backup: snapshot (--force; 175 groups, 32939 files, 195.56 GB)

README.md

@@ -2,15 +2,16 @@
 license: apache-2.0
 language:
 - en
-pretty_name: mytts-en-data
+pretty_name: "mytts-en-v1-data: training data of the mytts-en-v1 English TTS"
 task_categories:
 - text-to-speech
 tags:
 - speech
 - tts
+- synthetic-speech
 - dacvae
 - latents
-- synthetic-speech
+- repa
 size_categories:
 - 1M<n<10M
 source_datasets:
@@ -18,94 +19,243 @@ source_datasets:
 viewer: false
 ---
 
-# mytts-en-data
+# mytts-en-v1-data
+
+**The prepared training data of the English text-to-speech model [VoiceHub/mytts-en-v1](https://huggingface.co/VoiceHub/mytts-en-v1).** Every clip of [SynDataLab-EN/echo-clones-4m-en](https://huggingface.co/datasets/SynDataLab-EN/echo-clones-4m-en) (about 3.98 M synthetic
+English speech clips of 4,000 voices) turned into the compact numbers the model learns from, plus the checks used to pick
+the good clips.
 
-Derived training data of the English flow-matching TTS of the dacvae-next project (`mytts`, a DiT on Semantic-DACVAE
-latents), made from [SynDataLab-EN/echo-clones-4m-en](https://huggingface.co/datasets/SynDataLab-EN/echo-clones-4m-en)
-(about 3.98 M EchoTTS clips, about 11,150 h, 4,000 synthetic voices). It holds no raw training audio: per clip, the
-codec latents, the REPA targets of an SSL teacher and ASR / silence annotations, plus per-group manifests that point back
-to the source rows. It is also the project's backup: `scripts/hub_backup.py` uploads every committed ingest group and
-verifies each file against the Hub's metadata (size, plus the LFS sha256 or, for small files, the git blob sha1); before
-each status update it checks that every file of `backup/manifest.jsonl` is on the Hub with that size and hash. The LFS
-sha256 is the Hub's record of the upload, not a re-read of the stored bytes: `restore` re-hashes what it downloads. The
-backup runs every 50 ingest groups; the status below is updated at every pass.
+| | |
+|---|---|
+| Model trained on it | [VoiceHub/mytts-en-v1](https://huggingface.co/VoiceHub/mytts-en-v1) |
+| Made from | [SynDataLab-EN/echo-clones-4m-en](https://huggingface.co/datasets/SynDataLab-EN/echo-clones-4m-en) (revision `aebc7c787c64`) |
+| Status | **being processed**: 2,795 of 7,967 source files (35.1 %), updated 2026-10-01 00:23:26 UTC |
+| Size now | 32,939 files, 195.6 GB (about 525.5 GB when complete) |
+| Licence | apache-2.0 (see [Provenance and licence](#provenance-and-licence)) |
+| Code | [kadirnar/dacvae-next](https://github.com/kadirnar/dacvae-next/tree/roadmap/en-echo) (private at the moment) |
 
-## Status
+## What is in it
 
-Last pass: 2026-09-30T19:42:54Z (53 new committed groups (>= 50)), code `350d3e0`.
-32,549 files, 156.0 GB, 136 ingest groups.
+For every clip the model needs three things. The first two come at 25 frames per second, so frame *i* of the latents lines
+up with frame *i* of the REPA targets:
 
-| ingest | progress | groups | stream |
+| feature | in plain words | exact form | folder |
 |---|---|---|---|
-| `echo_en` (tts_en) | 2,171 / 7,967 source files | 136 groups committed | streaming |
+| **Codec latents** (what the model learns to generate) | a compressed form of the audio; the codec's decoder turns it back into sound | [Semantic-DACVAE](https://huggingface.co/Aratako/Semantic-DACVAE-Japanese) posterior mean, 25 Hz x 128 numbers, float16 (codec `Aratako/Semantic-DACVAE-Japanese@737fbad2a679f880`) | `tts_en/latents/` |
+| **REPA targets** (a training aid) | what a large speech-understanding model hears in each frame; the TTS model is nudged to match it, which makes it learn faster | [w2v-BERT 2.0](https://huggingface.co/facebook/w2v-bert-2.0) layer 16, reduced from 1,024 to 128 numbers with PCA (`tts_en/ssl/pca.npz`), float16 | `tts_en/ssl/` |
+| **Text and quality checks** | the sentence each clip says, what a speech recognizer heard, where the silences are; used to drop bad clips | the source text (`.meta.parquet`, manifests); [Parakeet CTC 1.1B](https://huggingface.co/nvidia/parakeet-ctc-1.1b) transcript + CER / WER / NLL of every clip; Whisper-large-v3-turbo on a 2 % sample; silence at the edges, the longest pause and whether the clip hit EchoTTS's length cap | `tts_en/annotations/` |
 
-| root | files | size |
-|---|---|---|
-| `eval/en/` | 30,769 | 9.8 GB |
-| `tts_en/` | 1,713 | 138.8 GB |
-| `tts_en_hubl18/` | 25 | 2.5 GB |
-| `tts_en_hubl24/` | 25 | 2.5 GB |
-| `tts_en_hubl24_fp32/` | 17 | 2.5 GB |
+Also here: the original audio of the 236 **held-out voices** and the English **evaluation sets** (both evaluation only,
+see [below](#held-out-and-evaluation-only-parts)).
 
-## Contents
+## What it is not
 
-| path | what |
-|---|---|
-| `tts_en/latents/<source>/shard-<tag>-NNNNNN.{bin,idx.json,meta.parquet}` | DACVAE latents: `Aratako/Semantic-DACVAE-Japanese@737fbad2a679f880` posterior mean, 25 fps, 128 dims, float16. `meta.parquet`: text, speaker, duration, source id, bandwidth, licence per clip |
-| `tts_en/ssl/<source>/shard-<tag>-t0-NNNNNN.{bin,idx.json}` | REPA targets: `facebook/w2v-bert-2.0` layer 16, PCA-projected to 128 dims with `tts_en/ssl/pca.npz` (pca_sha `96b1b619c73f5a2b`), float16, aligned to the latent frames |
-| `tts_en/annotations/{asr2,asr,edges,silence}/<source>/part-<tag>.parquet` | `asr2`: Parakeet-CTC-1.1B transcript, CER / WER / NLL (every clip); `asr`: Whisper-large-v3-turbo (held-out clips and a 2 % key-hash sample); `edges`: leading / trailing silence; `silence`: longest silence run and the generation cap |
-| `tts_en/manifests/<source>/*.jsonl` | one line per clip and group: id, text, speaker, licence and the source row `hf://datasets/<repo>@<revision>/<file>#<row>` (held-out clips: their kept audio file) |
-| `tts_en/ingest/<source>.json`, `.ledger.jsonl` | the ingest state (run spec, committed source files, held-out counts) and its per-group ledger, as of the last pass |
-| `tts_en/catalogs/`, `latent_stats.json`, `splits.json`, `*.txt` | catalog snapshots, frozen latent statistics, file lists and held-out voice lists |
-| `tts_en_hubl18/`, `tts_en_hubl24/`, `tts_en_hubl24_fp32/` | HuBERT-large teacher roots of the Tier-1 teacher screen (EN-25): SSL targets of the 75 Tier-1 files and their catalogs |
-| `eval/en/` | English eval sets: echo held-out prompts / pairs and the kept held-out echo audio (`echo_heldout_audio/`), Seed-TTS test-en / zh, LibriSpeech(-PC) test-clean / dev-clean, AMI, validation gate audio, voice embeddings |
-| `backup/manifest.jsonl` | every file: path, size, sha256, group, local path, commit; `backup/status.json`; `backup/symlinks.json` (symlinks of the data roots) |
-
-`<source>` is `echo_en` (training) or `echo_en_heldout` (the 236 held-out voices, at most
-60 clips each, never trained on). `<tag>` = `g<time_ns>-<group>` names one ingest group
-(16 source parquet files, about 8,000 clips).
-
-## Provenance
-
-- Source: `SynDataLab-EN/echo-clones-4m-en` at revision `aebc7c787c64`, streamed by `scripts/ingest_stream.py` (each clip decoded
-  once, the raw files deleted after the coverage gate). Its audio was generated with EchoTTS (`jordand/echo-tts-base`)
-  from the reference voices of `SynData-2/echo-ref-speakers-4k-en` and the texts of `SynData-2/echo-4m-text-en`.
-- Models: codec `Aratako/Semantic-DACVAE-Japanese`, SSL teacher `facebook/w2v-bert-2.0` (and `facebook/hubert-large-ll60k`
-  in the teacher-screen roots), ASR `nvidia/parakeet-ctc-1.1b` and `openai/whisper-large-v3-turbo`.
-- Eval sets: Seed-TTS-eval (`zhaochenyang20/seed-tts-eval` mirror), LibriSpeech / LibriSpeech-PC (`openslr/librispeech_asr`),
-  AMI (`edinburghcstr/ami`), and the held-out voices of echo-clones-4m-en.
-
-## Licence
-
-apache-2.0 (owner decision of 2026-09-29; the licence stated on the echo-clones-4m-en card). The generator of that
-audio, EchoTTS (`jordand/echo-tts-base`), is itself licensed cc-by-nc-sa-4.0 (see `configs/release/en_provenance.yaml`
-in the project); the reference voices and texts state no licence. The eval audio that comes from other corpora keeps its
-own licence and attribution: LibriSpeech and the Seed-TTS-eval mirror are CC BY 4.0, the AMI Meeting Corpus is CC BY 4.0
-(University of Edinburgh / AMI consortium). The models named above have their own licences.
-
-## Reading the shards
+- **Not audio to listen to.** Latents are numbers. To hear the speech: play the source clips in
+  [the source's dataset viewer](https://huggingface.co/datasets/SynDataLab-EN/echo-clones-4m-en/viewer), listen to the model's samples on [VoiceHub/mytts-en-v1](https://huggingface.co/VoiceHub/mytts-en-v1), or decode
+  latents with the codec ([Hear a clip](#hear-a-clip)). The only audio here is the held-out voices' clips
+  ([`eval/en/echo_heldout_audio/`](https://huggingface.co/datasets/VoiceHub/mytts-en-v1-data/tree/main/eval/en/echo_heldout_audio): open a `.wav` to play it) and the
+  evaluation sets.
+- **Not the source dataset.** The original audio and texts are [SynDataLab-EN/echo-clones-4m-en](https://huggingface.co/datasets/SynDataLab-EN/echo-clones-4m-en); every clip here points back to its source row
+  (`tts_en/manifests/`, and `source` / `id` in each `.meta.parquet`).
+- **Not the training list by itself.** It holds every clip. The clips a model trains on (after the quality filters) are
+  listed in a catalog under `tts_en/catalogs/` ([The catalog](#the-catalog-the-exact-training-list)).
+- **Not usable with another codec.** The latents belong to this exact codec version; the REPA targets to `tts_en/ssl/pca.npz`.
+- **No real recordings in the training data.** All of the training speech is synthetic (EchoTTS voice clones). Only the
+  evaluation sets under `eval/en/` hold real recordings (LibriSpeech, AMI, the Common Voice prompts of Seed-TTS-eval).
+
+## Size and progress
+
+| | now | when complete |
+|---|---:|---:|
+| Source files (parquet) processed | 2,795 / 7,967 (35.1 %) | 7,967 |
+| Source clips read | 1,397,250 | about 3.98 M |
+| Groups uploaded (16 source files each) | 175 | 498 |
+| Held-out clips kept (236 voices, at most 60 each) | 14,160 | at most 14,160 |
+| Files in this repo | 32,939 | - |
+| Size | 195.6 GB | about 525.5 GB |
+
+The data is still being processed ([EN-12](https://github.com/kadirnar/dacvae-next/issues/36)); this page is written again after every 50 new groups.
+Last update: 2026-10-01 00:23:26 UTC (--force).
+
+## Layout
+
+| folder | what is inside | used for | files | size |
+|---|---|---|---:|---:|
+| `tts_en/latents/echo_en/` | codec latents of every training clip: `shard-<tag>-NNNNNN.bin` (float16, 25 frames/s x 128 numbers), `.idx.json` (where each clip starts and how long it is), `.meta.parquet` (text, voice, duration, source row of each clip) | training | 525 | 87.9 GB |
+| `tts_en/ssl/echo_en/` | REPA targets of every training clip: w2v-BERT 2.0 layer 16, PCA to 128 numbers, float16, one row per latent frame (`shard-<tag>-t0-NNNNNN.{bin,idx.json}`) | training | 350 | 87.7 GB |
+| `tts_en/ssl/` | `pca.npz`: the PCA that turns w2v-BERT's 1,024 numbers into 128; `info.json` | training | 2 | 529.8 kB |
+| `tts_en/annotations/asr2/echo_en/` | Parakeet CTC 1.1B transcript of every clip with CER / WER / NLL against its text (`part-<tag>.parquet`) | training (clip filter) | 175 | 160.6 MB |
+| `tts_en/annotations/asr/echo_en/` | Whisper-large-v3-turbo transcript of a 2 % sample: a cross-check of the Parakeet scores | analysis | 175 | 8.0 MB |
+| `tts_en/annotations/edges/echo_en/` | silence at the start and the end of every clip | training (clip edges) | 175 | 69.0 MB |
+| `tts_en/annotations/silence/echo_en/` | longest pause inside every clip, the generation length cap and whether the clip reached it (EchoTTS failure clips) | training (clip filter) | 175 | 37.0 MB |
+| `tts_en/manifests/` | one JSON line per clip: id, text, voice, licence and its source row `hf://datasets/<source>@<revision>/<file>#<row>` (held-out clips: their kept audio file) | provenance | 175 | 532.6 MB |
+| `tts_en/catalogs/` | catalogs: the filtered clip lists a run trains on (`<name>/catalog.parquet` + `version.json`, build log, leak check) | training | 8 | 9.8 MB |
+| `tts_en/` | `latent_stats.json` (the frozen latent normalization of the model), `splits.json`, file lists and held-out voice lists | training | 7 | 213.7 kB |
+| `tts_en/ingest/` | the ingest's state (run spec, finished source files) and ledger (one line per group), as of the last upload | bookkeeping | 2 | 646.9 kB |
+| `tts_en/manifests_local/` | the Tier-1 screening subset's manifest (local raw paths; provenance only) | bookkeeping | 1 | 12.9 MB |
+| `tts_en/latents/echo_en_heldout/` | codec latents of the held-out voices' clips | **held out: evaluation only** | 111 | 946.5 MB |
+| `tts_en/ssl/echo_en_heldout/` | REPA targets of the held-out voices' clips | **held out: evaluation only** | 74 | 944.3 MB |
+| `tts_en/annotations/*/echo_en_heldout/` | the four annotations of the held-out clips (Whisper on every one) | **held out: evaluation only** | 148 | 5.0 MB |
+| `eval/en/echo_heldout_audio/` | the original audio of up to 60 clips per held-out voice (`<voice>/<id>.wav`): the prompts and references of the echo tests | **held out: evaluation only** | 14,160 | 5.5 GB |
+| `eval/en/echo_heldout/` | the echo tests of the held-out voices: echo-dev / echo-test lists, prompt draws, prompt pairs (with their prompt audio) | evaluation only | 212 | 111.4 MB |
+| `eval/en/echo_heldout_ann/` | annotations and manifests of the kept held-out audio | evaluation only | 7 | 1.6 MB |
+| `eval/en/seedtts_testset/` | Seed-TTS-eval test sets (`en`: Common Voice prompts; `zh`), CC BY 4.0 | evaluation only | 5,134 | 1.2 GB |
+| `eval/en/LibriSpeech/` | LibriSpeech test-clean / dev-clean audio (the LibriSpeech-PC test), CC BY 4.0 | evaluation only | 5,512 | 727.8 MB |
+| `eval/en/ami/` | AMI Meeting Corpus clips (`ihm` headset, `sdm` distant microphone), CC BY 4.0 | evaluation only | 996 | 336.5 MB |
+| `eval/en/validation/` | validation gates: evaluation runs (scores and audio) of the real test audio and of its codec resynthesis; they check the evaluation pipeline | evaluation only | 4,732 | 1.8 GB |
+| `eval/en/voice_emb/` | speaker embeddings of eval prompts and references | evaluation only | 8 | 83.2 MB |
+| `eval/en/` | evaluation lists (`*.jsonl`: Seed-TTS test-en dev / test, LibriSpeech-PC, AMI) and the protect list (test sentences no training catalog may hold) | evaluation only | 8 | 3.0 MB |
+| `tts_en_hubl18/` | HuBERT-large layer 18 REPA targets of the screening subset (EN-25 teacher screen) | experiment (not used by v1) | 25 | 2.5 GB |
+| `tts_en_hubl24/` | HuBERT-large layer 24 REPA targets of the screening subset (EN-25) | experiment (not used by v1) | 25 | 2.5 GB |
+| `tts_en_hubl24_fp32/` | the same, computed in fp32 (EN-25) | experiment (not used by v1) | 17 | 2.5 GB |
+| `backup/` | `manifest.jsonl` (every file: size, sha256, group), `status.json` (the numbers on this page), `symlinks.json` | bookkeeping | 3 | - |
+
+`<tag>` = `g<time_ns>-<group>` names one ingest group (16 source files, about 8,000 clips). Every file of a group carries
+its tag, so a group can be downloaded on its own. `echo_en` is the training source, `echo_en_heldout` the held-out voices.
+
+## Held-out and evaluation-only parts
+
+Do **not** train on these if you want results that compare with ours:
+
+- **Held-out voices**: every `echo_en_heldout` folder and `eval/en/echo_heldout_audio/`. 236 of the 4,000 voices
+  ([list](https://huggingface.co/datasets/VoiceHub/mytts-en-v1-data/blob/main/tts_en/heldout_speakers.txt)) never enter training; at most 60 clips of each are kept for the echo
+  tests (EN-08).
+- **Evaluation sets**: everything under `eval/en/`. Seed-TTS test-en, LibriSpeech(-PC) and AMI come from other corpora
+  and keep their own licences.
+- **Echo-unseen voices**: 177 more voices ([list](https://github.com/kadirnar/dacvae-next/blob/roadmap/en-echo/configs/eval/en/echo_unseen_speakers.txt), EN-56) are inside `echo_en`
+  but no training catalog holds them (the catalogs already leave them out, and the trainer refuses a catalog that does not).
+
+## How to load
+
+### Download
+
+All of it is large (195.6 GB now). One group is enough to try the data:
 
 ```python
-import json, numpy as np
-idx = json.load(open("tts_en/latents/echo_en/shard-<tag>-000000.idx.json"))
-mm = np.memmap("tts_en/latents/echo_en/shard-<tag>-000000.bin", dtype=idx["dtype"], mode="r").reshape(-1, idx["dim"])
-i = 0  # item i: key idx["keys"][i]
-z = np.asarray(mm[idx["offsets"][i] : idx["offsets"][i] + idx["n_frames"][i]], np.float32)  # [frames, 128] at 25 fps
+from huggingface_hub import snapshot_download
+
+tag = "g1790656525369654176-00001"  # any group; backup/status.json lists them all
+local = snapshot_download("VoiceHub/mytts-en-v1-data", repo_type="dataset", local_dir="mytts-en-v1-data",
+                          allow_patterns=[f"tts_en/*/echo_en/*{tag}*", "tts_en/ssl/pca.npz", "tts_en/latent_stats.json"])
 ```
 
-With the project code: `mytts.data.shards.ShardReader`, `build_catalog` and `TTSDataset` read the restored layout as is.
+Everything: `allow_patterns=["tts_en/*"]` (the training data) or no filter, or the verified restore below.
 
-## Restore
+### Read one clip (numpy only)
+
+```python
+import json
+
+import numpy as np
+import pyarrow.parquet as pq
+
+shard = f"{local}/tts_en/latents/echo_en/shard-{tag}-000000"
+idx = json.load(open(shard + ".idx.json"))  # keys, offsets, n_frames, dim, dtype
+meta = pq.read_table(shard + ".meta.parquet").to_pylist()  # one row per clip: key, text, speaker, duration, ...
+lat = np.memmap(shard + ".bin", dtype=idx["dtype"], mode="r").reshape(-1, idx["dim"])
+i = 0
+z = np.asarray(lat[idx["offsets"][i] : idx["offsets"][i] + idx["n_frames"][i]], np.float32)  # [frames, 128], 25 frames per second
+print(meta[i]["text"], z.shape)
+```
+
+The REPA targets of the same clip: `tts_en/ssl/echo_en/shard-<tag>-t0-000000.{bin,idx.json}`, same key, same frames.
+
+### With the project code (`mytts`)
+
+```python
+from mytts.config import DataConfig
+from mytts.data.catalog import CatalogFilters, build_catalog
+from mytts.data.dataset import Catalog, TTSDataset
+
+root = f"{local}/tts_en"
+cat_dir, version = build_catalog(root, "my_subset", CatalogFilters(), num_workers=2, verbose=False)  # index what you downloaded
+cat = Catalog(cat_dir, split=None)
+ds = TTSDataset(cat, root, DataConfig(use_ssl=True), latent_dim=128, ssl_dim=(version.filters or {}).get("ssl_dim"), train=False)
+item = ds[0]
+print(cat.get(0, "text"), item["latent"].shape, item["ssl"].shape)  # [frames, 128] each
+```
+
+### The catalog: the exact training list
+
+`tts_en/catalogs/<name>/catalog.parquet` has one row per clip that passed the quality filters: key, text, voice, length,
+where its latents and REPA targets are (paths relative to `tts_en/`) and its split (`train` / `val`); `version.json` records
+the filters and the hours. `echo_en_v1` is the catalog of the main run (uploaded after the stream completes):
+Parakeet CER <= 0.3 (rescored with the en-v2 text normalization), no long pause (> 6 s), no clip stuck at the length cap with
+silence, the echo-unseen voices left out, 0.1 % held back as `val`. `en_t1` is the 88.8 h screening subset.
+Catalogs here: `en_t1`, `en_t1_x56`. To train on exactly what the model saw, download all of `tts_en/`, then
+`Catalog(f"{root}/catalogs/echo_en_v1", split="train")`.
+
+### Hear a clip
+
+```python
+import soundfile as sf
+import torch
+
+from mytts.codec import DACVAECodec  # needs pip install git+https://github.com/facebookresearch/dacvae
+
+codec = DACVAECodec.load("Aratako/Semantic-DACVAE-Japanese@737fbad2a679f880")  # the codec the latents were made with
+wav = codec.decode(torch.from_numpy(z)[None])[0].numpy()  # z from 'Read one clip'; 48 kHz
+sf.write("clip.wav", wav, 48000)
+```
+
+## How it was made
+
+1. **Stream** ([`scripts/ingest_stream.py`](https://github.com/kadirnar/dacvae-next/blob/roadmap/en-echo/scripts/ingest_stream.py), [EN-12, issue #36](https://github.com/kadirnar/dacvae-next/issues/36)): the source's
+   7,967 parquet files (1.51 TB) are downloaded 16 at a time (one *group*), in a fixed shuffled order.
+2. **Decode** each clip once. Clips of the 236 held-out voices go to `echo_en_heldout` (the first 60 of each voice,
+   with their audio kept); all others to `echo_en`.
+3. **Encode** on the GPU: the codec latents (48 kHz audio, exact fp32 encoder, stored as float16) and the w2v-BERT 2.0
+   layer-16 features, projected to 128 numbers with a PCA fitted once (the REPA targets).
+4. **Annotate**: Parakeet CTC transcript and CER / WER / NLL for every clip, Whisper on the held-out clips and a 2 %
+   sample, silence at the clip edges, the longest pause and the length cap.
+5. **Check and commit**: a group counts only when every output covers its clips (the coverage gate); then it gets its
+   line in the ledger (`tts_en/ingest/`) and its downloaded source files are deleted.
+6. **Upload** ([`scripts/hub_backup.py`](https://github.com/kadirnar/dacvae-next/blob/roadmap/en-echo/scripts/hub_backup.py)): after every 50 committed groups the new files come here,
+   each checked against the Hub's own record of it (size and sha256).
+7. **Curate** (after the stream, `scripts/recipes/en_m4.sh catalog`): the filters above make the training catalog `echo_en_v1`.
+
+## Provenance and licence
+
+- **This dataset: apache-2.0** (owner decision, 2026-09-29).
+- **Source**: [SynDataLab-EN/echo-clones-4m-en](https://huggingface.co/datasets/SynDataLab-EN/echo-clones-4m-en); its card states **apache-2.0**. Its speech was generated with EchoTTS
+  ([jordand/echo-tts-base](https://huggingface.co/jordand/echo-tts-base), **cc-by-nc-sa-4.0** on its card) from the reference voices of
+  [SynData-2/echo-ref-speakers-4k-en](https://huggingface.co/datasets/SynData-2/echo-ref-speakers-4k-en) and the texts of
+  [SynData-2/echo-4m-text-en](https://huggingface.co/datasets/SynData-2/echo-4m-text-en) (neither card states a licence). Check these terms
+  yourself before any commercial use.
+- **Models used to make the features** (each under its own licence; none of their weights are in this repo):
+  Aratako/Semantic-DACVAE-Japanese (MIT), facebook/w2v-bert-2.0 (MIT), nvidia/parakeet-ctc-1.1b (CC BY 4.0),
+  openai/whisper-large-v3-turbo (MIT); facebook/hubert-large-ll60k (Apache-2.0) for the `tts_en_hubl*` roots only.
+- **Evaluation audio** under `eval/en/` keeps its own licence and attribution: Seed-TTS-eval (CC BY 4.0, Common Voice
+  prompts), LibriSpeech (CC BY 4.0), AMI Meeting Corpus (CC BY 4.0, University of Edinburgh / AMI consortium).
+
+## Backup and restore (maintainers)
+
+This repo is also the project's backup of its data roots. `scripts/hub_backup.py` uploads only what the ingest has
+committed, checks every file after its commit (size, plus the LFS sha256 or, for small files, the git blob sha1) and, before
+every status update, that every file of `backup/manifest.jsonl` is on the repo with that size and hash. The LFS sha256 is
+the Hub's record of the upload, not a re-read of the stored bytes: `restore` hashes what it downloads.
 
 ```bash
-python scripts/hub_backup.py restore --out /workspace/restore_check --sample-groups 3   # restore test: a random sample, verified
-python scripts/hub_backup.py restore --all --out /workspace/data_restore              # everything, verified, into a scratch root
+python scripts/hub_backup.py restore --repo VoiceHub/mytts-en-v1-data --out /workspace/restore_check --sample-groups 3   # a verified sample
+python scripts/hub_backup.py restore --repo VoiceHub/mytts-en-v1-data --all --out /workspace/data_restore              # everything, verified
 # then, with the ingest and its job_guard stopped and any existing /workspace/data/<root> moved aside:
 for r in tts_en tts_en_hubl18 tts_en_hubl24 tts_en_hubl24_fp32; do mv -T /workspace/data_restore/$r /workspace/data/$r; done
 mkdir -p /workspace/data/eval && mv -T /workspace/data_restore/eval/en /workspace/data/eval/en
 ```
 
 `restore` writes only into an empty directory (`--into-empty-only` / `--overwrite-existing` allow a non-empty one) and
-always refuses the data root of a running ingest (a held `ingest/*.lock`, or a running `ingest_stream.py` / `job_guard.py`
-naming it): it would replace the live ingest state and ledger. The manifests and the ingest state record absolute paths
-under `/workspace/data`, so they hold again after the swap, and the ingest resumes from `tts_en/ingest/echo_en.json`.
+always refuses the data root of a running ingest. The manifests and the ingest state record absolute paths under
+`/workspace/data`, so they hold again after the swap, and the ingest resumes from `tts_en/ingest/echo_en.json`.
+
+## Links
+
+- Model: [VoiceHub/mytts-en-v1](https://huggingface.co/VoiceHub/mytts-en-v1)
+- Source audio and texts: [SynDataLab-EN/echo-clones-4m-en](https://huggingface.co/datasets/SynDataLab-EN/echo-clones-4m-en)
+- Code: [kadirnar/dacvae-next](https://github.com/kadirnar/dacvae-next/tree/roadmap/en-echo) (branch `roadmap/en-echo`): `scripts/ingest_stream.py`, `mytts/data/`,
+  `scripts/hub_backup.py`
+- The project's earlier pages (screening phase, kept as an archive; not needed to use this data):
+  [VoiceHub/mytts-en](https://huggingface.co/VoiceHub/mytts-en) (screening checkpoints),
+  [VoiceHub/mytts-en-samples](https://huggingface.co/datasets/VoiceHub/mytts-en-samples) (their samples),
+  [VoiceHub/mytts-en-ablations](https://huggingface.co/datasets/VoiceHub/mytts-en-ablations) (screening runs and results),
+  [VoiceHub/mytts-en-eval](https://huggingface.co/datasets/VoiceHub/mytts-en-eval) (evaluation lists),
+  [VoiceHub/mytts-en-base](https://huggingface.co/datasets/VoiceHub/mytts-en-base) (the main run's internal training state)

backup/manifest.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:754fbaa492521a8c4ef1717bc5183d0a4fd76250dc221c44d7c7c8392d7547b0
-size 14091135
+oid sha256:d76ba8d01ae34c84b7bf7e7707ddedf99577a4995f049839ad1f4f638ec3b80c
+size 14274307

backup/status.json

@@ -1,12 +1,122 @@
 {
+ "by_folder": {
+  "eval/en/": {
+   "bytes": 3008896,
+   "files": 8
+  },
+  "eval/en/LibriSpeech/": {
+   "bytes": 727801100,
+   "files": 5512
+  },
+  "eval/en/ami/": {
+   "bytes": 336498064,
+   "files": 996
+  },
+  "eval/en/echo_heldout/": {
+   "bytes": 111356939,
+   "files": 212
+  },
+  "eval/en/echo_heldout_ann/": {
+   "bytes": 1561445,
+   "files": 7
+  },
+  "eval/en/echo_heldout_audio/": {
+   "bytes": 5526438476,
+   "files": 14160
+  },
+  "eval/en/seedtts_testset/": {
+   "bytes": 1235334210,
+   "files": 5134
+  },
+  "eval/en/validation/": {
+   "bytes": 1764667161,
+   "files": 4732
+  },
+  "eval/en/voice_emb/": {
+   "bytes": 83155012,
+   "files": 8
+  },
+  "tts_en/": {
+   "bytes": 213710,
+   "files": 7
+  },
+  "tts_en/annotations/*/echo_en_heldout/": {
+   "bytes": 5038028,
+   "files": 148
+  },
+  "tts_en/annotations/asr/echo_en/": {
+   "bytes": 8044049,
+   "files": 175
+  },
+  "tts_en/annotations/asr2/echo_en/": {
+   "bytes": 160610439,
+   "files": 175
+  },
+  "tts_en/annotations/edges/echo_en/": {
+   "bytes": 69036333,
+   "files": 175
+  },
+  "tts_en/annotations/silence/echo_en/": {
+   "bytes": 36970509,
+   "files": 175
+  },
+  "tts_en/catalogs/": {
+   "bytes": 9784799,
+   "files": 8
+  },
+  "tts_en/ingest/": {
+   "bytes": 646857,
+   "files": 2
+  },
+  "tts_en/latents/echo_en/": {
+   "bytes": 87877435766,
+   "files": 525
+  },
+  "tts_en/latents/echo_en_heldout/": {
+   "bytes": 946491003,
+   "files": 111
+  },
+  "tts_en/manifests/": {
+   "bytes": 532640363,
+   "files": 175
+  },
+  "tts_en/manifests_local/": {
+   "bytes": 12892342,
+   "files": 1
+  },
+  "tts_en/ssl/": {
+   "bytes": 529822,
+   "files": 2
+  },
+  "tts_en/ssl/echo_en/": {
+   "bytes": 87701452247,
+   "files": 350
+  },
+  "tts_en/ssl/echo_en_heldout/": {
+   "bytes": 944333793,
+   "files": 74
+  },
+  "tts_en_hubl18/": {
+   "bytes": 2488799323,
+   "files": 25
+  },
+  "tts_en_hubl24/": {
+   "bytes": 2488799324,
+   "files": 25
+  },
+  "tts_en_hubl24_fp32/": {
+   "bytes": 2488810713,
+   "files": 17
+  }
+ },
  "by_root": {
   "eval/en": {
    "bytes": 9789821303,
    "files": 30769
   },
   "tts_en": {
-   "bytes": 138753808438,
-   "files": 1713
+   "bytes": 178306120060,
+   "files": 2103
   },
   "tts_en_hubl18": {
    "bytes": 2488799323,
@@ -23,7 +133,7 @@
  },
  "by_tier": {
   "snapshot": {
-   "bytes": 504346,
+   "bytes": 646857,
    "files": 2
   },
   "static": {
@@ -31,12 +141,16 @@
    "files": 16694
   },
   "stream": {
-   "bytes": 144256321895,
-   "files": 15853
+   "bytes": 183808491006,
+   "files": 16243
   }
  },
- "bytes": 156010039101,
- "code": "350d3e0",
+ "bytes": 195562350723,
+ "catalogs": [
+  "en_t1",
+  "en_t1_x56"
+ ],
+ "code": "46e3e71",
  "excludes": [
   "*.tmp",
   "*.lock",
@@ -45,7 +159,7 @@
   "*/__pycache__/*",
   "eval/en/wavlm_large_finetune.pth"
  ],
- "files": 32549,
+ "files": 32939,
  "groups": [
   "g1790656525369654176-00001",
   "g1790660204553794962-00001",
@@ -182,7 +296,46 @@
   "g1790793988324182523-00077",
   "g1790794409390069260-00078",
   "g1790794824606348797-00079",
-  "g1790795253423643335-00080"
+  "g1790795253423643335-00080",
+  "g1790795674907814590-00081",
+  "g1790796086663306331-00082",
+  "g1790796515212226687-00083",
+  "g1790796939145621360-00084",
+  "g1790797374990856046-00085",
+  "g1790797793607907173-00086",
+  "g1790798216366059599-00087",
+  "g1790798646226334116-00088",
+  "g1790799081591774935-00089",
+  "g1790799507316476385-00090",
+  "g1790799916976564320-00091",
+  "g1790800332868688638-00092",
+  "g1790800762842797072-00093",
+  "g1790801171860199664-00094",
+  "g1790801596739530817-00095",
+  "g1790802023623748965-00096",
+  "g1790802457100038480-00097",
+  "g1790802898650332673-00098",
+  "g1790803312132765599-00099",
+  "g1790803734610677405-00100",
+  "g1790804155136641606-00101",
+  "g1790804582331808779-00102",
+  "g1790804983262340791-00103",
+  "g1790805394925553412-00104",
+  "g1790805816445752126-00105",
+  "g1790806239126363055-00106",
+  "g1790806658236550275-00107",
+  "g1790807090211887955-00108",
+  "g1790807525369930532-00109",
+  "g1790807955985732463-00110",
+  "g1790808374365825905-00111",
+  "g1790808817520484970-00112",
+  "g1790809248786950639-00113",
+  "g1790809673774631850-00114",
+  "g1790810100801439072-00115",
+  "g1790810547718308685-00116",
+  "g1790810962071312713-00117",
+  "g1790811390355893879-00118",
+  "g1790811821693925778-00119"
  ],
  "includes": [],
  "ingests": {
@@ -193,10 +346,12 @@
     "edges",
     "silence"
    ],
+   "asr2_model": "nvidia/parakeet-ctc-1.1b",
    "codec_id": "Aratako/Semantic-DACVAE-Japanese@737fbad2a679f880",
    "complete": false,
-   "done_files": 2171,
-   "groups_committed": 136,
+   "done_files": 2795,
+   "group_size": 16,
+   "groups_committed": 175,
    "held_out_clips": 14160,
    "holdout": {
     "list_sha": "38848d0bea53f777",
@@ -204,14 +359,14 @@
     "n_ids": 236,
     "source": "echo_en_heldout"
    },
-   "ledger_lines": 137,
-   "ledger_sha256": "e782f5b4a91803625171d669caa6ca2d31da8acc13bd88aac3d3ea52cb6acfa2",
+   "ledger_lines": 176,
+   "ledger_sha256": "841480f5b10a84ded1eb4f5a5cd08edcd768a8cfa63301b7c07e0c0d210d5230",
    "n_files": 7967,
    "repo": "SynDataLab-EN/echo-clones-4m-en",
    "revision": "aebc7c787c64e918b0130c04d3878f17a602c1df",
    "root": "tts_en",
-   "rows_seen": 1085250,
-   "state_sha256": "27d335876918df8985faae85936a2323804f8bb553061392e4f7036d0f27fafc",
+   "rows_seen": 1397250,
+   "state_sha256": "1236746c32138e3a7962a2143fd2c20ed6617e0737dff1bd8045a98336eacb9a",
    "teachers": [
     {
      "dim": 128,
@@ -219,25 +374,28 @@
      "model": "facebook/w2v-bert-2.0",
      "pca_sha": "96b1b619c73f5a2b"
     }
-   ]
+   ],
+   "whisper_frac": 0.02,
+   "whisper_model": "openai/whisper-large-v3-turbo"
   }
  },
  "partial": false,
  "pass": {
-  "commits": 15,
+  "commits": 11,
   "failures": 0,
   "lost_files": [],
   "n_lost_files": 0,
-  "reason": "53 new committed groups (>= 50)",
+  "reason": "--force",
   "repaired_on_hub": 0,
-  "seconds": 1482.2,
-  "upload_mb_per_s": 38.2,
-  "upload_s": 1399.3,
-  "uploaded_bytes": 53433468018,
-  "uploaded_files": 530
+  "seconds": 1677.5,
+  "upload_mb_per_s": 24.6,
+  "upload_s": 1609.6,
+  "uploaded_bytes": 39552169111,
+  "uploaded_files": 390
  },
- "reason": "53 new committed groups (>= 50)",
- "repo": "VoiceHub/mytts-en-data",
+ "reason": "--force",
+ "renamed_from": "VoiceHub/mytts-en-data",
+ "repo": "VoiceHub/mytts-en-v1-data",
  "repo_type": "dataset",
  "roots": {
   "eval/en": "/workspace/data/eval/en",
@@ -247,5 +405,5 @@
   "tts_en_hubl24_fp32": "/workspace/data/tts_en_hubl24_fp32"
  },
  "symlinks": 9,
- "updated": "2026-09-30T19:42:54Z"
+ "updated": "2026-10-01T00:23:26Z"
 }