ACappellaSet model release

.gitattributes

@@ -0,0 +1,35 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text

LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 ACappellaSet contributors
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

README.md

@@ -0,0 +1,100 @@
+---
+license: cc-by-nc-4.0
+tags:
+- audio
+- music-source-separation
+- source-separation
+- htdemucs
+- acappella
+inference: false
+---
+
+# ACappellaSet models
+
+Model weights and release metadata for a cappella source separation.
+
+**Repository status: private.** The active release contains **29 checkpoints**: three fixed-output models, twelve original part-specific specialists, and fourteen conditioned selections, including two trained with AI-generated singer augmentation. Downloads require repository access.
+**Weight license: [CC BY-NC 4.0](licenses/CC-BY-NC-4.0.txt)**, subject to the upstream notices and third-party rights described in [WEIGHTS_NOTICE.md](WEIGHTS_NOTICE.md).
+
+- **Checkpoint paths, hashes, configurations, and validation selections:** [manifest.json](manifest.json)
+- **Code and installation:** [ACappellaSet on GitHub](https://github.com/pting126/ACappellaSet)
+- **Inference:** [fixed-output and part-specific models](https://github.com/pting126/ACappellaSet/blob/main/docs/inference.md), [shared query-conditioned models](https://github.com/pting126/ACappellaSet/blob/main/docs/conditioned-inference.md).
+- **Training and configurations:** [published code and recipes](https://github.com/pting126/ACappellaSet).
+- **Dataset:** [ACappellaSet](https://huggingface.co/datasets/AcappellaSet/ACappellaSet) is released with Manual Approval for Non-commercial Research Use. Dataset access approval does not grant access to this private model repository, and Dataset Terms do not relicense these model weights.
+
+## Fixed-output and part-specific models
+
+These models use the [HTDemucs-MSS architecture](https://github.com/pting126/ACappellaSet/blob/main/models/demucs4ht.py), with parallel waveform and spectrogram branches coupled by a cross-domain Transformer. Fixed-output models predict six vocal parts simultaneously. Each part-specific model has a two-output head for its designated target and residual, with a separate checkpoint per vocal part; no query input is required.
+
+JC denotes JaCappella; SK denotes natural Seagull-K recordings; JC + SK denotes joint training. Each checkpoint below links to its file.
+
+| Model | Training corpus | Checkpoints |
+| --- | --- | --- |
+| Fixed-output HTDemucs (6 outputs) | JC | [Fixed_6stem_htdemucs_JC](fixed/JC/legacy/Fixed_6stem_htdemucs_JC.pt) |
+| Fixed-output HTDemucs (6 outputs) | SK | [Fixed_6stem_htdemucs_SK](fixed/SK/legacy/Fixed_6stem_htdemucs_SK.pt) |
+| Fixed-output HTDemucs (6 outputs) | JC + SK | [Fixed_6stem_htdemucs_SKnJC](fixed/SKnJC/legacy/Fixed_6stem_htdemucs_SKnJC.pt) |
+| Part-specific HTDemucs (mixture input) | JC | [Soprano](specialists/JC/legacy/soprano/epoch_0162.pt), [Alto](specialists/JC/legacy/alto/epoch_0223.pt), [Tenor](specialists/JC/legacy/tenor/epoch_0119.pt), [Bass](specialists/JC/legacy/bass/epoch_0110.pt), [VP](specialists/JC/legacy/vp/epoch_0045.pt), [Lead](specialists/JC/legacy/lead/epoch_0149.pt) |
+| Part-specific HTDemucs (mixture input) | SK | [Soprano](specialists/SK/legacy/soprano/epoch_0220.pt), [Alto](specialists/SK/legacy/alto/epoch_0237.pt), [Tenor](specialists/SK/legacy/tenor/epoch_0214.pt), [Bass](specialists/SK/legacy/bass/epoch_0182.pt), [VP](specialists/SK/legacy/vp/epoch_0205.pt), [Baritone](specialists/SK/legacy/baritone/epoch_0043.pt) |
+
+## Shared query-conditioned
+
+These models extend [HTDemucs with query-conditioned FiLM layers](https://github.com/pting126/ACappellaSet/blob/main/models/conditioned_core.py) in the waveform and spectrogram branches before the cross-domain Transformer. Queries use a learned vocal-part embedding, a matching fine-tuned ReDimNet2 singer-enrollment encoder, or both. A shared model extracts the requested target and its residual; the full-remix and single-clone models use the singer-conditioned architecture with different training data.
+
+Checkpoint labels show the selected training step and validation criterion. “Pooled” means JC + SK validation.
+
+| Model | Training data | Checkpoint (validation criterion) |
+| --- | --- | --- |
+| Vocal-part-conditioned HTDemucs | SK | [4k](conditioned/role/train-sk-natural/sisdr-v1/no-encoder/step-0004000.pt) (JC inclusive); [198k](conditioned/role/train-sk-natural/sisdr-v1/no-encoder/step-0198000.pt) (SK shared) |
+| Vocal-part-conditioned HTDemucs | JC | [119k](conditioned/role/train-jc-natural/sisdr-v1/no-encoder/step-0119000.pt) (SK inclusive); [141k](conditioned/role/train-jc-natural/sisdr-v1/no-encoder/step-0141000.pt) (JC shared) |
+| Singer-conditioned HTDemucs | SK | [12k](conditioned/enrollment/train-sk-natural/sisdr-v1/redimnet2-finetuned/step-0012000.pt) (JC shared, JC inclusive); [106k](conditioned/enrollment/train-sk-natural/sisdr-v1/redimnet2-finetuned/step-0106000.pt) (pooled shared, pooled inclusive); [179k](conditioned/enrollment/train-sk-natural/sisdr-v1/redimnet2-finetuned/step-0179000.pt) (SK shared) |
+| Vocal-part + singer-conditioned HTDemucs | SK | [7k](conditioned/dual/train-sk-natural/sisdr-v1/redimnet2-finetuned/step-0007000.pt) (JC inclusive) |
+| Singer-conditioned HTDemucs | JC | [184k](conditioned/enrollment/train-jc-natural/sisdr-v1/redimnet2-finetuned/step-0184000.pt) (SK inclusive) |
+| Vocal-part + singer-conditioned HTDemucs | JC | [188k](conditioned/dual/train-jc-natural/sisdr-v1/redimnet2-finetuned/step-0188000.pt) (SK inclusive) |
+| Singer-conditioned HTDemucs | JC + SK | [195k](conditioned/enrollment/train-jc-sk-natural/sisdr-v1/redimnet2-finetuned/step-0195000.pt) (JC shared, pooled shared); [198k](conditioned/enrollment/train-jc-sk-natural/sisdr-v1/redimnet2-finetuned/step-0198000.pt) (SK shared) |
+| Singer-conditioned HTDemucs | SK + full-remix singer augmentation | [131k](conditioned/enrollment/train-sk-full-remix/sisdr-v1/redimnet2-finetuned/step-0131000.pt) (pooled inclusive) |
+| Singer-conditioned HTDemucs | SK + single-clone singer augmentation | [48k](conditioned/enrollment/train-sk-single-clone/sisdr-v1/redimnet2-finetuned/step-0048000.pt) (pooled inclusive) |
+
+Conditioned runs were trained through 200k updates; the released files were selected using validation scores. Training data and validation corpus are separate attributes.
+
+Shared aggregates retained S/A/T/B/VP pairs rather than averaging five part means. Inclusive additionally includes JC Lead or SK Bar. In the machine-readable registry, `combined_*` selection labels mean pooled JC + SK validation.
+
+## Download and inference
+
+```bash
+hf auth login
+hf download AcappellaSet/ACappellaSet-Models --local-dir ./ACappellaSet-Models
+```
+
+To download one checkpoint, add its file path from the tables above after the repository name. Follow the linked GitHub installation and inference guides and pass the downloaded file with `--checkpoint`. The registry records `inference_model_id` and, for shared query-conditioned models, the available `selection_labels`. Released files contain inference tensors only (plus the selected step for query-conditioned models) and load with `weights_only=True`. Checkpoint and configuration hashes are verified before strict loading.
+
+Fixed-output and part-specific models take mixture audio. Shared query-conditioned models additionally use a vocal-part query, singer-enrollment audio, or both, according to their conditioning type. All 29 checkpoints have registered inference adapters. Singer- and dual-conditioned exports include their matching fine-tuned encoder tensors.
+
+### Output conventions
+
+The fixed-output JC + SK model returns `vp`, `bass`, `soprano`, `alto`, `tenor`, and `other`.
+Its `other` output has JC Lead / SK Baritone training semantics; these are not
+independent outputs. Part-specific specialists use residual index 0 and target
+index 1. Shared query-conditioned models use target index 0 and residual index 1. Preserve each checkpoint's native source order and inference settings.
+
+## Layout
+
+```text
+manifest.json
+fixed/{JC,SK,SKnJC}/legacy/
+specialists/{JC,SK}/legacy/{part}/
+conditioned/
+configs/
+```
+
+`model.json` and conditioned step JSON files record checkpoint identity, configuration, and license. `bank.json` maps specialist roles to model IDs. The registry provides released-file SHA-256 hashes. Configurations are included under `configs/`; maintained executable code is on GitHub.
+
+Validation-selected models record the validation corpus, shared/inclusive scope, training cutoff, and selected step. Multiple selectors may refer to the same file. The two AI-augmented entries identify their augmentation datasets explicitly.
+
+## Licenses and citation
+
+Original repository documentation and metadata use the [MIT license](LICENSE).
+All 29 model checkpoints use [CC BY-NC 4.0](licenses/CC-BY-NC-4.0.txt). Preserve attribution and indicate changes when sharing. This project license applies only to rights the licensors can grant; upstream notices and third-party rights remain applicable. See [weight notices](WEIGHTS_NOTICE.md). Dataset audio is separate.
+Upstream architecture attribution is maintained in the GitHub repository's
+[third-party notices](https://github.com/pting126/ACappellaSet/blob/main/THIRD_PARTY_NOTICES.md).
+
+Cite the accompanying paper when a verified citation becomes available; do not infer publication status from this repository.

WEIGHTS_NOTICE.md

@@ -0,0 +1,18 @@
+# Model weight licensing status
+
+**Project weight license: [Creative Commons Attribution-NonCommercial 4.0 International](licenses/CC-BY-NC-4.0.txt) (CC BY-NC 4.0).** This applies to all 29 project checkpoints identified in [manifest.json](manifest.json), only to rights the licensors are authorized to grant. Credit ACappellaSet and preserve upstream attribution; indicate modifications when sharing. It does not grant additional third-party rights or change upstream license terms. The full-remix and single-clone singer-augmentation checkpoints were trained with voice-converted singer augmentation generated using Moises; their project license does not grant additional rights controlled by service providers or other third parties.
+
+## Upstream notices
+
+Demucs source code uses MIT; preserve the [Meta notice](licenses/Demucs-MIT.txt).
+Pretrained initialization identities and tensor correspondence are recorded in the [initialization provenance](https://github.com/pting126/ACappellaSet/blob/main/third_party/initialization.json).
+
+Historical fixed and specialist training uses Music-Source-Separation-Training;
+preserve its [MIT notice](licenses/MSS-MIT.txt).
+Singer- and dual-conditioned models include fine-tuned WeSpeaker ReDimNet2
+B6-LM encoder weights; preserve the upstream [Apache license](licenses/WeSpeaker-Apache-2.0.txt)
+and [WeSpeaker attribution](https://huggingface.co/Wespeaker/wespeaker-voxceleb-redimnet2-B6-LM).
+Project fine-tuning changes pretrained tensors. All released checkpoints are inference-only exports, retaining every separator and matching encoder tensor while omitting optimizer state, training arguments, and internal filesystem paths.
+
+Original documentation and metadata retain their [MIT license](LICENSE).
+Dataset audio, corpus terms, and third-party rights remain separate.

conditioned/README.md

@@ -0,0 +1,9 @@
+# Shared query-conditioned models
+
+The release contains **14 selected checkpoints from 9 training runs**, including SK full-remix singer augmentation (131k) and SK single-clone singer augmentation (48k). Both augmentation models use pooled-inclusive validation selection. Single-clone refers to the augmentation dataset, not disabled waveform augmentation.
+
+See the [model card](../README.md#shared-query-conditioned) and [manifest](../manifest.json) for conditioning modes, training data, selected steps, validation corpora/scopes, and checkpoint hashes.
+
+[Public inference adapters](https://github.com/pting126/ACappellaSet/blob/main/docs/conditioned-inference.md) support all 14 selections and nine model recipes. Checkpoints include the matching fine-tuned encoder parameters for singer and dual models.
+
+Released files contain only the selected step and original model/encoder tensors; they load with `weights_only=True`. Preserve the registered enrollment-audio preprocessing. All 14 conditioned weights use CC BY-NC 4.0; see [weight notices](../WEIGHTS_NOTICE.md).

conditioned/dual/train-jc-natural/sisdr-v1/redimnet2-finetuned/step-0188000.json

@@ -0,0 +1,38 @@
+{
+  "id": "l4-step0188000",
+  "family": "conditioned",
+  "research_id": "L4",
+  "description": "HTDemucs | vocal-part + singer-enrollment-conditioned | jc-natural | redimnet2-finetuned | step 188000 (L4)",
+  "conditioning": "dual",
+  "training_data": "jc-natural",
+  "encoder": "redimnet2-finetuned",
+  "step": 188000,
+  "selection_horizon_updates": 200000,
+  "selection_labels": [
+    "sk_inclusive"
+  ],
+  "paper_tables": [
+    "target_conditioning"
+  ],
+  "release_inclusion": "confirmed",
+  "artifact_status": "published",
+  "public_inference_status": "published_export_strict_loading_verified",
+  "planned_checkpoint_path": "conditioned/dual/train-jc-natural/sisdr-v1/redimnet2-finetuned/step-0188000.pt",
+  "original_checkpoint_sha256": "724af8d2906fc08de54b7484df69143a3a7b2c62c412e6e473b979bc34cd74d0",
+  "original_checkpoint_filename": "best_step0188000.pt",
+  "released_checkpoint_sha256": "511dc6729da436c5fe72960e9a1de4968578fbb985fd00db5eb653e2d605dfb9",
+  "download_url": "https://huggingface.co/AcappellaSet/ACappellaSet-Models/resolve/main/conditioned/dual/train-jc-natural/sisdr-v1/redimnet2-finetuned/step-0188000.pt",
+  "weight_license": "cc-by-nc-4.0",
+  "original_hash_verification": "PASS, full-file SHA-256, 2026-10-14",
+  "intended_weight_license": "cc-by-nc-4.0",
+  "license_review_status": "assigned; applicable upstream notices preserved",
+  "notice_path": "WEIGHTS_NOTICE.md",
+  "config": "configs/conditioned/L4.json",
+  "config_sha256": "7c46503ca0ddc32623e184e801fe411184ff3a3d42dcddaf34106fb33d45f2cf",
+  "inference_model_id": "Dual_conditioned_htdemucs_JC",
+  "artifact_format": "acappellaset-conditioned-inference-v1",
+  "checkpoint_contents": "Inference-only checkpoint; all original model and matching encoder tensors preserved; no optimizer state, training arguments, or internal paths.",
+  "released_checkpoint_bytes": 161978465,
+  "weights_only": true,
+  "validation_report": "https://github.com/pting126/ACappellaSet/blob/main/docs/validation/inference-exports-2026-10-22.json"
+}

conditioned/dual/train-jc-natural/sisdr-v1/redimnet2-finetuned/step-0188000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:511dc6729da436c5fe72960e9a1de4968578fbb985fd00db5eb653e2d605dfb9
+size 161978465

conditioned/dual/train-sk-natural/sisdr-v1/redimnet2-finetuned/step-0007000.json

@@ -0,0 +1,38 @@
+{
+  "id": "k7-step0007000",
+  "family": "conditioned",
+  "research_id": "K7",
+  "description": "HTDemucs | vocal-part + singer-enrollment-conditioned | sk-natural | redimnet2-finetuned | step 7000 (K7)",
+  "conditioning": "dual",
+  "training_data": "sk-natural",
+  "encoder": "redimnet2-finetuned",
+  "step": 7000,
+  "selection_horizon_updates": 200000,
+  "selection_labels": [
+    "jc_inclusive"
+  ],
+  "paper_tables": [
+    "target_conditioning"
+  ],
+  "release_inclusion": "confirmed",
+  "artifact_status": "published",
+  "public_inference_status": "published_export_strict_loading_verified",
+  "planned_checkpoint_path": "conditioned/dual/train-sk-natural/sisdr-v1/redimnet2-finetuned/step-0007000.pt",
+  "original_checkpoint_sha256": "0646527189b1221552c01c918cff01173e98077dc515a6fe9f795ffab00c24ee",
+  "original_checkpoint_filename": "best_step0007000.pt",
+  "released_checkpoint_sha256": "528d75380b58ef43e266a0c1f6f3cd86883710d2a95ff9c492724790193fa6ab",
+  "download_url": "https://huggingface.co/AcappellaSet/ACappellaSet-Models/resolve/main/conditioned/dual/train-sk-natural/sisdr-v1/redimnet2-finetuned/step-0007000.pt",
+  "weight_license": "cc-by-nc-4.0",
+  "original_hash_verification": "PASS, full-file SHA-256, 2026-10-14",
+  "intended_weight_license": "cc-by-nc-4.0",
+  "notice_path": "WEIGHTS_NOTICE.md",
+  "license_review_status": "assigned; applicable upstream notices preserved",
+  "config": "configs/conditioned/K7.json",
+  "config_sha256": "5cbc1736ca73762cd34603e55cc3850e2cbf2872388898f710a4ff3e66b117dc",
+  "inference_model_id": "Dual_conditioned_htdemucs_SK",
+  "artifact_format": "acappellaset-conditioned-inference-v1",
+  "checkpoint_contents": "Inference-only checkpoint; all original model and matching encoder tensors preserved; no optimizer state, training arguments, or internal paths.",
+  "released_checkpoint_bytes": 161978465,
+  "weights_only": true,
+  "validation_report": "https://github.com/pting126/ACappellaSet/blob/main/docs/validation/inference-exports-2026-10-22.json"
+}

conditioned/dual/train-sk-natural/sisdr-v1/redimnet2-finetuned/step-0007000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:528d75380b58ef43e266a0c1f6f3cd86883710d2a95ff9c492724790193fa6ab
+size 161978465

conditioned/enrollment/train-jc-natural/sisdr-v1/redimnet2-finetuned/step-0184000.json

@@ -0,0 +1,38 @@
+{
+  "id": "l3-step0184000",
+  "family": "conditioned",
+  "research_id": "L3",
+  "description": "HTDemucs | singer-enrollment-conditioned | jc-natural | redimnet2-finetuned | step 184000 (L3)",
+  "conditioning": "singer",
+  "training_data": "jc-natural",
+  "encoder": "redimnet2-finetuned",
+  "step": 184000,
+  "selection_horizon_updates": 200000,
+  "selection_labels": [
+    "sk_inclusive"
+  ],
+  "paper_tables": [
+    "target_conditioning"
+  ],
+  "release_inclusion": "confirmed",
+  "artifact_status": "published",
+  "public_inference_status": "published_export_strict_loading_verified",
+  "planned_checkpoint_path": "conditioned/enrollment/train-jc-natural/sisdr-v1/redimnet2-finetuned/step-0184000.pt",
+  "original_checkpoint_sha256": "9f49b7d59575658dcfb9667395d998e8fc34a935b3995b099593e99ee1de5226",
+  "original_checkpoint_filename": "best_step0184000.pt",
+  "released_checkpoint_sha256": "7021317a39953018ea31d9f07578bbe4da9e3646ca62e1e59891abba57edd94d",
+  "download_url": "https://huggingface.co/AcappellaSet/ACappellaSet-Models/resolve/main/conditioned/enrollment/train-jc-natural/sisdr-v1/redimnet2-finetuned/step-0184000.pt",
+  "weight_license": "cc-by-nc-4.0",
+  "original_hash_verification": "PASS, full-file SHA-256, 2026-10-14",
+  "intended_weight_license": "cc-by-nc-4.0",
+  "license_review_status": "assigned; applicable upstream notices preserved",
+  "notice_path": "WEIGHTS_NOTICE.md",
+  "config": "configs/conditioned/L3.json",
+  "config_sha256": "a0bab6c22d17b9b23eca07c5cd3f2f41ffcb01d807ea78a19f3c131eccd91438",
+  "inference_model_id": "Singer_conditioned_htdemucs_JC",
+  "artifact_format": "acappellaset-conditioned-inference-v1",
+  "checkpoint_contents": "Inference-only checkpoint; all original model and matching encoder tensors preserved; no optimizer state, training arguments, or internal paths.",
+  "released_checkpoint_bytes": 161878741,
+  "weights_only": true,
+  "validation_report": "https://github.com/pting126/ACappellaSet/blob/main/docs/validation/inference-exports-2026-10-22.json"
+}

conditioned/enrollment/train-jc-natural/sisdr-v1/redimnet2-finetuned/step-0184000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7021317a39953018ea31d9f07578bbe4da9e3646ca62e1e59891abba57edd94d
+size 161878741

conditioned/enrollment/train-jc-sk-natural/sisdr-v1/redimnet2-finetuned/step-0195000.json

@@ -0,0 +1,44 @@
+{
+  "id": "m3-step0195000",
+  "family": "conditioned",
+  "research_id": "M3",
+  "description": "HTDemucs | singer-enrollment-conditioned | jc-sk-natural | redimnet2-finetuned | step 195000 (M3)",
+  "conditioning": "singer",
+  "training_data": "jc-sk-natural",
+  "encoder": "redimnet2-finetuned",
+  "step": 195000,
+  "selection_horizon_updates": 200000,
+  "selection_labels": [
+    "jc_shared",
+    "combined_shared"
+  ],
+  "paper_tables": [
+    "development_a"
+  ],
+  "release_inclusion": "confirmed",
+  "artifact_status": "published",
+  "public_inference_status": "published_export_strict_loading_verified",
+  "planned_checkpoint_path": "conditioned/enrollment/train-jc-sk-natural/sisdr-v1/redimnet2-finetuned/step-0195000.pt",
+  "original_checkpoint_sha256": "badc60d5bb2f2cf0cca92b11dc98c5a4de51fcd41312fff43607f7664ba5f9df",
+  "original_checkpoint_filename": "best_step0195000.pt",
+  "released_checkpoint_sha256": "12a92b20a7b1fed9614c48b925eaeb2f563428201d5c270c11e3f7c3df7359e7",
+  "download_url": "https://huggingface.co/AcappellaSet/ACappellaSet-Models/resolve/main/conditioned/enrollment/train-jc-sk-natural/sisdr-v1/redimnet2-finetuned/step-0195000.pt",
+  "weight_license": "cc-by-nc-4.0",
+  "original_hash_verification": "PASS, full-file SHA-256, 2026-10-14",
+  "supersedes": {
+    "step": 173000,
+    "sha256": "f99b7ee04a026c2c628bcfaac3ddbe7a5ffabc95bb96e9d0889e3079e8db1b06",
+    "preserve_historical": true
+  },
+  "intended_weight_license": "cc-by-nc-4.0",
+  "license_review_status": "assigned; applicable upstream notices preserved",
+  "notice_path": "WEIGHTS_NOTICE.md",
+  "config": "configs/conditioned/M3.json",
+  "config_sha256": "fee6def3080111793b362b2b9fda724720ddd13f28b2596f2a3eeb95fb8168d0",
+  "inference_model_id": "Singer_conditioned_htdemucs_SKnJC",
+  "artifact_format": "acappellaset-conditioned-inference-v1",
+  "checkpoint_contents": "Inference-only checkpoint; all original model and matching encoder tensors preserved; no optimizer state, training arguments, or internal paths.",
+  "released_checkpoint_bytes": 161878741,
+  "weights_only": true,
+  "validation_report": "https://github.com/pting126/ACappellaSet/blob/main/docs/validation/inference-exports-2026-10-22.json"
+}

conditioned/enrollment/train-jc-sk-natural/sisdr-v1/redimnet2-finetuned/step-0195000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:12a92b20a7b1fed9614c48b925eaeb2f563428201d5c270c11e3f7c3df7359e7
+size 161878741

conditioned/enrollment/train-jc-sk-natural/sisdr-v1/redimnet2-finetuned/step-0198000.json

@@ -0,0 +1,43 @@
+{
+  "id": "m3-step0198000",
+  "family": "conditioned",
+  "research_id": "M3",
+  "description": "HTDemucs | singer-enrollment-conditioned | jc-sk-natural | redimnet2-finetuned | step 198000 (M3)",
+  "conditioning": "singer",
+  "training_data": "jc-sk-natural",
+  "encoder": "redimnet2-finetuned",
+  "step": 198000,
+  "selection_horizon_updates": 200000,
+  "selection_labels": [
+    "sk_shared"
+  ],
+  "paper_tables": [
+    "development_a"
+  ],
+  "release_inclusion": "confirmed",
+  "artifact_status": "published",
+  "public_inference_status": "published_export_strict_loading_verified",
+  "planned_checkpoint_path": "conditioned/enrollment/train-jc-sk-natural/sisdr-v1/redimnet2-finetuned/step-0198000.pt",
+  "original_checkpoint_sha256": "6997e980ed2dded55ba1db2d5e88b71ad7024782125b2bd8dcf41df71189d6b4",
+  "original_checkpoint_filename": "best_step0198000.pt",
+  "released_checkpoint_sha256": "67c68a7f12076db456166edab0120f94947806689eca50da546f7adac260d77b",
+  "download_url": "https://huggingface.co/AcappellaSet/ACappellaSet-Models/resolve/main/conditioned/enrollment/train-jc-sk-natural/sisdr-v1/redimnet2-finetuned/step-0198000.pt",
+  "weight_license": "cc-by-nc-4.0",
+  "original_hash_verification": "PASS, full-file SHA-256, 2026-10-14",
+  "supersedes": {
+    "step": 159000,
+    "sha256": "03e74a6ae7f56d647c66e582f368f714b5504f8c657e78fb7f6c2142d0ae8ab1",
+    "preserve_historical": true
+  },
+  "intended_weight_license": "cc-by-nc-4.0",
+  "license_review_status": "assigned; applicable upstream notices preserved",
+  "notice_path": "WEIGHTS_NOTICE.md",
+  "config": "configs/conditioned/M3.json",
+  "config_sha256": "fee6def3080111793b362b2b9fda724720ddd13f28b2596f2a3eeb95fb8168d0",
+  "inference_model_id": "Singer_conditioned_htdemucs_SKnJC",
+  "artifact_format": "acappellaset-conditioned-inference-v1",
+  "checkpoint_contents": "Inference-only checkpoint; all original model and matching encoder tensors preserved; no optimizer state, training arguments, or internal paths.",
+  "released_checkpoint_bytes": 161878741,
+  "weights_only": true,
+  "validation_report": "https://github.com/pting126/ACappellaSet/blob/main/docs/validation/inference-exports-2026-10-22.json"
+}

conditioned/enrollment/train-jc-sk-natural/sisdr-v1/redimnet2-finetuned/step-0198000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:67c68a7f12076db456166edab0120f94947806689eca50da546f7adac260d77b
+size 161878741

conditioned/enrollment/train-sk-full-remix/sisdr-v1/redimnet2-finetuned/step-0131000.json

@@ -0,0 +1,38 @@
+{
+  "id": "k6-step0131000",
+  "family": "conditioned",
+  "research_id": "K6",
+  "description": "HTDemucs | singer-enrollment-conditioned | sk-full-remix | redimnet2-finetuned | step 131000 (K6)",
+  "conditioning": "singer",
+  "training_data": "sk-full-remix",
+  "encoder": "redimnet2-finetuned",
+  "step": 131000,
+  "selection_horizon_updates": 200000,
+  "selection_labels": [
+    "combined_inclusive"
+  ],
+  "paper_tables": [
+    "development_b"
+  ],
+  "release_inclusion": "confirmed",
+  "artifact_status": "published",
+  "public_inference_status": "published_export_strict_loading_verified",
+  "planned_checkpoint_path": "conditioned/enrollment/train-sk-full-remix/sisdr-v1/redimnet2-finetuned/step-0131000.pt",
+  "original_checkpoint_sha256": "ff74b497341edcfe6a6357749ed3e6bff795f1740ef022f4003f1241f198e517",
+  "original_checkpoint_filename": "best_step0131000.pt",
+  "released_checkpoint_sha256": "893326c96ac17b9cfed1cd5126176763f6a0a525115e3322c9a701ef232a9e69",
+  "download_url": "https://huggingface.co/AcappellaSet/ACappellaSet-Models/resolve/main/conditioned/enrollment/train-sk-full-remix/sisdr-v1/redimnet2-finetuned/step-0131000.pt",
+  "weight_license": "cc-by-nc-4.0",
+  "original_hash_verification": "PASS, full-file SHA-256, 2026-10-14",
+  "intended_weight_license": "cc-by-nc-4.0",
+  "license_review_status": "assigned; applicable upstream notices preserved",
+  "artifact_format": "acappellaset-conditioned-inference-v1",
+  "notice_path": "WEIGHTS_NOTICE.md",
+  "checkpoint_contents": "Inference-only checkpoint; all original model and matching encoder tensors preserved; no optimizer state, training arguments, or internal paths.",
+  "config": "configs/conditioned/Singer_conditioned_htdemucs_SK_full_remix.json",
+  "config_sha256": "fb94dcf47cd96b4ee07b0975996a300d9c0938e3a24b855cfbc273fe60f58c8a",
+  "released_checkpoint_bytes": 161878741,
+  "inference_model_id": "Singer_conditioned_htdemucs_SK_full_remix",
+  "weights_only": true,
+  "validation_report": "https://github.com/pting126/ACappellaSet/blob/main/docs/validation/inference-exports-2026-10-22.json"
+}

conditioned/enrollment/train-sk-full-remix/sisdr-v1/redimnet2-finetuned/step-0131000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:893326c96ac17b9cfed1cd5126176763f6a0a525115e3322c9a701ef232a9e69
+size 161878741

conditioned/enrollment/train-sk-natural/sisdr-v1/redimnet2-finetuned/step-0012000.json

@@ -0,0 +1,40 @@
+{
+  "id": "k5-step0012000",
+  "family": "conditioned",
+  "research_id": "K5",
+  "description": "HTDemucs | singer-enrollment-conditioned | sk-natural | redimnet2-finetuned | step 12000 (K5)",
+  "conditioning": "singer",
+  "training_data": "sk-natural",
+  "encoder": "redimnet2-finetuned",
+  "step": 12000,
+  "selection_horizon_updates": 200000,
+  "selection_labels": [
+    "jc_shared",
+    "jc_inclusive"
+  ],
+  "paper_tables": [
+    "development_a",
+    "target_conditioning"
+  ],
+  "release_inclusion": "confirmed",
+  "artifact_status": "published",
+  "public_inference_status": "published_export_strict_loading_verified",
+  "planned_checkpoint_path": "conditioned/enrollment/train-sk-natural/sisdr-v1/redimnet2-finetuned/step-0012000.pt",
+  "original_checkpoint_sha256": "6f8e199008a542904330495dab28d7f56d8c380a6959ea4e460b22450c56ca43",
+  "original_checkpoint_filename": "best_step0012000.pt",
+  "released_checkpoint_sha256": "89c5a8829aca66cdbb0033338dc83d769e06f98300f68464a666b87f73e80abf",
+  "download_url": "https://huggingface.co/AcappellaSet/ACappellaSet-Models/resolve/main/conditioned/enrollment/train-sk-natural/sisdr-v1/redimnet2-finetuned/step-0012000.pt",
+  "weight_license": "cc-by-nc-4.0",
+  "original_hash_verification": "PASS, full-file SHA-256, 2026-10-14",
+  "intended_weight_license": "cc-by-nc-4.0",
+  "notice_path": "WEIGHTS_NOTICE.md",
+  "license_review_status": "assigned; applicable upstream notices preserved",
+  "config": "configs/conditioned/K5.json",
+  "config_sha256": "fb94dcf47cd96b4ee07b0975996a300d9c0938e3a24b855cfbc273fe60f58c8a",
+  "inference_model_id": "Singer_conditioned_htdemucs_SK",
+  "artifact_format": "acappellaset-conditioned-inference-v1",
+  "checkpoint_contents": "Inference-only checkpoint; all original model and matching encoder tensors preserved; no optimizer state, training arguments, or internal paths.",
+  "released_checkpoint_bytes": 161878741,
+  "weights_only": true,
+  "validation_report": "https://github.com/pting126/ACappellaSet/blob/main/docs/validation/inference-exports-2026-10-22.json"
+}

conditioned/enrollment/train-sk-natural/sisdr-v1/redimnet2-finetuned/step-0012000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:89c5a8829aca66cdbb0033338dc83d769e06f98300f68464a666b87f73e80abf
+size 161878741

conditioned/enrollment/train-sk-natural/sisdr-v1/redimnet2-finetuned/step-0106000.json

@@ -0,0 +1,40 @@
+{
+  "id": "k5-step0106000",
+  "family": "conditioned",
+  "research_id": "K5",
+  "description": "HTDemucs | singer-enrollment-conditioned | sk-natural | redimnet2-finetuned | step 106000 (K5)",
+  "conditioning": "singer",
+  "training_data": "sk-natural",
+  "encoder": "redimnet2-finetuned",
+  "step": 106000,
+  "selection_horizon_updates": 200000,
+  "selection_labels": [
+    "combined_shared",
+    "combined_inclusive"
+  ],
+  "paper_tables": [
+    "development_a",
+    "development_b"
+  ],
+  "release_inclusion": "confirmed",
+  "artifact_status": "published",
+  "public_inference_status": "published_export_strict_loading_verified",
+  "planned_checkpoint_path": "conditioned/enrollment/train-sk-natural/sisdr-v1/redimnet2-finetuned/step-0106000.pt",
+  "original_checkpoint_sha256": "df846b039ddd5e032f221adcf3de33a9a3f546f21b47f0bc3d30df891ab4ca93",
+  "original_checkpoint_filename": "best_step0106000.pt",
+  "released_checkpoint_sha256": "b28626e30a12ba43337b3cec3da2558685ddcc649fe34dd78e7e71590e922b29",
+  "download_url": "https://huggingface.co/AcappellaSet/ACappellaSet-Models/resolve/main/conditioned/enrollment/train-sk-natural/sisdr-v1/redimnet2-finetuned/step-0106000.pt",
+  "weight_license": "cc-by-nc-4.0",
+  "original_hash_verification": "PASS, full-file SHA-256, 2026-10-14",
+  "intended_weight_license": "cc-by-nc-4.0",
+  "notice_path": "WEIGHTS_NOTICE.md",
+  "license_review_status": "assigned; applicable upstream notices preserved",
+  "config": "configs/conditioned/K5.json",
+  "config_sha256": "fb94dcf47cd96b4ee07b0975996a300d9c0938e3a24b855cfbc273fe60f58c8a",
+  "inference_model_id": "Singer_conditioned_htdemucs_SK",
+  "artifact_format": "acappellaset-conditioned-inference-v1",
+  "checkpoint_contents": "Inference-only checkpoint; all original model and matching encoder tensors preserved; no optimizer state, training arguments, or internal paths.",
+  "released_checkpoint_bytes": 161878741,
+  "weights_only": true,
+  "validation_report": "https://github.com/pting126/ACappellaSet/blob/main/docs/validation/inference-exports-2026-10-22.json"
+}

conditioned/enrollment/train-sk-natural/sisdr-v1/redimnet2-finetuned/step-0106000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b28626e30a12ba43337b3cec3da2558685ddcc649fe34dd78e7e71590e922b29
+size 161878741

conditioned/enrollment/train-sk-natural/sisdr-v1/redimnet2-finetuned/step-0179000.json

@@ -0,0 +1,38 @@
+{
+  "id": "k5-step0179000",
+  "family": "conditioned",
+  "research_id": "K5",
+  "description": "HTDemucs | singer-enrollment-conditioned | sk-natural | redimnet2-finetuned | step 179000 (K5)",
+  "conditioning": "singer",
+  "training_data": "sk-natural",
+  "encoder": "redimnet2-finetuned",
+  "step": 179000,
+  "selection_horizon_updates": 200000,
+  "selection_labels": [
+    "sk_shared"
+  ],
+  "paper_tables": [
+    "development_a"
+  ],
+  "release_inclusion": "confirmed",
+  "artifact_status": "published",
+  "public_inference_status": "published_export_strict_loading_verified",
+  "planned_checkpoint_path": "conditioned/enrollment/train-sk-natural/sisdr-v1/redimnet2-finetuned/step-0179000.pt",
+  "original_checkpoint_sha256": "eeeac981bacdd063207a0258fb60625edd2b60e94bc544852a1c2459223a04d3",
+  "original_checkpoint_filename": "best_step0179000.pt",
+  "released_checkpoint_sha256": "537531081da4c7c1a7d5568f8847bfaab1df3d54658486381bdd7f3d2f7f8d7d",
+  "download_url": "https://huggingface.co/AcappellaSet/ACappellaSet-Models/resolve/main/conditioned/enrollment/train-sk-natural/sisdr-v1/redimnet2-finetuned/step-0179000.pt",
+  "weight_license": "cc-by-nc-4.0",
+  "original_hash_verification": "PASS, full-file SHA-256, 2026-10-14",
+  "intended_weight_license": "cc-by-nc-4.0",
+  "notice_path": "WEIGHTS_NOTICE.md",
+  "license_review_status": "assigned; applicable upstream notices preserved",
+  "config": "configs/conditioned/K5.json",
+  "config_sha256": "fb94dcf47cd96b4ee07b0975996a300d9c0938e3a24b855cfbc273fe60f58c8a",
+  "inference_model_id": "Singer_conditioned_htdemucs_SK",
+  "artifact_format": "acappellaset-conditioned-inference-v1",
+  "checkpoint_contents": "Inference-only checkpoint; all original model and matching encoder tensors preserved; no optimizer state, training arguments, or internal paths.",
+  "released_checkpoint_bytes": 161878741,
+  "weights_only": true,
+  "validation_report": "https://github.com/pting126/ACappellaSet/blob/main/docs/validation/inference-exports-2026-10-22.json"
+}

conditioned/enrollment/train-sk-natural/sisdr-v1/redimnet2-finetuned/step-0179000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:537531081da4c7c1a7d5568f8847bfaab1df3d54658486381bdd7f3d2f7f8d7d
+size 161878741

conditioned/enrollment/train-sk-single-clone/sisdr-v1/redimnet2-finetuned/step-0048000.json

@@ -0,0 +1,38 @@
+{
+  "id": "k6-no-aug-step0048000",
+  "family": "conditioned",
+  "research_id": "K6_no_aug",
+  "description": "HTDemucs | singer-enrollment-conditioned | sk-single-clone | redimnet2-finetuned | step 48000 (K6_no_aug)",
+  "conditioning": "singer",
+  "training_data": "sk-single-clone",
+  "encoder": "redimnet2-finetuned",
+  "step": 48000,
+  "selection_horizon_updates": 200000,
+  "selection_labels": [
+    "combined_inclusive"
+  ],
+  "paper_tables": [
+    "development_b"
+  ],
+  "release_inclusion": "confirmed",
+  "artifact_status": "published",
+  "public_inference_status": "published_export_strict_loading_verified",
+  "planned_checkpoint_path": "conditioned/enrollment/train-sk-single-clone/sisdr-v1/redimnet2-finetuned/step-0048000.pt",
+  "original_checkpoint_sha256": "79ee3dd239925ea2c075641eadc0e13c001a81f9f367f8ee6881c5e8a3881901",
+  "original_checkpoint_filename": "best_step0048000.pt",
+  "released_checkpoint_sha256": "00fc378e77c7148f75e853ba8e84f246aba1557c30ec3e5f205e4227099c5c65",
+  "download_url": "https://huggingface.co/AcappellaSet/ACappellaSet-Models/resolve/main/conditioned/enrollment/train-sk-single-clone/sisdr-v1/redimnet2-finetuned/step-0048000.pt",
+  "weight_license": "cc-by-nc-4.0",
+  "original_hash_verification": "PASS, full-file SHA-256, 2026-10-14",
+  "intended_weight_license": "cc-by-nc-4.0",
+  "license_review_status": "assigned; applicable upstream notices preserved",
+  "artifact_format": "acappellaset-conditioned-inference-v1",
+  "notice_path": "WEIGHTS_NOTICE.md",
+  "checkpoint_contents": "Inference-only checkpoint; all original model and matching encoder tensors preserved; no optimizer state, training arguments, or internal paths.",
+  "config": "configs/conditioned/Singer_conditioned_htdemucs_SK_single_clone.json",
+  "config_sha256": "fb94dcf47cd96b4ee07b0975996a300d9c0938e3a24b855cfbc273fe60f58c8a",
+  "released_checkpoint_bytes": 161878741,
+  "inference_model_id": "Singer_conditioned_htdemucs_SK_single_clone",
+  "weights_only": true,
+  "validation_report": "https://github.com/pting126/ACappellaSet/blob/main/docs/validation/inference-exports-2026-10-22.json"
+}

conditioned/enrollment/train-sk-single-clone/sisdr-v1/redimnet2-finetuned/step-0048000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:00fc378e77c7148f75e853ba8e84f246aba1557c30ec3e5f205e4227099c5c65
+size 161878741

conditioned/role/train-jc-natural/sisdr-v1/no-encoder/step-0119000.json

@@ -0,0 +1,38 @@
+{
+  "id": "l1-step0119000",
+  "family": "conditioned",
+  "research_id": "L1",
+  "description": "HTDemucs | vocal-part-conditioned | jc-natural | no-encoder | step 119000 (L1)",
+  "conditioning": "part",
+  "training_data": "jc-natural",
+  "encoder": "no-encoder",
+  "step": 119000,
+  "selection_horizon_updates": 200000,
+  "selection_labels": [
+    "sk_inclusive"
+  ],
+  "paper_tables": [
+    "target_conditioning"
+  ],
+  "release_inclusion": "confirmed",
+  "artifact_status": "published",
+  "public_inference_status": "published_export_strict_loading_verified",
+  "planned_checkpoint_path": "conditioned/role/train-jc-natural/sisdr-v1/no-encoder/step-0119000.pt",
+  "original_checkpoint_sha256": "5de579abaf11546eab1fcfa29566cf4934ece6c5fe7b23df56b6af27c3be3723",
+  "original_checkpoint_filename": "best_step0119000.pt",
+  "released_checkpoint_sha256": "6af03b62d8e5f419f72787b7909c3e82ad7d3194fa43c920a47a064374f63e8b",
+  "download_url": "https://huggingface.co/AcappellaSet/ACappellaSet-Models/resolve/main/conditioned/role/train-jc-natural/sisdr-v1/no-encoder/step-0119000.pt",
+  "weight_license": "cc-by-nc-4.0",
+  "original_hash_verification": "PASS, full-file SHA-256, 2026-10-14",
+  "intended_weight_license": "cc-by-nc-4.0",
+  "license_review_status": "assigned; applicable upstream notices preserved",
+  "notice_path": "WEIGHTS_NOTICE.md",
+  "config": "configs/conditioned/L1.json",
+  "config_sha256": "45fca96de4f1849d076368538dd789f97e5e0a4ab4443c6c0922e4505c2ac08a",
+  "inference_model_id": "Part_conditioned_htdemucs_JC",
+  "artifact_format": "acappellaset-conditioned-inference-v1",
+  "checkpoint_contents": "Inference-only checkpoint; all original model and matching encoder tensors preserved; no optimizer state, training arguments, or internal paths.",
+  "released_checkpoint_bytes": 109889741,
+  "weights_only": true,
+  "validation_report": "https://github.com/pting126/ACappellaSet/blob/main/docs/validation/inference-exports-2026-10-22.json"
+}

conditioned/role/train-jc-natural/sisdr-v1/no-encoder/step-0119000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6af03b62d8e5f419f72787b7909c3e82ad7d3194fa43c920a47a064374f63e8b
+size 109889741

conditioned/role/train-jc-natural/sisdr-v1/no-encoder/step-0141000.json

@@ -0,0 +1,38 @@
+{
+  "id": "l1-step0141000",
+  "family": "conditioned",
+  "research_id": "L1",
+  "description": "HTDemucs | vocal-part-conditioned | jc-natural | no-encoder | step 141000 (L1)",
+  "conditioning": "part",
+  "training_data": "jc-natural",
+  "encoder": "no-encoder",
+  "step": 141000,
+  "selection_horizon_updates": 200000,
+  "selection_labels": [
+    "jc_shared"
+  ],
+  "paper_tables": [
+    "cross_corpus"
+  ],
+  "release_inclusion": "confirmed",
+  "artifact_status": "published",
+  "public_inference_status": "published_export_strict_loading_verified",
+  "planned_checkpoint_path": "conditioned/role/train-jc-natural/sisdr-v1/no-encoder/step-0141000.pt",
+  "original_checkpoint_sha256": "92e19968bef2ce84d3800141afc50cc47141bd016adfcc07b70e705df7408004",
+  "original_checkpoint_filename": "best_step0141000.pt",
+  "released_checkpoint_sha256": "f9fdca2de1f89b650b603d386f39128b4b06b0ebd4a59469f57ce256e8c26065",
+  "download_url": "https://huggingface.co/AcappellaSet/ACappellaSet-Models/resolve/main/conditioned/role/train-jc-natural/sisdr-v1/no-encoder/step-0141000.pt",
+  "weight_license": "cc-by-nc-4.0",
+  "original_hash_verification": "PASS, full-file SHA-256, 2026-10-14",
+  "intended_weight_license": "cc-by-nc-4.0",
+  "license_review_status": "assigned; applicable upstream notices preserved",
+  "notice_path": "WEIGHTS_NOTICE.md",
+  "config": "configs/conditioned/L1.json",
+  "config_sha256": "45fca96de4f1849d076368538dd789f97e5e0a4ab4443c6c0922e4505c2ac08a",
+  "inference_model_id": "Part_conditioned_htdemucs_JC",
+  "artifact_format": "acappellaset-conditioned-inference-v1",
+  "checkpoint_contents": "Inference-only checkpoint; all original model and matching encoder tensors preserved; no optimizer state, training arguments, or internal paths.",
+  "released_checkpoint_bytes": 109889741,
+  "weights_only": true,
+  "validation_report": "https://github.com/pting126/ACappellaSet/blob/main/docs/validation/inference-exports-2026-10-22.json"
+}

conditioned/role/train-jc-natural/sisdr-v1/no-encoder/step-0141000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f9fdca2de1f89b650b603d386f39128b4b06b0ebd4a59469f57ce256e8c26065
+size 109889741

conditioned/role/train-sk-natural/sisdr-v1/no-encoder/step-0004000.json

@@ -0,0 +1,38 @@
+{
+  "id": "k1-step0004000",
+  "family": "conditioned",
+  "research_id": "K1",
+  "description": "HTDemucs | vocal-part-conditioned | sk-natural | no-encoder | step 4000 (K1)",
+  "conditioning": "part",
+  "training_data": "sk-natural",
+  "encoder": "no-encoder",
+  "step": 4000,
+  "selection_horizon_updates": 200000,
+  "selection_labels": [
+    "jc_inclusive"
+  ],
+  "paper_tables": [
+    "target_conditioning"
+  ],
+  "release_inclusion": "confirmed",
+  "artifact_status": "published",
+  "public_inference_status": "published_export_strict_loading_verified",
+  "planned_checkpoint_path": "conditioned/role/train-sk-natural/sisdr-v1/no-encoder/step-0004000.pt",
+  "original_checkpoint_sha256": "979ba17521f398fb1b2a3e0ccad29615999b50e87d76347b2682212a61b49427",
+  "original_checkpoint_filename": "best_step0004000.pt",
+  "released_checkpoint_sha256": "31821cfe956fc6a7527e718d6702bcd71b84a8881484c1339753bca7207b316b",
+  "download_url": "https://huggingface.co/AcappellaSet/ACappellaSet-Models/resolve/main/conditioned/role/train-sk-natural/sisdr-v1/no-encoder/step-0004000.pt",
+  "weight_license": "cc-by-nc-4.0",
+  "original_hash_verification": "PASS, full-file SHA-256, 2026-10-14",
+  "intended_weight_license": "cc-by-nc-4.0",
+  "notice_path": "WEIGHTS_NOTICE.md",
+  "license_review_status": "assigned; applicable upstream notices preserved",
+  "config": "configs/conditioned/K1.json",
+  "config_sha256": "a2c358f669574dafa25403b5d6e4a0ee0fc3cf237e03399886761f95fd7c0fbd",
+  "inference_model_id": "Part_conditioned_htdemucs_SK",
+  "artifact_format": "acappellaset-conditioned-inference-v1",
+  "checkpoint_contents": "Inference-only checkpoint; all original model and matching encoder tensors preserved; no optimizer state, training arguments, or internal paths.",
+  "released_checkpoint_bytes": 109889741,
+  "weights_only": true,
+  "validation_report": "https://github.com/pting126/ACappellaSet/blob/main/docs/validation/inference-exports-2026-10-22.json"
+}

conditioned/role/train-sk-natural/sisdr-v1/no-encoder/step-0004000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:31821cfe956fc6a7527e718d6702bcd71b84a8881484c1339753bca7207b316b
+size 109889741

conditioned/role/train-sk-natural/sisdr-v1/no-encoder/step-0198000.json

@@ -0,0 +1,43 @@
+{
+  "id": "k1-step0198000",
+  "family": "conditioned",
+  "research_id": "K1",
+  "description": "HTDemucs | vocal-part-conditioned | sk-natural | no-encoder | step 198000 (K1)",
+  "conditioning": "part",
+  "training_data": "sk-natural",
+  "encoder": "no-encoder",
+  "step": 198000,
+  "selection_horizon_updates": 200000,
+  "selection_labels": [
+    "sk_shared"
+  ],
+  "paper_tables": [
+    "cross_corpus"
+  ],
+  "release_inclusion": "confirmed",
+  "artifact_status": "published",
+  "public_inference_status": "published_export_strict_loading_verified",
+  "planned_checkpoint_path": "conditioned/role/train-sk-natural/sisdr-v1/no-encoder/step-0198000.pt",
+  "original_checkpoint_sha256": "fbd19ff91075877839bb76b022d1b5ad018a980a26603ea9cb4cca7aac0feb6a",
+  "original_checkpoint_filename": "best_step0198000.pt",
+  "released_checkpoint_sha256": "9c59bfca391faf96e6d3f130d27dd75f4ed5f0eb1b950ed75361b343e926af14",
+  "download_url": "https://huggingface.co/AcappellaSet/ACappellaSet-Models/resolve/main/conditioned/role/train-sk-natural/sisdr-v1/no-encoder/step-0198000.pt",
+  "weight_license": "cc-by-nc-4.0",
+  "original_hash_verification": "PASS, full-file SHA-256, 2026-10-14",
+  "supersedes": {
+    "step": 184000,
+    "sha256": "261daa7899501229143136c46a501097caba9ac903650bc9c475c3ee049407b0",
+    "preserve_historical": true
+  },
+  "intended_weight_license": "cc-by-nc-4.0",
+  "notice_path": "WEIGHTS_NOTICE.md",
+  "license_review_status": "assigned; applicable upstream notices preserved",
+  "config": "configs/conditioned/K1.json",
+  "config_sha256": "a2c358f669574dafa25403b5d6e4a0ee0fc3cf237e03399886761f95fd7c0fbd",
+  "inference_model_id": "Part_conditioned_htdemucs_SK",
+  "artifact_format": "acappellaset-conditioned-inference-v1",
+  "checkpoint_contents": "Inference-only checkpoint; all original model and matching encoder tensors preserved; no optimizer state, training arguments, or internal paths.",
+  "released_checkpoint_bytes": 109889741,
+  "weights_only": true,
+  "validation_report": "https://github.com/pting126/ACappellaSet/blob/main/docs/validation/inference-exports-2026-10-22.json"
+}

conditioned/role/train-sk-natural/sisdr-v1/no-encoder/step-0198000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c59bfca391faf96e6d3f130d27dd75f4ed5f0eb1b950ed75361b343e926af14
+size 109889741

configs/conditioned/K1.json

@@ -0,0 +1,31 @@
+{
+  "conditioning": "role",
+  "source_mode": "target_residual",
+  "channels": "stereo",
+  "seconds": 10.0,
+  "enroll_seconds": 10.0,
+  "emb_dim": 192,
+  "role_embedding_dim": 16,
+  "embedding_backend": "wespeaker_redimnet2",
+  "freeze_external_embedding": true,
+  "ht_channels": 48,
+  "ht_depth": 4,
+  "ht_growth": 2,
+  "ht_bottom_channels": 0,
+  "ht_dconv_mode": 3,
+  "ht_t_layers": 5,
+  "ht_nfft": 4096,
+  "sample_rate": 44100,
+  "hop_seconds": 4.0,
+  "training_roles": [
+    "A",
+    "B",
+    "Bar",
+    "S",
+    "T",
+    "VP"
+  ],
+  "untrained_zero_role": "Lead",
+  "encoder_config": "configs/conditioned/redimnet2.yaml",
+  "encoder_config_sha256": "57a348c24d56f7bf26794f8bcca2765984726ab56363b2d3e5ef3a27b56524fc"
+}

configs/conditioned/K5.json

@@ -0,0 +1,31 @@
+{
+  "conditioning": "enrollment",
+  "source_mode": "target_residual",
+  "channels": "stereo",
+  "seconds": 10.0,
+  "enroll_seconds": 10.0,
+  "emb_dim": 192,
+  "role_embedding_dim": 16,
+  "embedding_backend": "wespeaker_redimnet2",
+  "freeze_external_embedding": false,
+  "ht_channels": 48,
+  "ht_depth": 4,
+  "ht_growth": 2,
+  "ht_bottom_channels": 0,
+  "ht_dconv_mode": 3,
+  "ht_t_layers": 5,
+  "ht_nfft": 4096,
+  "sample_rate": 44100,
+  "hop_seconds": 4.0,
+  "training_roles": [
+    "A",
+    "B",
+    "Bar",
+    "S",
+    "T",
+    "VP"
+  ],
+  "untrained_zero_role": "Lead",
+  "encoder_config": "configs/conditioned/redimnet2.yaml",
+  "encoder_config_sha256": "57a348c24d56f7bf26794f8bcca2765984726ab56363b2d3e5ef3a27b56524fc"
+}

configs/conditioned/K7.json

@@ -0,0 +1,31 @@
+{
+  "conditioning": "dual",
+  "source_mode": "target_residual",
+  "channels": "stereo",
+  "seconds": 10.0,
+  "enroll_seconds": 10.0,
+  "emb_dim": 192,
+  "role_embedding_dim": 16,
+  "embedding_backend": "wespeaker_redimnet2",
+  "freeze_external_embedding": false,
+  "ht_channels": 48,
+  "ht_depth": 4,
+  "ht_growth": 2,
+  "ht_bottom_channels": 0,
+  "ht_dconv_mode": 3,
+  "ht_t_layers": 5,
+  "ht_nfft": 4096,
+  "sample_rate": 44100,
+  "hop_seconds": 4.0,
+  "training_roles": [
+    "A",
+    "B",
+    "Bar",
+    "S",
+    "T",
+    "VP"
+  ],
+  "untrained_zero_role": "Lead",
+  "encoder_config": "configs/conditioned/redimnet2.yaml",
+  "encoder_config_sha256": "57a348c24d56f7bf26794f8bcca2765984726ab56363b2d3e5ef3a27b56524fc"
+}

configs/conditioned/L1.json

@@ -0,0 +1,31 @@
+{
+  "conditioning": "role",
+  "source_mode": "target_residual",
+  "channels": "stereo",
+  "seconds": 10.0,
+  "enroll_seconds": 10.0,
+  "emb_dim": 192,
+  "role_embedding_dim": 16,
+  "embedding_backend": "wespeaker_redimnet2",
+  "freeze_external_embedding": true,
+  "ht_channels": 48,
+  "ht_depth": 4,
+  "ht_growth": 2,
+  "ht_bottom_channels": 0,
+  "ht_dconv_mode": 3,
+  "ht_t_layers": 5,
+  "ht_nfft": 4096,
+  "sample_rate": 44100,
+  "hop_seconds": 4.0,
+  "training_roles": [
+    "A",
+    "B",
+    "Lead",
+    "S",
+    "T",
+    "VP"
+  ],
+  "untrained_zero_role": "Bar",
+  "encoder_config": "configs/conditioned/redimnet2.yaml",
+  "encoder_config_sha256": "57a348c24d56f7bf26794f8bcca2765984726ab56363b2d3e5ef3a27b56524fc"
+}

configs/conditioned/L3.json

@@ -0,0 +1,31 @@
+{
+  "conditioning": "enrollment",
+  "source_mode": "target_residual",
+  "channels": "stereo",
+  "seconds": 10.0,
+  "enroll_seconds": 10.0,
+  "emb_dim": 192,
+  "role_embedding_dim": 16,
+  "embedding_backend": "wespeaker_redimnet2",
+  "freeze_external_embedding": false,
+  "ht_channels": 48,
+  "ht_depth": 4,
+  "ht_growth": 2,
+  "ht_bottom_channels": 0,
+  "ht_dconv_mode": 3,
+  "ht_t_layers": 5,
+  "ht_nfft": 4096,
+  "sample_rate": 44100,
+  "hop_seconds": 4.0,
+  "training_roles": [
+    "A",
+    "B",
+    "Lead",
+    "S",
+    "T",
+    "VP"
+  ],
+  "untrained_zero_role": "Bar",
+  "encoder_config": "configs/conditioned/redimnet2.yaml",
+  "encoder_config_sha256": "57a348c24d56f7bf26794f8bcca2765984726ab56363b2d3e5ef3a27b56524fc"
+}

configs/conditioned/L4.json

@@ -0,0 +1,31 @@
+{
+  "conditioning": "dual",
+  "source_mode": "target_residual",
+  "channels": "stereo",
+  "seconds": 10.0,
+  "enroll_seconds": 10.0,
+  "emb_dim": 192,
+  "role_embedding_dim": 16,
+  "embedding_backend": "wespeaker_redimnet2",
+  "freeze_external_embedding": false,
+  "ht_channels": 48,
+  "ht_depth": 4,
+  "ht_growth": 2,
+  "ht_bottom_channels": 0,
+  "ht_dconv_mode": 3,
+  "ht_t_layers": 5,
+  "ht_nfft": 4096,
+  "sample_rate": 44100,
+  "hop_seconds": 4.0,
+  "training_roles": [
+    "A",
+    "B",
+    "Lead",
+    "S",
+    "T",
+    "VP"
+  ],
+  "untrained_zero_role": "Bar",
+  "encoder_config": "configs/conditioned/redimnet2.yaml",
+  "encoder_config_sha256": "57a348c24d56f7bf26794f8bcca2765984726ab56363b2d3e5ef3a27b56524fc"
+}

configs/conditioned/M3.json

@@ -0,0 +1,32 @@
+{
+  "conditioning": "enrollment",
+  "source_mode": "target_residual",
+  "channels": "stereo",
+  "seconds": 10.0,
+  "enroll_seconds": 10.0,
+  "emb_dim": 192,
+  "role_embedding_dim": 16,
+  "embedding_backend": "wespeaker_redimnet2",
+  "freeze_external_embedding": false,
+  "ht_channels": 48,
+  "ht_depth": 4,
+  "ht_growth": 2,
+  "ht_bottom_channels": 0,
+  "ht_dconv_mode": 3,
+  "ht_t_layers": 5,
+  "ht_nfft": 4096,
+  "sample_rate": 44100,
+  "hop_seconds": 4.0,
+  "training_roles": [
+    "A",
+    "B",
+    "Bar",
+    "Lead",
+    "S",
+    "T",
+    "VP"
+  ],
+  "untrained_zero_role": null,
+  "encoder_config": "configs/conditioned/redimnet2.yaml",
+  "encoder_config_sha256": "57a348c24d56f7bf26794f8bcca2765984726ab56363b2d3e5ef3a27b56524fc"
+}

configs/conditioned/Singer_conditioned_htdemucs_SK_full_remix.json

@@ -0,0 +1,31 @@
+{
+  "conditioning": "enrollment",
+  "source_mode": "target_residual",
+  "channels": "stereo",
+  "seconds": 10.0,
+  "enroll_seconds": 10.0,
+  "emb_dim": 192,
+  "role_embedding_dim": 16,
+  "embedding_backend": "wespeaker_redimnet2",
+  "freeze_external_embedding": false,
+  "ht_channels": 48,
+  "ht_depth": 4,
+  "ht_growth": 2,
+  "ht_bottom_channels": 0,
+  "ht_dconv_mode": 3,
+  "ht_t_layers": 5,
+  "ht_nfft": 4096,
+  "sample_rate": 44100,
+  "hop_seconds": 4.0,
+  "training_roles": [
+    "A",
+    "B",
+    "Bar",
+    "S",
+    "T",
+    "VP"
+  ],
+  "untrained_zero_role": "Lead",
+  "encoder_config": "configs/conditioned/redimnet2.yaml",
+  "encoder_config_sha256": "57a348c24d56f7bf26794f8bcca2765984726ab56363b2d3e5ef3a27b56524fc"
+}

configs/conditioned/Singer_conditioned_htdemucs_SK_single_clone.json

@@ -0,0 +1,31 @@
+{
+  "conditioning": "enrollment",
+  "source_mode": "target_residual",
+  "channels": "stereo",
+  "seconds": 10.0,
+  "enroll_seconds": 10.0,
+  "emb_dim": 192,
+  "role_embedding_dim": 16,
+  "embedding_backend": "wespeaker_redimnet2",
+  "freeze_external_embedding": false,
+  "ht_channels": 48,
+  "ht_depth": 4,
+  "ht_growth": 2,
+  "ht_bottom_channels": 0,
+  "ht_dconv_mode": 3,
+  "ht_t_layers": 5,
+  "ht_nfft": 4096,
+  "sample_rate": 44100,
+  "hop_seconds": 4.0,
+  "training_roles": [
+    "A",
+    "B",
+    "Bar",
+    "S",
+    "T",
+    "VP"
+  ],
+  "untrained_zero_role": "Lead",
+  "encoder_config": "configs/conditioned/redimnet2.yaml",
+  "encoder_config_sha256": "57a348c24d56f7bf26794f8bcca2765984726ab56363b2d3e5ef3a27b56524fc"
+}

configs/conditioned/redimnet2.yaml

@@ -0,0 +1,147 @@
+checkpoint: exp/redimnet2/expB6_v2/-LM/models/model_0.pt
+data_type: shard
+dataloader_args:
+  batch_size: 16
+  drop_last: true
+  num_workers: 4
+  pin_memory: false
+  prefetch_factor: 4
+dataset_args:
+  aug_prob: 0.0
+  cmvn: false
+  filter: false
+  frontend: tfmel
+  num_frms: 600
+  resample_rate: 16000
+  sample_num_per_epoch: 0
+  shuffle: true
+  shuffle_args:
+    shuffle_size: 2500
+  spec_aug: false
+  speed_perturb: false
+  tfmel_args:
+    do_preemph: true
+    do_spec_aug: false
+    eps: 1.0e-08
+    f_max: 7600
+    f_min: 20
+    hop_length: 160
+    n_fft: 512
+    n_mels: 72
+    norm_signal: true
+    win_length: 400
+enable_amp: true
+exp_dir: exp/redimnet2/expB6_v2/-LM
+gpus:
+- 2
+- 3
+grad_accum_steps: 8
+log_batch_interval: 100
+loss: CrossEntropyLoss
+loss_args: {}
+margin_scheduler: MarginScheduler
+margin_update:
+  epoch_iter: 34125
+  final_margin: 0.3
+  fix_start_epoch: 0
+  increase_start_epoch: 0
+  increase_type: exp
+  initial_margin: 0.3
+  update_margin: false
+model: ReDimNet2Wrap
+model_args:
+  C: 64
+  F: 72
+  block_1d_type: conv+att
+  block_2d_type: basic_resnet
+  causal: none
+  compress_tconvs: true
+  emb_bn: false
+  embed_dim: 192
+  feat_dim: 72
+  fm_weigthing_type: NC
+  global_context_att: true
+  group_divisor: 1
+  hop_length: 160
+  out_channels: 224
+  pooling_func: ASTP
+  return_2d_output: true
+  spec: null
+  spec_in_channels: 1
+  stages_setup:
+  - - - 1
+      - 1
+    - 3
+    - 3
+    - - - 3
+        - 3
+    - 64
+  - - - 2
+      - 1
+    - 4
+    - 2
+    - - - 3
+        - 3
+    - 64
+  - - - 1
+      - 2
+    - 5
+    - 2
+    - - - 3
+        - 3
+    - 48
+  - - - 2
+      - 1
+    - 5
+    - 1
+    - - - 3
+        - 3
+    - 48
+  - - - 1
+      - 2
+    - 4
+    - 0.75
+    - - - 3
+        - 3
+    - 32
+  - - - 2
+      - 1
+    - 3
+    - 0.5
+    - - - 3
+        - 3
+    - 24
+model_init: null
+noise_data: data/voxceleb/musan/lmdb
+num_avg: 1
+num_epochs: 5
+optimizer: SGD
+optimizer_args:
+  lr: 0.0001
+  momentum: 0.9
+  nesterov: true
+  weight_decay: 2.0e-05
+projection_args:
+  do_lm: false
+  easy_margin: false
+  embed_dim: 192
+  lanbuda: 0.7
+  margin_type: C
+  num_class: 5994
+  project_type: sphereface2
+  scale: 32.0
+  t: 3
+reverb_data: data/voxceleb/rirs/lmdb
+save_epoch_interval: 1
+scheduler: ExponentialDecrease
+scheduler_args:
+  epoch_iter: 34125
+  final_lr: 1.0e-05
+  initial_lr: 0.0001
+  num_epochs: 5
+  scale_ratio: 4.0
+  warm_from_zero: false
+  warm_up_epoch: 0
+seed: 42
+train_data: data/voxceleb/vox2_dev/shard.list
+train_label: data/voxceleb/vox2_dev/utt2spk

configs/fixed/JC.yaml

@@ -0,0 +1,104 @@
+audio:
+  chunk_size: 441000 # 44100 * 10s
+  min_mean_abs: 0.001
+  hop_length: 1024
+
+training:
+  batch_size: 8
+  gradient_accumulation_steps: 1
+  grad_clip: 0
+  segment: 10
+  shift: 1
+  samplerate: 44100
+  channels: 2
+  normalize: true
+  instruments: ['vp', 'bass', 'soprano', 'alto', 'tenor', 'lead_vocal']
+  target_instrument: null
+  num_epochs: 1000
+  num_steps: 1000
+  optimizer: adam
+  lr: 9.0e-05
+  patience: 2
+  reduce_factor: 0.95
+  q: 0.95
+  coarse_loss_clip: true
+  ema_momentum: 0.999
+  other_fix: false
+  use_amp: true
+
+augmentations:
+  enable: true
+  loudness: true
+  loudness_min: 0.5
+  loudness_max: 1.5
+  mixup: true
+  mixup_probs: [0.2, 0.02]
+  mixup_loudness_min: 0.5
+  mixup_loudness_max: 1.5
+  all:
+    channel_shuffle: 0.5
+    random_inverse: 0.1
+    random_polarity: 0.5
+
+inference:
+  num_overlap: 4
+  batch_size: 8
+
+model: htdemucs
+
+htdemucs:
+  channels: 48
+  channels_time:
+  growth: 2
+  num_subbands: 1
+  nfft: 4096
+  wiener_iters: 0
+  end_iters: 0
+  wiener_residual: false
+  cac: true
+  depth: 4
+  rewrite: true
+  multi_freqs: []
+  multi_freqs_depth: 3
+  freq_emb: 0.2
+  emb_scale: 10
+  emb_smooth: true
+  kernel_size: 8
+  stride: 4
+  time_stride: 2
+  context: 1
+  context_enc: 0
+  norm_starts: 4
+  norm_groups: 4
+  dconv_mode: 3
+  dconv_depth: 2
+  dconv_comp: 8
+  dconv_init: 1e-3
+  bottom_channels: 0
+  t_layers: 5
+  t_hidden_scale: 4.0
+  t_heads: 8
+  t_dropout: 0.0
+  t_layer_scale: True
+  t_gelu: True
+  t_emb: sin
+  t_max_positions: 10000
+  t_max_period: 10000.0
+  t_weight_pos_embed: 1.0
+  t_cape_mean_normalize: True
+  t_cape_augment: True
+  t_cape_glob_loc_scale: [5000.0, 1.0, 1.4]
+  t_sin_random_shift: 0
+  t_norm_in: True
+  t_norm_in_group: False
+  t_group_norm: False
+  t_norm_first: True
+  t_norm_out: True
+  t_weight_decay: 0.0
+  t_lr:
+  t_sparse_self_attn: False
+  t_sparse_cross_attn: False
+  t_mask_type: diag
+  t_mask_random_seed: 42
+  t_sparse_attn_window: 400
+  t_global_window: 100

configs/fixed/SK.yaml

@@ -0,0 +1,104 @@
+audio:
+  chunk_size: 441000 # 44100 * 10s
+  min_mean_abs: 0.001
+  hop_length: 1024
+
+training:
+  batch_size: 8
+  gradient_accumulation_steps: 1
+  grad_clip: 0
+  segment: 10
+  shift: 1
+  samplerate: 44100
+  channels: 2
+  normalize: true
+  instruments: ['vp', 'bass', 'soprano', 'alto', 'tenor', 'baritone']
+  target_instrument: null
+  num_epochs: 1000
+  num_steps: 1000
+  optimizer: adam
+  lr: 9.0e-05
+  patience: 2
+  reduce_factor: 0.95
+  q: 0.95
+  coarse_loss_clip: true
+  ema_momentum: 0.999
+  other_fix: false
+  use_amp: true
+
+augmentations:
+  enable: true
+  loudness: true
+  loudness_min: 0.5
+  loudness_max: 1.5
+  mixup: true
+  mixup_probs: [0.2, 0.02]
+  mixup_loudness_min: 0.5
+  mixup_loudness_max: 1.5
+  all:
+    channel_shuffle: 0.5
+    random_inverse: 0.1
+    random_polarity: 0.5
+
+inference:
+  num_overlap: 4
+  batch_size: 8
+
+model: htdemucs
+
+htdemucs:
+  channels: 48
+  channels_time:
+  growth: 2
+  num_subbands: 1
+  nfft: 4096
+  wiener_iters: 0
+  end_iters: 0
+  wiener_residual: false
+  cac: true
+  depth: 4
+  rewrite: true
+  multi_freqs: []
+  multi_freqs_depth: 3
+  freq_emb: 0.2
+  emb_scale: 10
+  emb_smooth: true
+  kernel_size: 8
+  stride: 4
+  time_stride: 2
+  context: 1
+  context_enc: 0
+  norm_starts: 4
+  norm_groups: 4
+  dconv_mode: 3
+  dconv_depth: 2
+  dconv_comp: 8
+  dconv_init: 1e-3
+  bottom_channels: 0
+  t_layers: 5
+  t_hidden_scale: 4.0
+  t_heads: 8
+  t_dropout: 0.0
+  t_layer_scale: True
+  t_gelu: True
+  t_emb: sin
+  t_max_positions: 10000
+  t_max_period: 10000.0
+  t_weight_pos_embed: 1.0
+  t_cape_mean_normalize: True
+  t_cape_augment: True
+  t_cape_glob_loc_scale: [5000.0, 1.0, 1.4]
+  t_sin_random_shift: 0
+  t_norm_in: True
+  t_norm_in_group: False
+  t_group_norm: False
+  t_norm_first: True
+  t_norm_out: True
+  t_weight_decay: 0.0
+  t_lr:
+  t_sparse_self_attn: False
+  t_sparse_cross_attn: False
+  t_mask_type: diag
+  t_mask_random_seed: 42
+  t_sparse_attn_window: 400
+  t_global_window: 100

configs/fixed/combined_ep333.yaml

@@ -0,0 +1,104 @@
+audio:
+  chunk_size: 441000 # 44100 * 10s
+  min_mean_abs: 0.001
+  hop_length: 1024
+
+training:
+  batch_size: 8
+  gradient_accumulation_steps: 1
+  grad_clip: 0
+  segment: 10
+  shift: 1
+  samplerate: 44100
+  channels: 2
+  normalize: true
+  instruments: ['vp', 'bass', 'soprano', 'alto', 'tenor', 'other']
+  target_instrument: null
+  num_epochs: 1000
+  num_steps: 1000
+  optimizer: adam
+  lr: 9.0e-05
+  patience: 2
+  reduce_factor: 0.95
+  q: 0.95
+  coarse_loss_clip: true
+  ema_momentum: 0.999
+  other_fix: false
+  use_amp: true
+
+augmentations:
+  enable: true
+  loudness: true
+  loudness_min: 0.5
+  loudness_max: 1.5
+  mixup: true
+  mixup_probs: [0.2, 0.02]
+  mixup_loudness_min: 0.5
+  mixup_loudness_max: 1.5
+  all:
+    channel_shuffle: 0.5
+    random_inverse: 0.1
+    random_polarity: 0.5
+
+inference:
+  num_overlap: 4
+  batch_size: 8
+
+model: htdemucs
+
+htdemucs:
+  channels: 48
+  channels_time:
+  growth: 2
+  num_subbands: 1
+  nfft: 4096
+  wiener_iters: 0
+  end_iters: 0
+  wiener_residual: false
+  cac: true
+  depth: 4
+  rewrite: true
+  multi_freqs: []
+  multi_freqs_depth: 3
+  freq_emb: 0.2
+  emb_scale: 10
+  emb_smooth: true
+  kernel_size: 8
+  stride: 4
+  time_stride: 2
+  context: 1
+  context_enc: 0
+  norm_starts: 4
+  norm_groups: 4
+  dconv_mode: 3
+  dconv_depth: 2
+  dconv_comp: 8
+  dconv_init: 1e-3
+  bottom_channels: 0
+  t_layers: 5
+  t_hidden_scale: 4.0
+  t_heads: 8
+  t_dropout: 0.0
+  t_layer_scale: True
+  t_gelu: True
+  t_emb: sin
+  t_max_positions: 10000
+  t_max_period: 10000.0
+  t_weight_pos_embed: 1.0
+  t_cape_mean_normalize: True
+  t_cape_augment: True
+  t_cape_glob_loc_scale: [5000.0, 1.0, 1.4]
+  t_sin_random_shift: 0
+  t_norm_in: True
+  t_norm_in_group: False
+  t_group_norm: False
+  t_norm_first: True
+  t_norm_out: True
+  t_weight_decay: 0.0
+  t_lr:
+  t_sparse_self_attn: False
+  t_sparse_cross_attn: False
+  t_mask_type: diag
+  t_mask_random_seed: 42
+  t_sparse_attn_window: 400
+  t_global_window: 100

configs/specialists/JC/alto.yaml

@@ -0,0 +1,104 @@
+audio:
+  chunk_size: 441000 # 44100 * 10s
+  min_mean_abs: 0.001
+  hop_length: 1024
+
+training:
+  batch_size: 8
+  gradient_accumulation_steps: 1
+  grad_clip: 0
+  segment: 10
+  shift: 1
+  samplerate: 44100
+  channels: 2
+  normalize: true
+  instruments: ['nonalto', 'alto']
+  target_instrument: null
+  num_epochs: 1000
+  num_steps: 1000
+  optimizer: adam
+  lr: 9.0e-05
+  patience: 2
+  reduce_factor: 0.95
+  q: 0.95
+  coarse_loss_clip: true
+  ema_momentum: 0.999
+  other_fix: false
+  use_amp: true
+
+augmentations:
+  enable: true
+  loudness: true
+  loudness_min: 0.5
+  loudness_max: 1.5
+  mixup: true
+  mixup_probs: [0.2, 0.02]
+  mixup_loudness_min: 0.5
+  mixup_loudness_max: 1.5
+  all:
+    channel_shuffle: 0.5
+    random_inverse: 0.1
+    random_polarity: 0.5
+
+inference:
+  num_overlap: 4
+  batch_size: 8
+
+model: htdemucs
+
+htdemucs:
+  channels: 48
+  channels_time:
+  growth: 2
+  num_subbands: 1
+  nfft: 4096
+  wiener_iters: 0
+  end_iters: 0
+  wiener_residual: false
+  cac: true
+  depth: 4
+  rewrite: true
+  multi_freqs: []
+  multi_freqs_depth: 3
+  freq_emb: 0.2
+  emb_scale: 10
+  emb_smooth: true
+  kernel_size: 8
+  stride: 4
+  time_stride: 2
+  context: 1
+  context_enc: 0
+  norm_starts: 4
+  norm_groups: 4
+  dconv_mode: 3
+  dconv_depth: 2
+  dconv_comp: 8
+  dconv_init: 1e-3
+  bottom_channels: 0
+  t_layers: 5
+  t_hidden_scale: 4.0
+  t_heads: 8
+  t_dropout: 0.0
+  t_layer_scale: True
+  t_gelu: True
+  t_emb: sin
+  t_max_positions: 10000
+  t_max_period: 10000.0
+  t_weight_pos_embed: 1.0
+  t_cape_mean_normalize: True
+  t_cape_augment: True
+  t_cape_glob_loc_scale: [5000.0, 1.0, 1.4]
+  t_sin_random_shift: 0
+  t_norm_in: True
+  t_norm_in_group: False
+  t_group_norm: False
+  t_norm_first: True
+  t_norm_out: True
+  t_weight_decay: 0.0
+  t_lr:
+  t_sparse_self_attn: False
+  t_sparse_cross_attn: False
+  t_mask_type: diag
+  t_mask_random_seed: 42
+  t_sparse_attn_window: 400
+  t_global_window: 100

configs/specialists/JC/bass.yaml

@@ -0,0 +1,104 @@
+audio:
+  chunk_size: 441000 # 44100 * 10s
+  min_mean_abs: 0.001
+  hop_length: 1024
+
+training:
+  batch_size: 8
+  gradient_accumulation_steps: 1
+  grad_clip: 0
+  segment: 10
+  shift: 1
+  samplerate: 44100
+  channels: 2
+  normalize: true
+  instruments: ['nonbass', 'bass']
+  target_instrument: null
+  num_epochs: 1000
+  num_steps: 1000
+  optimizer: adam
+  lr: 9.0e-05
+  patience: 2
+  reduce_factor: 0.95
+  q: 0.95
+  coarse_loss_clip: true
+  ema_momentum: 0.999
+  other_fix: false
+  use_amp: true
+
+augmentations:
+  enable: true
+  loudness: true
+  loudness_min: 0.5
+  loudness_max: 1.5
+  mixup: true
+  mixup_probs: [0.2, 0.02]
+  mixup_loudness_min: 0.5
+  mixup_loudness_max: 1.5
+  all:
+    channel_shuffle: 0.5
+    random_inverse: 0.1
+    random_polarity: 0.5
+
+inference:
+  num_overlap: 4
+  batch_size: 8
+
+model: htdemucs
+
+htdemucs:
+  channels: 48
+  channels_time:
+  growth: 2
+  num_subbands: 1
+  nfft: 4096
+  wiener_iters: 0
+  end_iters: 0
+  wiener_residual: false
+  cac: true
+  depth: 4
+  rewrite: true
+  multi_freqs: []
+  multi_freqs_depth: 3
+  freq_emb: 0.2
+  emb_scale: 10
+  emb_smooth: true
+  kernel_size: 8
+  stride: 4
+  time_stride: 2
+  context: 1
+  context_enc: 0
+  norm_starts: 4
+  norm_groups: 4
+  dconv_mode: 3
+  dconv_depth: 2
+  dconv_comp: 8
+  dconv_init: 1e-3
+  bottom_channels: 0
+  t_layers: 5
+  t_hidden_scale: 4.0
+  t_heads: 8
+  t_dropout: 0.0
+  t_layer_scale: True
+  t_gelu: True
+  t_emb: sin
+  t_max_positions: 10000
+  t_max_period: 10000.0
+  t_weight_pos_embed: 1.0
+  t_cape_mean_normalize: True
+  t_cape_augment: True
+  t_cape_glob_loc_scale: [5000.0, 1.0, 1.4]
+  t_sin_random_shift: 0
+  t_norm_in: True
+  t_norm_in_group: False
+  t_group_norm: False
+  t_norm_first: True
+  t_norm_out: True
+  t_weight_decay: 0.0
+  t_lr:
+  t_sparse_self_attn: False
+  t_sparse_cross_attn: False
+  t_mask_type: diag
+  t_mask_random_seed: 42
+  t_sparse_attn_window: 400
+  t_global_window: 100

configs/specialists/JC/lead_vocal.yaml

@@ -0,0 +1,104 @@
+audio:
+  chunk_size: 441000 # 44100 * 10s
+  min_mean_abs: 0.001
+  hop_length: 1024
+
+training:
+  batch_size: 8
+  gradient_accumulation_steps: 1
+  grad_clip: 0
+  segment: 10
+  shift: 1
+  samplerate: 44100
+  channels: 2
+  normalize: true
+  instruments: ['nonlead_vocal', 'lead_vocal']
+  target_instrument: null
+  num_epochs: 1000
+  num_steps: 1000
+  optimizer: adam
+  lr: 9.0e-05
+  patience: 2
+  reduce_factor: 0.95
+  q: 0.95
+  coarse_loss_clip: true
+  ema_momentum: 0.999
+  other_fix: false
+  use_amp: true
+
+augmentations:
+  enable: true
+  loudness: true
+  loudness_min: 0.5
+  loudness_max: 1.5
+  mixup: true
+  mixup_probs: [0.2, 0.02]
+  mixup_loudness_min: 0.5
+  mixup_loudness_max: 1.5
+  all:
+    channel_shuffle: 0.5
+    random_inverse: 0.1
+    random_polarity: 0.5
+
+inference:
+  num_overlap: 4
+  batch_size: 8
+
+model: htdemucs
+
+htdemucs:
+  channels: 48
+  channels_time:
+  growth: 2
+  num_subbands: 1
+  nfft: 4096
+  wiener_iters: 0
+  end_iters: 0
+  wiener_residual: false
+  cac: true
+  depth: 4
+  rewrite: true
+  multi_freqs: []
+  multi_freqs_depth: 3
+  freq_emb: 0.2
+  emb_scale: 10
+  emb_smooth: true
+  kernel_size: 8
+  stride: 4
+  time_stride: 2
+  context: 1
+  context_enc: 0
+  norm_starts: 4
+  norm_groups: 4
+  dconv_mode: 3
+  dconv_depth: 2
+  dconv_comp: 8
+  dconv_init: 1e-3
+  bottom_channels: 0
+  t_layers: 5
+  t_hidden_scale: 4.0
+  t_heads: 8
+  t_dropout: 0.0
+  t_layer_scale: True
+  t_gelu: True
+  t_emb: sin
+  t_max_positions: 10000
+  t_max_period: 10000.0
+  t_weight_pos_embed: 1.0
+  t_cape_mean_normalize: True
+  t_cape_augment: True
+  t_cape_glob_loc_scale: [5000.0, 1.0, 1.4]
+  t_sin_random_shift: 0
+  t_norm_in: True
+  t_norm_in_group: False
+  t_group_norm: False
+  t_norm_first: True
+  t_norm_out: True
+  t_weight_decay: 0.0
+  t_lr:
+  t_sparse_self_attn: False
+  t_sparse_cross_attn: False
+  t_mask_type: diag
+  t_mask_random_seed: 42
+  t_sparse_attn_window: 400
+  t_global_window: 100

configs/specialists/JC/soprano.yaml

@@ -0,0 +1,104 @@
+audio:
+  chunk_size: 441000 # 44100 * 10s
+  min_mean_abs: 0.001
+  hop_length: 1024
+
+training:
+  batch_size: 8
+  gradient_accumulation_steps: 1
+  grad_clip: 0
+  segment: 10
+  shift: 1
+  samplerate: 44100
+  channels: 2
+  normalize: true
+  instruments: ['nonsoprano', 'soprano']
+  target_instrument: null
+  num_epochs: 1000
+  num_steps: 1000
+  optimizer: adam
+  lr: 9.0e-05
+  patience: 2
+  reduce_factor: 0.95
+  q: 0.95
+  coarse_loss_clip: true
+  ema_momentum: 0.999
+  other_fix: false
+  use_amp: true
+
+augmentations:
+  enable: true
+  loudness: true
+  loudness_min: 0.5
+  loudness_max: 1.5
+  mixup: true
+  mixup_probs: [0.2, 0.02]
+  mixup_loudness_min: 0.5
+  mixup_loudness_max: 1.5
+  all:
+    channel_shuffle: 0.5
+    random_inverse: 0.1
+    random_polarity: 0.5
+
+inference:
+  num_overlap: 4
+  batch_size: 8
+
+model: htdemucs
+
+htdemucs:
+  channels: 48
+  channels_time:
+  growth: 2
+  num_subbands: 1
+  nfft: 4096
+  wiener_iters: 0
+  end_iters: 0
+  wiener_residual: false
+  cac: true
+  depth: 4
+  rewrite: true
+  multi_freqs: []
+  multi_freqs_depth: 3
+  freq_emb: 0.2
+  emb_scale: 10
+  emb_smooth: true
+  kernel_size: 8
+  stride: 4
+  time_stride: 2
+  context: 1
+  context_enc: 0
+  norm_starts: 4
+  norm_groups: 4
+  dconv_mode: 3
+  dconv_depth: 2
+  dconv_comp: 8
+  dconv_init: 1e-3
+  bottom_channels: 0
+  t_layers: 5
+  t_hidden_scale: 4.0
+  t_heads: 8
+  t_dropout: 0.0
+  t_layer_scale: True
+  t_gelu: True
+  t_emb: sin
+  t_max_positions: 10000
+  t_max_period: 10000.0
+  t_weight_pos_embed: 1.0
+  t_cape_mean_normalize: True
+  t_cape_augment: True
+  t_cape_glob_loc_scale: [5000.0, 1.0, 1.4]
+  t_sin_random_shift: 0
+  t_norm_in: True
+  t_norm_in_group: False
+  t_group_norm: False
+  t_norm_first: True
+  t_norm_out: True
+  t_weight_decay: 0.0
+  t_lr:
+  t_sparse_self_attn: False
+  t_sparse_cross_attn: False
+  t_mask_type: diag
+  t_mask_random_seed: 42
+  t_sparse_attn_window: 400
+  t_global_window: 100