
67 "sha256": "6f10008ee13c315a26b49be31ae56316fd4b339ade843bf74bbb2da3480071db"
68 },
69 {
70 "name": "blocks.0.net.2.bias",
71 "shape": [
72 64
73 ],
74 "dtype": "float16",
75 "offset_bytes": 41088,
76 "nbytes": 128,
77 "sha256": "380e7438d6bc9814f74347ca449cd9ad8c47a6e382a85318a26ff3d45f5eda57"
78 },
79 {
80 "name": "blocks.0.net.2.weight",
81 "shape": [
82 64,
83 64,
84 5
85 ],
86 "dtype": "float16",
87 "offset_bytes": 41216,
88 "nbytes": 40960,
89 "sha256": "929727647f9f9a9c240151b99434bb6736817450cf4d22c7009dbb27e0fbea0e"
90 },
91 {
92 "name": "blocks.0.scale",
93 "shape": [
94 1
95 ],
96 "dtype": "float16",
97 "offset_bytes": 82176,
98 "nbytes": 2,
99 "sha256": "79642068e50edc738354e84b872e9a970c6df27e9e523877d4e269b645005c9a"
100 },
101 {
102 "name": "blocks.1.net.0.bias",
103 "shape": [
104 64
105 ],
106 "dtype": "float16",
107 "offset_bytes": 82178,
108 "nbytes": 128,
109 "sha256": "5ee13289f19840d2804b5128f217d6b8bfa017f896c111739d0f134f6fb0a9c7"
110 },
111 {
112 "name": "blocks.1.net.0.weight",
113 "shape": [
114 64,
115 64,
116 5
117 ],
118 "dtype": "float16",
119 "offset_bytes": 82306,
120 "nbytes": 40960,
121 "sha256": "11a4f7bbb4f1d272b3896f6ae8f884d209686df81c3759cd23de659d4c637c83"
122 },
123 {
124 "name": "blocks.1.net.2.bias",
125 "shape": [
126 64
127 ],
128 "dtype": "float16",
129 "offset_bytes": 123266,
130 "nbytes": 128,
131 "sha256": "ae1c979031d2d0c55f9145a6b84759c4cd5aae6511fbf41996556aecb1ba1220"
132 },
133 {
134 "name": "blocks.1.net.2.weight",
135 "shape": [
136 64,
137 64,
138 5
139 ],
140 "dtype": "float16",
141 "offset_bytes": 123394,
142 "nbytes": 40960,
143 "sha256": "5a7af8be71fb78892cf77115f108faa1ac8d7b1a014565fd0e5a54aa454c086e"
144 },
145 {
146 "name": "blocks.1.scale",
147 "shape": [
148 1
149 ],
150 "dtype": "float16",
151 "offset_bytes": 164354,
152 "nbytes": 2,
153 "sha256": "a2cd0941bc1bf85d41bdd01a7ac21492419b87eb17509fd2f48bb4c81cb82bb0"
154 },
155 {
156 "name": "blocks.2.net.0.bias",
157 "shape": [
158 64
159 ],
160 "dtype": "float16",
161 "offset_bytes": 164356,
162 "nbytes": 128,
163 "sha256": "5d6ab6b9c44795edce49467e1a66402fd5584df73fcdd8101c949cd4b5f870cf"
164 },
165 {
166 "name": "blocks.2.net.0.weight",
167 "shape": [
168 64,
169 64,
170 5
171 ],
172 "dtype": "float16",
173 "offset_bytes": 164484,
174 "nbytes": 40960,
175 "sha256": "71ab2ef9cd360d55bf82805729b5a792de34fa64ec3c71727f60276a8ec9483a"
176 },
177 {
178 "name": "blocks.2.net.2.bias",
179 "shape": [
180 64
181 ],
182 "dtype": "float16",
183 "offset_bytes": 205444,
184 "nbytes": 128,
185 "sha256": "ab6c5f912e7366a23cbaff55e53518b960cc426d416d2da747b23bc30ca3e020"
186 },
187 {
188 "name": "blocks.2.net.2.weight",
189 "shape": [
190 64,
191 64,
192 5
193 ],
194 "dtype": "float16",
195 "offset_bytes": 205572,
196 "nbytes": 40960,
197 "sha256": "dcdd13cc89ab671d4e525cd058413d42b98f03c379990b83a395546ac1bbd974"
198 },
199 {
200 "name": "blocks.2.scale",
201 "shape": [
202 1
203 ],
204 "dtype": "float16",
205 "offset_bytes": 246532,
206 "nbytes": 2,
207 "sha256": "97594727964ad1b8ea9201270d2a5c21f8dbfd5ca709e16a85e6df44007f6074"
208 },
209 {
210 "name": "embedding.weight",
211 "shape": [
212 142,
213 64
214 ],
215 "dtype": "float16",
216 "offset_bytes": 246534,
217 "nbytes": 18176,
218 "sha256": "5c7338621df4ca8549c0623c4ca8346bdc67f81193f5801e87188235e3cdd1f9"
219 },
220 {
221 "name": "input_proj.bias",
222 "shape": [
223 64
224 ],
225 "dtype": "float16",
226 "offset_bytes": 264710,
227 "nbytes": 128,
228 "sha256": "479834dcccb6118657b71bde9205b73cefe740a8c37eb5d9e741139085b63040"
229 },
230 {
231 "name": "input_proj.weight",
232 "shape": [
233 64,
234 67,
235 1
236 ],
237 "dtype": "float16",
238 "offset_bytes": 264838,
239 "nbytes": 8576,
240 "sha256": "cd2ee0b42e13b064329d7bf9767eaacd435f7b1ef4877258e88e45551b4028a9"
241 },
242 {
243 "name": "output.bias",
244 "shape": [
245 1
246 ],
247 "dtype": "float16",
248 "offset_bytes": 273414,
249 "nbytes": 2,
250 "sha256": "e9de523f005a2b36aba07fee12173980e15edbf71749509d993644cbc506754c"
251 },
252 {
253 "name": "output.weight",
254 "shape": [
255 1,
256 64,
257 1
258 ],
259 "dtype": "float16",
260 "offset_bytes": 273416,
261 "nbytes": 128,
262 "sha256": "13e939082b578453533ecda678d8065498574e9baca775da13a37dfa104133c7"
263 }
264 ]
265 },
266 "acoustic": {
267 "checkpoint": "artifacts/voices/de_DE/joint/latent-student.pt",
268 "checkpoint_sha256": "82d3ca50c21bf40f44bb802c1fa228b04262893ec881166b362e05f4fef6d782",
269 "config": {
270 "architecture": "token_context",
271 "vocab_size": 142,
272 "hidden": 96,
273 "depth": 4,
274 "kernel_size": 5,
275 "out_channels": 192,
276 "token_depth": 3
277 },
278 "parameters": 697831,
279 "tensors": [
280 {
281 "name": "embedding.weight",
282 "shape": [
283 142,
284 96
285 ],
286 "dtype": "float16",
287 "offset_bytes": 273544,
288 "nbytes": 27264,
289 "sha256": "6473e5c9ff6ac78aab27af84be2efecccd080d292c013457ed29cab3c878d4f9"
290 },
291 {
292 "name": "frame_blocks.0.net.0.bias",
293 "shape": [
294 96
295 ],
296 "dtype": "float16",
297 "offset_bytes": 300808,
298 "nbytes": 192,
299 "sha256": "2fba91850eb30df974270b020957bb73b81bade4ee7b3a6ce1b6e0e672fd393d"
300 },
301 {
302 "name": "frame_blocks.0.net.0.weight",
303 "shape": [
304 96,
305 96,
306 5
307 ],
308 "dtype": "float16",
309 "offset_bytes": 301000,
310 "nbytes": 92160,
311 "sha256": "67e9ab5087a25ea89ce944aa40c525f8da2eaa2a5f8d744e57d4976fb0c37e10"
312 },
313 {
314 "name": "frame_blocks.0.net.2.bias",
315 "shape": [
316 96
317 ],
318 "dtype": "float16",
319 "offset_bytes": 393160,
320 "nbytes": 192,
321 "sha256": "e9e1ccf5d88852fe5ffecbd55c2bab630cc8f503ff3a58e3e74f93a34f1b9863"
322 },
323 {
324 "name": "frame_blocks.0.net.2.weight",
325 "shape": [
326 96,
327 96,
328 5
329 ],
330 "dtype": "float16",
331 "offset_bytes": 393352,
332 "nbytes": 92160,
333 "sha256": "8a4bf87647017068ac4cc1f7b16e9884db6f152ede5ec89eb4d7d2ec02d3c556"
334 },
335 {
336 "name": "frame_blocks.0.scale",
337 "shape": [
338 1
339 ],
340 "dtype": "float16",
341 "offset_bytes": 485512,
342 "nbytes": 2,
343 "sha256": "cfe1b8990b0b7ecacfc8b03a53e09b0206f39d9de1a2c5723fce567db434d9c3"
344 },
345 {
346 "name": "frame_blocks.1.net.0.bias",
347 "shape": [
348 96
349 ],
350 "dtype": "float16",
351 "offset_bytes": 485514,
352 "nbytes": 192,
353 "sha256": "d40d6c34f0e2a3a99b156092820da6b7ffedc8c673613084f2529e280caf50e5"
354 },
355 {
356 "name": "frame_blocks.1.net.0.weight",
357 "shape": [
358 96,
359 96,
360 5
361 ],
362 "dtype": "float16",
363 "offset_bytes": 485706,
364 "nbytes": 92160,
365 "sha256": "4b44ffdb07c22210b27540c87bcf24c0a4ab16da1b72da056fb1b166591ba377"
366 },
367 {
368 "name": "frame_blocks.1.net.2.bias",
369 "shape": [
370 96
371 ],
372 "dtype": "float16",
373 "offset_bytes": 577866,
374 "nbytes": 192,
375 "sha256": "c771da4c777bcd3feef1dc2d208ecc8aaf924a65e1d3738174a5c52f348514c2"
376 },
377 {
378 "name": "frame_blocks.1.net.2.weight",
379 "shape": [
380 96,
381 96,
382 5
383 ],
384 "dtype": "float16",
385 "offset_bytes": 578058,
386 "nbytes": 92160,
387 "sha256": "2778e93aed7e104cee65e514a818cde4090c11de7c4ce5add2d463ca04171f37"
388 },
389 {
390 "name": "frame_blocks.1.scale",
391 "shape": [
392 1
393 ],
394 "dtype": "float16",
395 "offset_bytes": 670218,
396 "nbytes": 2,
397 "sha256": "5ba88d4e9124946940346279f643335341f6843f538be1f63035874767bada0e"
398 },
399 {
400 "name": "frame_blocks.2.net.0.bias",
401 "shape": [
402 96
403 ],
404 "dtype": "float16",
405 "offset_bytes": 670220,
406 "nbytes": 192,
407 "sha256": "1ca0163546f7ff9ddf061942effc731a3c75f762f52f912abf70255b2e8ffca8"
408 },
409 {
410 "name": "frame_blocks.2.net.0.weight",
411 "shape": [
412 96,
413 96,
414 5
415 ],
416 "dtype": "float16",
417 "offset_bytes": 670412,
418 "nbytes": 92160,
419 "sha256": "45f0e961ae260d90acf88a80e3f27ac54d24609cd58d62142bc63bf8906d9629"
420 },
421 {
422 "name": "frame_blocks.2.net.2.bias",
423 "shape": [
424 96
425 ],
426 "dtype": "float16",
427 "offset_bytes": 762572,
428 "nbytes": 192,
429 "sha256": "41f9f4ae0e0c55f3e05530dd7ebdc5850fadb6bdd955d5461efcba34dc0aeca0"
430 },
431 {
432 "name": "frame_blocks.2.net.2.weight",
433 "shape": [
434 96,
435 96,
436 5
437 ],
438 "dtype": "float16",
439 "offset_bytes": 762764,
440 "nbytes": 92160,
441 "sha256": "66e53dde609d50bbf3b3f00c333745845868f590ff724b722d469e5e2285db1f"
442 },
443 {
444 "name": "frame_blocks.2.scale",
445 "shape": [
446 1
447 ],
448 "dtype": "float16",
449 "offset_bytes": 854924,
450 "nbytes": 2,
451 "sha256": "5d54ba1ca6248e0ded00b2f133a479baf75f415dbd5d8838fc4c629dc5204563"
452 },
453 {
454 "name": "frame_blocks.3.net.0.bias",
455 "shape": [
456 96
457 ],
458 "dtype": "float16",
459 "offset_bytes": 854926,
460 "nbytes": 192,
461 "sha256": "65c523a5c1f0e6c5b836ae945c2a70bcb41a8d58f249a844601a8a1ed06d4260"
462 },
463 {
464 "name": "frame_blocks.3.net.0.weight",
465 "shape": [
466 96,
467 96,
468 5
469 ],
470 "dtype": "float16",
471 "offset_bytes": 855118,
472 "nbytes": 92160,
473 "sha256": "099a17d1ebbae770161c0a2e42920da19d0bd88ddd082b6f5278b76cb282dcd8"
474 },
475 {
476 "name": "frame_blocks.3.net.2.bias",
477 "shape": [
478 96
479 ],
480 "dtype": "float16",
481 "offset_bytes": 947278,
482 "nbytes": 192,
483 "sha256": "ca9ce9d9f8f7ce50488f46e1f1359a32129aa1ae05af497ba9a54670575cde15"
484 },
485 {
486 "name": "frame_blocks.3.net.2.weight",
487 "shape": [
488 96,
489 96,
490 5
491 ],
492 "dtype": "float16",
493 "offset_bytes": 947470,
494 "nbytes": 92160,
495 "sha256": "4575915869a10149e90b514be9e89e39d941bf10405d6e4879dab03337142d79"
496 },
497 {
498 "name": "frame_blocks.3.scale",
499 "shape": [
500 1
501 ],
502 "dtype": "float16",
503 "offset_bytes": 1039630,
504 "nbytes": 2,
505 "sha256": "619c61c11e257c40d9706d4c98ed2896d636571375a72195b55883416fa3545d"
506 },
507 {
508 "name": "frame_input_proj.bias",
509 "shape": [
510 96
511 ],
512 "dtype": "float16",
513 "offset_bytes": 1039632,
514 "nbytes": 192,
515 "sha256": "d7cd1198edc4357b6e0a409d3f32aade9e8169ab323dc0cc42d8f019d3eae5e7"
516 },
517 {
518 "name": "frame_input_proj.weight",
519 "shape": [
520 96,
521 99,
522 1
523 ],
524 "dtype": "float16",
525 "offset_bytes": 1039824,
526 "nbytes": 19008,
527 "sha256": "a72d99762c2a3e46b2a2f4a46b3ad6fc00d65fe368b3e10e3c418e8d8942d9bb"
528 },
529 {
530 "name": "output.bias",
531 "shape": [
532 192
533 ],
534 "dtype": "float16",
535 "offset_bytes": 1058832,
536 "nbytes": 384,
537 "sha256": "f852a0073e8f21b1a2c95e9713223c257abc8281d3f1c4e52482029005549941"
538 },
539 {
540 "name": "output.weight",
541 "shape": [
542 192,
543 96,
544 1
545 ],
546 "dtype": "float16",
547 "offset_bytes": 1059216,
548 "nbytes": 36864,
549 "sha256": "62dc502cd254c36e90ee61cd67bcb477ab8058d6d77d0f37fccc716edf64c4ae"
550 },
551 {
552 "name": "token_blocks.0.net.0.bias",
553 "shape": [
554 96
555 ],
556 "dtype": "float16",
557 "offset_bytes": 1096080,
558 "nbytes": 192,
559 "sha256": "fd9e90a257f3c08d7ca3b54a10eae9a5b3f5c6463ee8d9c161bac0ea75686c12"
560 },
561 {
562 "name": "token_blocks.0.net.0.weight",
563 "shape": [
564 96,
565 96,
566 5
567 ],
568 "dtype": "float16",
569 "offset_bytes": 1096272,
570 "nbytes": 92160,
571 "sha256": "5c09f6e9b8541676664965daa85e5a9bc8ab83b92e9f64ce1653aa9b2e0e3c29"
572 },
573 {
574 "name": "token_blocks.0.net.2.bias",
575 "shape": [
576 96
577 ],
578 "dtype": "float16",
579 "offset_bytes": 1188432,
580 "nbytes": 192,
581 "sha256": "f31ea54242cc5ab79b39117992edb97636e08b6bd03c0c75e02494ffd0ff7a59"
582 },
583 {
584 "name": "token_blocks.0.net.2.weight",
585 "shape": [
586 96,
587 96,
588 5
589 ],
590 "dtype": "float16",
591 "offset_bytes": 1188624,
592 "nbytes": 92160,
593 "sha256": "44b2547531a5daab2dee8b714e3c3f4408cabd631873c22677225e2292a0689f"
594 },
595 {
596 "name": "token_blocks.0.scale",
597 "shape": [
598 1
599 ],
600 "dtype": "float16",
601 "offset_bytes": 1280784,
602 "nbytes": 2,
603 "sha256": "6f2845e2cf0d48b81a10faab93fd4ced79c5032662ab1d56e64f35ec48eeb5db"
604 },
605 {
606 "name": "token_blocks.1.net.0.bias",
607 "shape": [
608 96
609 ],
610 "dtype": "float16",
611 "offset_bytes": 1280786,
612 "nbytes": 192,
613 "sha256": "45f4e4af38604a2e1acb4a6a5a53a1228775d3940f3a348b07804d6cebfbf00e"
614 },
615 {
616 "name": "token_blocks.1.net.0.weight",
617 "shape": [
618 96,
619 96,
620 5
621 ],
622 "dtype": "float16",
623 "offset_bytes": 1280978,
624 "nbytes": 92160,
625 "sha256": "7c26370bbcc8cdd6b1f7953ad8778e996b721cc3f68f3c40f72aab1b0e30f037"
626 },
627 {
628 "name": "token_blocks.1.net.2.bias",
629 "shape": [
630 96
631 ],
632 "dtype": "float16",
633 "offset_bytes": 1373138,
634 "nbytes": 192,
635 "sha256": "4108e758858b5c11be0f41a4d457f7a316d3ec22027a6dc637bdef6cb6bea177"
636 },
637 {
638 "name": "token_blocks.1.net.2.weight",
639 "shape": [
640 96,
641 96,
642 5
643 ],
644 "dtype": "float16",
645 "offset_bytes": 1373330,
646 "nbytes": 92160,
647 "sha256": "22712078383757cb09d8b7e0fd8b132f2f2b00f29ca824b8c0bfab03f4cad818"
648 },
649 {
650 "name": "token_blocks.1.scale",
651 "shape": [
652 1
653 ],
654 "dtype": "float16",
655 "offset_bytes": 1465490,
656 "nbytes": 2,
657 "sha256": "eef0f49308adeee241f284cda2b22efedd652e869e9aedd27ffd3b95c6bd4060"
658 },
659 {
660 "name": "token_blocks.2.net.0.bias",
661 "shape": [
662 96
663 ],
664 "dtype": "float16",
665 "offset_bytes": 1465492,
666 "nbytes": 192,
667 "sha256": "6e33083dac5981e5be9cad33547b62c6e7871cc7267d0241594e961d840fd0ce"
668 },
669 {
670 "name": "token_blocks.2.net.0.weight",
671 "shape": [
672 96,
673 96,
674 5
675 ],
676 "dtype": "float16",
677 "offset_bytes": 1465684,
678 "nbytes": 92160,
679 "sha256": "a95168bcd6cc60e9231700ce94720643aac0a1de3996b571c11a7f2d4fdfa303"
680 },
681 {
682 "name": "token_blocks.2.net.2.bias",
683 "shape": [
684 96
685 ],
686 "dtype": "float16",
687 "offset_bytes": 1557844,
688 "nbytes": 192,
689 "sha256": "dd8311c36dac2b8dc44c029b5f2c1ce00682670c7282a8812285fab68c6670c7"
690 },
691 {
692 "name": "token_blocks.2.net.2.weight",
693 "shape": [
694 96,
695 96,
696 5
697 ],
698 "dtype": "float16",
699 "offset_bytes": 1558036,
700 "nbytes": 92160,
701 "sha256": "75f4b2e1efc25005d72237dbae16bca9ed75ce741e70864a0afb24cf660caa9c"
702 },
703 {
704 "name": "token_blocks.2.scale",
705 "shape": [
706 1
707 ],
708 "dtype": "float16",
709 "offset_bytes": 1650196,
710 "nbytes": 2,
711 "sha256": "45a0f494311e681e9d23d35770343794867de4bedd7dd21b53582c8b90a56099"
712 },
713 {
714 "name": "token_input_proj.bias",
715 "shape": [
716 96
717 ],
718 "dtype": "float16",
719 "offset_bytes": 1650198,
720 "nbytes": 192,
721 "sha256": "db96a751fd326ecc01b216d6439aa0528bb87ea4a51487ea823fff966f75bae1"
722 },
723 {
724 "name": "token_input_proj.weight",
725 "shape": [
726 96,
727 98,
728 1
729 ],
730 "dtype": "float16",
731 "offset_bytes": 1650390,
732 "nbytes": 18816,
733 "sha256": "930d475dfcf69a2319c8e8e4319a424764dd336d24509537d87214173b0b8b6f"
734 }
735 ]
736 },
737 "decoder": {
738 "checkpoint": "artifacts/voices/de_DE/joint/decoder-student.pt",
739 "checkpoint_sha256": "710e49a51a34a2571c760f35d0e23793684a44e3a50fa9110ac0c3e0d3f1e2ad",
740 "config": {
741 "in_channels": 192,
742 "source_in_channels": 192,
743 "channels": [
744 160,
745 80,
746 40,
747 20
748 ],
749 "res_layers": 1,
750 "variant": "piperlite",
751 "lrc_code_dim": 40,
752 "lrc_encoder_hidden": 64,
753 "rank_ratio": 0.5,
754 "activation": "leaky_relu",
755 "stage_affine": false,
756 "factorized_pre_rank": 0,
757 "piper_res_factor_rank_ratio": 0.0,
758 "res_bank_scale_mode": "kept",
759 "stage0_branches": [
760 0,
761 1,
762 2
763 ],
764 "stage1_branches": [
765 0,
766 1,
767 2
768 ],
769 "stage2_branches": [
770 0,
771 1,
772 2
773 ],
774 "stage3_branches": [
775 0,
776 1,
777 2
778 ],
779 "fsd_dim": 72,
780 "fsd_blocks": 5,
781 "fsd_film_rank": 12,
782 "fsd_head_rank": 48,
783 "wavehax_channels": 32,
784 "wavehax_blocks": 8,
785 "wavehax_mult_channels": 2,
786 "wavehax_kernel_freq": 7,
787 "wavehax_kernel_time": 7,
788 "wavehax_f0_channel": 80,
789 "wavehax_voiced_channel": 81,
790 "wavehax_sample_rate": 24000,
791 "wavehax_prior_power": 0.1,
792 "wavehax_prior_noise": 0.01,
793 "whx_channels": 32,
794 "whx_hidden": 64,
795 "whx_blocks": 8,
796 "whx_n_fft": 512,
797 "whx_f0_channel": 80,
798 "whx_voiced_channel": 81,
799 "whx_sample_rate": 24000,
800 "hift_width": 224,
801 "hift_gen_channels": 192,
802 "hift_style_dim": 64,
803 "hift_body_blocks": 3,
804 "hift_f0_channel": 80,
805 "hift_voiced_channel": 81,
806 "hift_sample_rate": 24000,
807 "stage_projection_bottlenecks": [],
808 "teacher_init_checkpoint": null,
809 "teacher_init_method": "first",
810 "teacher_init": null,
811 "init_decoder_checkpoint": "artifacts/voices/de_DE/decoder/decoder-student.pt",
812 "decoder_init": {
813 "checkpoint": "artifacts/voices/de_DE/decoder/decoder-student.pt",
814 "decoder_parameters": 730721,
815 "matched_config": {
816 "in_channels": 192,
817 "channels": [
818 160,
819 80,
820 40,
821 20
822 ],
823 "res_layers": 1,
824 "variant": "piperlite",
825 "rank_ratio": 0.5,
826 "activation": "leaky_relu",
827 "stage_affine": false,
828 "factorized_pre_rank": 0,
829 "piper_res_factor_rank_ratio": 0.0,
830 "res_bank_scale_mode": "kept",
831 "stage0_branches": [
832 0,
833 1,
834 2
835 ],
836 "stage1_branches": [
837 0,
838 1,
839 2
840 ],
841 "stage2_branches": [
842 0,
843 1,
844 2
845 ],
846 "stage3_branches": [
847 0,
848 1,
849 2
850 ],
851 "fsd_dim": 72,
852 "fsd_blocks": 5,
853 "fsd_film_rank": 12,
854 "fsd_head_rank": 48,
855 "wavehax_channels": 32,
856 "wavehax_blocks": 8,
857 "wavehax_mult_channels": 2,
858 "wavehax_kernel_freq": 7,
859 "wavehax_kernel_time": 7,
860 "wavehax_f0_channel": 80,
861 "wavehax_voiced_channel": 81,
862 "wavehax_sample_rate": 24000,
863 "wavehax_prior_power": 0.1,
864 "wavehax_prior_noise": 0.01,
865 "whx_channels": 32,
866 "whx_hidden": 64,
867 "whx_blocks": 8,
868 "whx_n_fft": 512,
869 "whx_f0_channel": 80,
870 "whx_voiced_channel": 81,
871 "whx_sample_rate": 24000,
872 "hift_width": 224,
873 "hift_gen_channels": 192,
874 "hift_style_dim": 64,
875 "hift_body_blocks": 3,
876 "hift_f0_channel": 80,
877 "hift_voiced_channel": 81,
878 "hift_sample_rate": 24000,
879 "stage_projection_bottlenecks": [],
880 "post_filter_channels": 0,
881 "post_filter_layers": 0,
882 "post_filter_kernel": 9,
883 "post_filter_scale": 0.25,
884 "pre_tanh_repair_channels": 0,
885 "pre_tanh_repair_layers": 0,
886 "pre_tanh_repair_kernel": 7,
887 "pre_tanh_repair_scale": 0.15,
888 "istft_n_fft": 512,
889 "istft_tail_head_rank": 20
890 },
891 "allowed_post_filter_mismatches": [],
892 "allowed_pre_tanh_repair_mismatches": [],
893 "allowed_activation_mismatches": [],
894 "missing_keys": [],
895 "unexpected_keys": []
896 },
897 "post_filter_channels": 0,
898 "post_filter_layers": 0,
899 "post_filter_kernel": 9,
900 "post_filter_scale": 0.25,
901 "pre_tanh_repair_channels": 0,
902 "pre_tanh_repair_layers": 0,
903 "pre_tanh_repair_kernel": 7,
904 "pre_tanh_repair_scale": 0.15,
905 "istft_n_fft": 512,
906 "istft_tail_head_rank": 20,
907 "ap_amplitude_weight": 0.0,
908 "ap_phase_weight": 0.0,
909 "ap_complex_weight": 0.0,
910 "spectral_head_init": "default",
911 "spectral_head_init_scale": 0.001,
912 "ap_amp_init_bias": -6.0,
913 "ap_phase_real_init_bias": 1.0,
914 "spectral_head_init_summary": null,
915 "feature_exact_weight": 0.0,
916 "feature_exact_keys": [
917 "pre",
918 "up0"
919 ],
920 "feature_exact_teacher_init": null,
921 "signature_pack_dir": "artifacts/voices/de_DE/signatures",
922 "exact_feature_pack_dir": null,
923 "signature_hint_weight": 0.05,
924 "signature_temporal_weight": 0.4,
925 "signature_phase_weight": 0.0,
926 "signature_phase_bins": 0,
927 "signature_keys": [
928 "stage0_mix",
929 "stage1_mix",
930 "stage2_mix",
931 "pre_tanh",
932 "audio"
933 ],
934 "bottleneck_code_checkpoint": null,
935 "bottleneck_code_weight": 0.0,
936 "bottleneck_code_keys": [],
937 "bottleneck_code_training_only_parameters": 0,
938 "echo_tail_weight": 0.0,
939 "echo_tail_min_ms": 8.0,
940 "echo_tail_max_ms": 40.0,
941 "echo_tail_lags": 9,
942 "echo_tail_margin": 0.02,
943 "paired_acoustic_residual_weight": 0.0,
944 "paired_acoustic_residual_max_scale": 0.25,
945 "lrc_pred_code_mix_prob": 0.0,
946 "lrc_pred_code_checkpoint": null,
947 "lrc_pred_code_residual_prob": 0.0,
948 "lrc_pred_code_residual_max_scale": 0.25,
949 "lrc_pred_code_cache": null
950 },
951 "parameters": 730721,
952 "tensors": [
953 {
954 "name": "post.bias",
955 "shape": [
956 1
957 ],
958 "dtype": "float16",
959 "offset_bytes": 1669206,
960 "nbytes": 2,
961 "sha256": "1b4d44143dc4935a27037e4a778395d6cb2ce46c6a413e5deaee6d00baa3f69b"
962 },
963 {
964 "name": "post.weight",
965 "shape": [
966 1,
967 20,
968 7
969 ],
970 "dtype": "float16",
971 "offset_bytes": 1669208,
972 "nbytes": 280,
973 "sha256": "66bea518b7ba96818872c3fdf5e7f7e43a42ebc4f8380e89cb9ecb3c93affb40"
974 },
975 {
976 "name": "pre.bias",
977 "shape": [
978 160
979 ],
980 "dtype": "float16",
981 "offset_bytes": 1669488,
982 "nbytes": 320,
983 "sha256": "ad2cb46f940b08faf091d657c2bc91dfd723777b8bb30e53b8ff6298492b952d"
984 },
985 {
986 "name": "pre.weight",
987 "shape": [
988 160,
989 192,
990 7
991 ],
992 "dtype": "float16",
993 "offset_bytes": 1669808,
994 "nbytes": 430080,
995 "sha256": "48bd00775541c729f229b9cc515a7c763306b6c6a3f8f00472e78017ef7302cc"
996 },
997 {
998 "name": "res0.0.blocks.0.conv1.bias",
999 "shape": [
1000 80
1001 ],
1002 "dtype": "float16",
1003 "offset_bytes": 2099888,
1004 "nbytes": 160,
1005 "sha256": "1dd46a27673a1211e978c851d6398e60030dd55044b0922004b51af8dcb46c18"
1006 },
1007 {
1008 "name": "res0.0.blocks.0.conv1.weight",
1009 "shape": [
1010 80,
1011 80,
1012 3
1013 ],
1014 "dtype": "float16",
1015 "offset_bytes": 2100048,
1016 "nbytes": 38400,
1017 "sha256": "b54672099e27715062dae273a61a42c7468c81b66fcd40d75d2e7dcde3fb0291"
1018 },
1019 {
1020 "name": "res0.0.blocks.0.conv2.bias",
1021 "shape": [
1022 80
1023 ],
1024 "dtype": "float16",
1025 "offset_bytes": 2138448,
1026 "nbytes": 160,
1027 "sha256": "8a9cedbd7537fe71c3eb8dd1ce2302275a747a3140855a966470a6e9d5476908"
1028 },
1029 {
1030 "name": "res0.0.blocks.0.conv2.weight",
1031 "shape": [
1032 80,
1033 80,
1034 3
1035 ],
1036 "dtype": "float16",
1037 "offset_bytes": 2138608,
1038 "nbytes": 38400,
1039 "sha256": "8d508c6372b8cccb37ca6ca541ea92f416b708da96e4f5bb0171197454ab44ed"
1040 },
1041 {
1042 "name": "res0.0.blocks.1.conv1.bias",
1043 "shape": [
1044 80
1045 ],
1046 "dtype": "float16",
1047 "offset_bytes": 2177008,
1048 "nbytes": 160,
1049 "sha256": "e2212a65ee0432bf074ac11a97bf0b618cc26e6575d5523b7d0b4115d65214b4"
1050 },
1051 {
1052 "name": "res0.0.blocks.1.conv1.weight",
1053 "shape": [
1054 80,
1055 80,
1056 5
1057 ],
1058 "dtype": "float16",
1059 "offset_bytes": 2177168,
1060 "nbytes": 64000,
1061 "sha256": "aca87da28039a4d7f83bddb37ecec8347add9cd9d9ba0c8062102919a91ded61"
1062 },
1063 {
1064 "name": "res0.0.blocks.1.conv2.bias",
1065 "shape": [
1066 80
1067 ],
1068 "dtype": "float16",
1069 "offset_bytes": 2241168,
1070 "nbytes": 160,
1071 "sha256": "114285e6a50a8427aeda002ef609d39e733829a686168b12e3e0ed033fc4e115"
1072 },
1073 {
1074 "name": "res0.0.blocks.1.conv2.weight",
1075 "shape": [
1076 80,
1077 80,
1078 5
1079 ],
1080 "dtype": "float16",
1081 "offset_bytes": 2241328,
1082 "nbytes": 64000,
1083 "sha256": "3a0170d0e7f2c2df433df909873d1452420bc3b6495f5dbb1d2b47a02825a8b9"
1084 },
1085 {
1086 "name": "res0.0.blocks.2.conv1.bias",
1087 "shape": [
1088 80
1089 ],
1090 "dtype": "float16",
1091 "offset_bytes": 2305328,
1092 "nbytes": 160,
1093 "sha256": "80323bdf4b2e3ad66e48a0609e1b3a70b5899045f726ad308d0327eb9b7431af"
1094 },
1095 {
1096 "name": "res0.0.blocks.2.conv1.weight",
1097 "shape": [
1098 80,
1099 80,
1100 7
1101 ],
1102 "dtype": "float16",
1103 "offset_bytes": 2305488,
1104 "nbytes": 89600,
1105 "sha256": "28ebd6afd885f09f4da367d3f955295b02248d1c4ee4e88586c566512e2fbb51"
1106 },
1107 {
1108 "name": "res0.0.blocks.2.conv2.bias",
1109 "shape": [
1110 80
1111 ],
1112 "dtype": "float16",
1113 "offset_bytes": 2395088,
1114 "nbytes": 160,
1115 "sha256": "481b227cfd24af4a380b562b4184535a7160f63b9d72393d28b6be05e76d028a"
1116 },
1117 {
1118 "name": "res0.0.blocks.2.conv2.weight",
1119 "shape": [
1120 80,
1121 80,
1122 7
1123 ],
1124 "dtype": "float16",
1125 "offset_bytes": 2395248,
1126 "nbytes": 89600,
1127 "sha256": "55587dd14a47fdc5d919cff44e8e2fdcd4fb8a7450f0e0c453ee301b6b9b1808"
1128 },
1129 {
1130 "name": "res1.0.blocks.0.conv1.bias",
1131 "shape": [
1132 40
1133 ],
1134 "dtype": "float16",
1135 "offset_bytes": 2484848,
1136 "nbytes": 80,
1137 "sha256": "93662730529500dec7ec6dbeee0a2e4196ecdb0bee3250b08186b7462b9dfa8c"
1138 },
1139 {
1140 "name": "res1.0.blocks.0.conv1.weight",
1141 "shape": [
1142 40,
1143 40,
1144 3
1145 ],
1146 "dtype": "float16",
1147 "offset_bytes": 2484928,
1148 "nbytes": 9600,
1149 "sha256": "3df19d06fbd7c077ae2ea1caa42f5cf5a04733695acdacbb7e684e03f1fe0210"
1150 },
1151 {
1152 "name": "res1.0.blocks.0.conv2.bias",
1153 "shape": [
1154 40
1155 ],
1156 "dtype": "float16",
1157 "offset_bytes": 2494528,
1158 "nbytes": 80,
1159 "sha256": "395510d2a03d187cc30e10490f950ab4004ea43e9273d2b640883200c807c5dd"
1160 },
1161 {
1162 "name": "res1.0.blocks.0.conv2.weight",
1163 "shape": [
1164 40,
1165 40,
1166 3
1167 ],
1168 "dtype": "float16",
1169 "offset_bytes": 2494608,
1170 "nbytes": 9600,
1171 "sha256": "6670ad43e39271224759399e36cbc1a12556614f7bfd1855a2d7341f619717f5"
1172 },
1173 {
1174 "name": "res1.0.blocks.1.conv1.bias",
1175 "shape": [
1176 40
1177 ],
1178 "dtype": "float16",
1179 "offset_bytes": 2504208,
1180 "nbytes": 80,
1181 "sha256": "1788d94cf7bc19d25b687e9574e2b3624c2c8659edde79a3bc05f7b80c0b637c"
1182 },
1183 {
1184 "name": "res1.0.blocks.1.conv1.weight",
1185 "shape": [
1186 40,
1187 40,
1188 5
1189 ],
1190 "dtype": "float16",
1191 "offset_bytes": 2504288,
1192 "nbytes": 16000,
1193 "sha256": "9a3a371b689ed47c4209805f13a0f879d8752838178ef18bcd491b75cce5ab77"
1194 },
1195 {
1196 "name": "res1.0.blocks.1.conv2.bias",
1197 "shape": [
1198 40
1199 ],
1200 "dtype": "float16",
1201 "offset_bytes": 2520288,
1202 "nbytes": 80,
1203 "sha256": "5db91928283aec8603d80c2803fe4ea39a3db2191266f42d3fc2a17a221d255e"
1204 },
1205 {
1206 "name": "res1.0.blocks.1.conv2.weight",
1207 "shape": [
1208 40,
1209 40,
1210 5
1211 ],
1212 "dtype": "float16",
1213 "offset_bytes": 2520368,
1214 "nbytes": 16000,
1215 "sha256": "09c2636d24e7f86ee40bbe01412d5210d260050b162d635e58dee997327bda5e"
1216 },
1217 {
1218 "name": "res1.0.blocks.2.conv1.bias",
1219 "shape": [
1220 40
1221 ],
1222 "dtype": "float16",
1223 "offset_bytes": 2536368,
1224 "nbytes": 80,
1225 "sha256": "7d267ed4875356ba4a283fd306c3f91e8b883ffaea306b8e45f6dff78de3d9c8"
1226 },
1227 {
1228 "name": "res1.0.blocks.2.conv1.weight",
1229 "shape": [
1230 40,
1231 40,
1232 7
1233 ],
1234 "dtype": "float16",
1235 "offset_bytes": 2536448,
1236 "nbytes": 22400,
1237 "sha256": "058efdc4aa8c66cdd462531a9192b18845797681edea3d6df51661fae26e9c54"
1238 },
1239 {
1240 "name": "res1.0.blocks.2.conv2.bias",
1241 "shape": [
1242 40
1243 ],
1244 "dtype": "float16",
1245 "offset_bytes": 2558848,
1246 "nbytes": 80,
1247 "sha256": "0744680f9c60259ee6cfd6c9287b12925cf6bca4841dbb0098087a9f01100443"
1248 },
1249 {
1250 "name": "res1.0.blocks.2.conv2.weight",
1251 "shape": [
1252 40,
1253 40,
1254 7
1255 ],
1256 "dtype": "float16",
1257 "offset_bytes": 2558928,
1258 "nbytes": 22400,
1259 "sha256": "6719a239e47a2b82d37395e400c3a69a46d42f1bcd04058dafe2453d9da1f94c"
1260 },
1261 {
1262 "name": "res2.0.blocks.0.conv1.bias",
1263 "shape": [
1264 20
1265 ],
1266 "dtype": "float16",
1267 "offset_bytes": 2581328,
1268 "nbytes": 40,
1269 "sha256": "7f5687207c3934cdb050ece740b8b8e55be7db0319f8a9c8ebc3197bea847308"
1270 },
1271 {
1272 "name": "res2.0.blocks.0.conv1.weight",
1273 "shape": [
1274 20,
1275 20,
1276 3
1277 ],
1278 "dtype": "float16",
1279 "offset_bytes": 2581368,
1280 "nbytes": 2400,
1281 "sha256": "fe637485d975b9025bfd78970657d891760d3085d9f4d7309c3f6075ea0751da"
1282 },
1283 {
1284 "name": "res2.0.blocks.0.conv2.bias",
1285 "shape": [
1286 20
1287 ],
1288 "dtype": "float16",
1289 "offset_bytes": 2583768,
1290 "nbytes": 40,
1291 "sha256": "84eae343a0d32a4f58945453b9c97c47962b894a04215b8a9fb06f474026da5c"
1292 },
1293 {
1294 "name": "res2.0.blocks.0.conv2.weight",
1295 "shape": [
1296 20,
1297 20,
1298 3
1299 ],
1300 "dtype": "float16",
1301 "offset_bytes": 2583808,
1302 "nbytes": 2400,
1303 "sha256": "91804d39e926627c53994e12708508e31d07e6dbe8bd765e32b5cdcf5da11e7c"
1304 },
1305 {
1306 "name": "res2.0.blocks.1.conv1.bias",
1307 "shape": [
1308 20
1309 ],
1310 "dtype": "float16",
1311 "offset_bytes": 2586208,
1312 "nbytes": 40,
1313 "sha256": "ad1cc3f9a9b7691ed54dfe321cd9bfc69bc68013c156a930f9a5d272f45b4f1d"
1314 },
1315 {
1316 "name": "res2.0.blocks.1.conv1.weight",
1317 "shape": [
1318 20,
1319 20,
1320 5
1321 ],
1322 "dtype": "float16",
1323 "offset_bytes": 2586248,
1324 "nbytes": 4000,
1325 "sha256": "0436f9458194d20735aab8870eb61656c8d7bc5850c264d98afd0ee035afd50b"
1326 },
1327 {
1328 "name": "res2.0.blocks.1.conv2.bias",
1329 "shape": [
1330 20
1331 ],
1332 "dtype": "float16",
1333 "offset_bytes": 2590248,
1334 "nbytes": 40,
1335 "sha256": "8a6c0e1f25d9244509581c129a2cd26a4d224ef1306553d5f20e0fe34820a909"
1336 },
1337 {
1338 "name": "res2.0.blocks.1.conv2.weight",
1339 "shape": [
1340 20,
1341 20,
1342 5
1343 ],
1344 "dtype": "float16",
1345 "offset_bytes": 2590288,
1346 "nbytes": 4000,
1347 "sha256": "7cf3930aa834be21bdf12c9beb42808c27c09909d9275f252dc4a889142a89ee"
1348 },
1349 {
1350 "name": "res2.0.blocks.2.conv1.bias",
1351 "shape": [
1352 20
1353 ],
1354 "dtype": "float16",
1355 "offset_bytes": 2594288,
1356 "nbytes": 40,
1357 "sha256": "ec8bd64bab5a7b187b345761c0a8fed093f59e7ecaa226e1016d8046629bcbb0"
1358 },
1359 {
1360 "name": "res2.0.blocks.2.conv1.weight",
1361 "shape": [
1362 20,
1363 20,
1364 7
1365 ],
1366 "dtype": "float16",
1367 "offset_bytes": 2594328,
1368 "nbytes": 5600,
1369 "sha256": "e0745f6fa143a437f63ed9b791c65a93cd47ad5745b80efbab45f78699fb5f90"
1370 },
1371 {
1372 "name": "res2.0.blocks.2.conv2.bias",
1373 "shape": [
1374 20
1375 ],
1376 "dtype": "float16",
1377 "offset_bytes": 2599928,
1378 "nbytes": 40,
1379 "sha256": "3186eee5ca0e5aa37c7dc4e619d677d1a631df07c5731faca0766d1e28798014"
1380 },
1381 {
1382 "name": "res2.0.blocks.2.conv2.weight",
1383 "shape": [
1384 20,
1385 20,
1386 7
1387 ],
1388 "dtype": "float16",
1389 "offset_bytes": 2599968,
1390 "nbytes": 5600,
1391 "sha256": "012b70b2d013782a3ed5b3d121561b3388abd008cdbe96143dd2e043c3088f93"
1392 },
1393 {
1394 "name": "up0.bias",
1395 "shape": [
1396 80
1397 ],
1398 "dtype": "float16",
1399 "offset_bytes": 2605568,
1400 "nbytes": 160,
1401 "sha256": "bd5b6a9df34632802f1a7e1580df1bb0b1c1b6ad1efe0144bd8aa54ebea5b8ef"
1402 },
1403 {
1404 "name": "up0.weight",
1405 "shape": [
1406 160,
1407 80,
1408 16
1409 ],
1410 "dtype": "float16",
1411 "offset_bytes": 2605728,
1412 "nbytes": 409600,
1413 "sha256": "445462548d3c72de7f93aca94ab9f4a710c4f817214ee026e8d1235eb4a16214"
1414 },
1415 {
1416 "name": "up1.bias",
1417 "shape": [
1418 40
1419 ],
1420 "dtype": "float16",
1421 "offset_bytes": 3015328,
1422 "nbytes": 80,
1423 "sha256": "3288de698f44df2cfa12f1a531f805a054e27a3d71fa280eeccb1e28f57fcff5"
1424 },
1425 {
1426 "name": "up1.weight",
1427 "shape": [
1428 80,
1429 40,
1430 16
1431 ],
1432 "dtype": "float16",
1433 "offset_bytes": 3015408,
1434 "nbytes": 102400,
1435 "sha256": "81e89b042141d0f0cd0a271effbd53f6eb87bf94e78ede35b45d1d8f35bc0b5e"
1436 },
1437 {
1438 "name": "up2.bias",
1439 "shape": [
1440 20
1441 ],
1442 "dtype": "float16",
1443 "offset_bytes": 3117808,
1444 "nbytes": 40,
1445 "sha256": "3b1cf2d14a46b08d7e092f7447f1b9ef5c485d1f5a55f319d27db13bee57e461"
1446 },
1447 {
1448 "name": "up2.weight",
1449 "shape": [
1450 40,
1451 20,
1452 8
1453 ],
1454 "dtype": "float16",
1455 "offset_bytes": 3117848,
1456 "nbytes": 12800,
1457 "sha256": "0dc984646ba5737de8ffba402a9503932caad7a8c8ee9f024986a79a402161da"
1458 }
1459 ]
1460 }
1461 },
1462 "runtime_ops": [
1463 "embedding_lookup",
1464 "conv1d_same",
1465 "conv1d_1x1",
1466 "conv_transpose1d",
1467 "silu",
1468 "leaky_relu",
1469 "tanh",
1470 "residual_add",
1471 "repeat_interleave",
1472 "masked_conv1d",
1473 "depthwise_conv1d",
1474 "low_rank_linear_or_conv1d"
1475 ],
1476 "runtime_boundary": {
1477 "neural_stack_self_contained": true,
1478 "arbitrary_text_self_contained": false,
1479 "accepted_runtime_input": "Piper phoneme ID sequence matching piper-phoneme-config.json",
1480 "output": "mono float waveform at manifest.sample_rate"
1481 },
1482 "frontend": {
1483 "status": "not fully self-contained for arbitrary text yet",
1484 "input_contract": "runtime may accept Piper phoneme ID sequences directly",
1485 "included_config": "piper-phoneme-config.json",
1486 "piper_phoneme_type": "espeak",
1487 "piper_espeak_voice": "de",
1488 "phoneme_id_map_entries": 152,
1489 "next_work": "embed or rewrite the Piper/eSpeak-compatible frontend, or keep a strict phoneme-ID API"
1490 },
1491 "size_budget": {
1492 "target_max_bytes": 4000000,
1493 "target_max_mib": 3.814697265625,
1494 "actual_model_asset_bytes": 3135467,
1495 "target_max_parameters": 1500000,
1496 "under_parameter_budget": false,
1497 "under_4mb_asset_budget": true
1498 },
1499 "limitations": [
1500 "This is a raw neural asset package, not a compiled standalone runtime binary.",
1501 "No PyTorch checkpoint, ONNX graph, or training state is included.",
1502 "Arbitrary text still requires a Piper/eSpeak-compatible frontend outside this package."
1503 ]
1504 }
1505