MIDI / transcriptionGPU workloadrubberband · Source beat_this · Source omnizart · Source BS-Roformer-Leap · Weights Rubber Band R3 + Beat This! + ChordMini BTC + Leap XE 90 + Leap Instrumental · Paper Rubber Band R3 + Beat This! + ChordMini BTC + Leap XE 90 + Leap Instrumental · Paper Rubber Band R3 + Beat This! + ChordMini BTC + Leap XE 90 + Leap Instrumental · Paper

Song BPM Alignment, Stems & DAW Export

Preserve original speed or choose a target BPM. Analyze beats, key and chords; optionally separate vocals and accompaniment. Export WAV, MIDI/CSV, song tags and verifiable analysis.

30 credits/useFile inputMax 100 MBEstimated 30-3300 seconds

Input

Supports inputs in formats such as mp3 / wav / flac / ogg / m4a / aac / aif / aiff / mp4 / mov / mkv

Supported formats

mp3wavflacoggm4aaacaifaiffmp4movmkv

Output

WAV / MIDI / CSV / JSON

Use case

Analyze tempo and harmony for arranging, mixing and DAW projects.

Quick feedback

Include the tool name, steps and the problem in your feedback. Screenshots can help us locate it.

Models, parameters, and sources

Rubber Band R3 + Beat This! + ChordMini BTC + Leap XE 90 + Leap InstrumentalBeat This! paper, Table 1Beat This!TelkNet · 2026-09-13Omnizart ChordV1 (raw)Omnizart ChordV1 (CSV)ChordMini BTC (CL full)ChordMini 2E1D (CL full)TelkNet · chord quality · 2026-09-132602.19778 · S4.T2Bortolozzo et al.†Li et al.†

Rubber Band R3 + Beat This! + ChordMini BTC + Leap XE 90 + Leap Instrumental

Rubber Band 4.0.0 source (pinned)Beat This! source (pinned)Omnizart source (pinned)Leap XE 90 vocal weights BS-RoFormer Leap Instrumental accompaniment weights Rubber Band Library official site Beat This! ISMIR 2024 paper Omnizart JOSS paper

Leave empty to preserve speed, or enter 30–300 BPM to change the actual audio tempo. Constant stretching does not force every detected beat onto a grid.
Choose the detected BPM, half-time or double-time, or enter the source BPM manually.
Meter of the song. Keep automatic detection, or pin 2 to 12 beats per bar when the downbeat grouping is known.

Paper and benchmark notes

Rubber Band Library: high-quality audio time-stretching and pitch-shifting

Official documentation - 2024

Official site Source (pinned)

Beat this! Accurate beat tracking without DBN postprocessing

ISMIR - 2024

Beat This! final0 detects beat and downbeat times. TelkNet derives BPM, a normalized beat grid, and fixed-versus-variable-tempo warnings; OpenKeyScan3 adds global key, Camelot, and Open Key notation. ChordMini BTC is the default; switch to 2E1D or Omnizart to compare. Chord MIDI provides harmony audition, not the original performance.

Paper

Omnizart: A General Toolbox for Automatic Music Transcription

JOSS - 2021

Analyze a song chord by chord and export the progression as MIDI plus a timestamped chord table. ChordMini BTC is the default; switch to 2E1D or Omnizart to compare. Chord MIDI provides harmony audition, not the original performance.

Paper

Official / repositories

Rubber Band 4.0.0 source (pinned)Beat This! source (pinned)Omnizart source (pinned)Leap XE 90 vocal weights BS-RoFormer Leap Instrumental accompaniment weights

Papers / technical notes

Rubber Band Library official site Beat This! ISMIR 2024 paper Omnizart JOSS paper

Parameter guide

Leave empty to preserve speed, or enter 30–300 BPM to change the actual audio tempo. Constant stretching does not force every detected beat onto a grid.
Change speed by one ratio across the song, preserving its internal tempo variations.
Change the speed of the whole song by one ratio; suitable for recordings with a steady beat.
Explicitly stretches between beats. Local speed can vary; review the reported range and timing warnings.
Choose the detected BPM, half-time or double-time, or enter the source BPM manually.
Meter of the song. Keep automatic detection, or pin 2 to 12 beats per bar when the downbeat grouping is known.
Empty target keeps the original speed. A numeric target changes the audio tempo; constant mode preserves natural timing differences.
Export WAV files at their original duration and pitch, together with a MIDI tempo map of the source song.
Export the target-speed version and the original-speed full song with its MIDI tempo map.
If the detected tempo is incorrect, enter the correct source BPM here.
Position of the first downbeat in seconds. Anchors the grid when automatic detection starts off-beat.

Workflow

1Select songs; leave target BPM blank to retain their speed or enter a new tempo.
2Choose whether to split vocals and accompaniment, then export the matching WAV, MIDI and analysis files.

What to check

Leave empty to preserve speed, or enter 30–300 BPM to change the actual audio tempo. Constant stretching does not force every detected beat onto a grid.

Models in this field

220 entries

Includes public test results for models used here and elsewhere. The complete source table is shown by default; conditions and metrics can also be selected for comparison. Scores from different test conditions are not directly comparable.

Only records with verified public inference implementations and matching weights are shown; algorithms requiring no weights are also eligible. Closed services, unreleased weights and unmatched experiment versions are excluded. This is not an all-model leaderboard.

Public weights do not establish a fully open-source license. Expand the data for declared weight licenses and access conditions. Undeclared licenses are marked separately.

BenchmarkRanking category

The selected model has no verified matching public result and has not been ranked.

TelkNet · 2026-09-13

Choose a dataset to view all metrics for each model. Rows follow the official rank or selected metric. Bold marks each column’s best comparable value, including ties. Different evaluation protocols are ranked separately; missing scores are shown as -.

Deployable entries 220 / source rows 220

Frozen 2026-09-13 evaluation: 285 inputs, 855 inferences. References exist for 208 GuitarSet and 48 SWD inputs; 29 local inputs have no independent truth and no accuracy score. All 55 groups are selectable, including solo, style, performer and reference type: performed, instructed and human. WCSR weights eligible duration; macro averages tracks; bar scores use reference bars. Expand each metric’s included/excluded duration. Omnizart raw and CSV remain separate. Results do not guarantee arbitrary-mix accuracy or zero training overlap.

Test settingRank by metric

TelkNet · 2026-09-13
Metric rank	Model	Test setting	root WCSR (%)↑	majmin WCSR (%)↑	triads WCSR (%)↑	sevenths WCSR (%)↑	tetrads WCSR (%)↑	tetrads_inv WCSR (%)↑	full_pitch_classes WCSR (%)↑	full_pitch_classes_inv WCSR (%)↑	root macro (%)↑	majmin macro (%)↑	triads macro (%)↑	sevenths macro (%)↑	tetrads macro (%)↑	tetrads_inv macro (%)↑	full_pitch_classes macro (%)↑	full_pitch_classes_inv macro (%)↑	Bar majmin macro (%)↑	Boundary F1 (±0.25 s, %)↑	Timeline coverage (%)	N	Explicit interval set (%)↑	11 (%)↑	7 (%)↑	9 (%)↑	aug (%)↑	dim (%)↑	dim7 (%)↑	hdim7 (%)↑	maj (%)↑	maj6 (%)↑	maj7 (%)↑	maj9 (%)↑	min (%)↑	min11 (%)↑	min6 (%)↑	min7 (%)↑	min9 (%)↑	minmaj7 (%)↑	sus2 (%)↑	sus4 (%)↑
1	ChordMini BTC (CL full) Public weights · Inference code checkpoints/btc_model_best.pth Inference codeWeight license not declared	GuitarSet 1.1.0 · comping · performed	81.707611 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	86.753748 Evaluation duration Included 3631.09785836346 s; excluded 1813.7882414097794 s	58.609709 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	79.801292 Evaluation duration Included 3502.102646060085 s; excluded 1942.7834537131555 s	52.045132 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	41.559627 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	52.045132 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	41.559627 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	80.737427 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	83.252875 Evaluation duration Included 3631.09785836346 s; excluded 1813.7882414097794 s	57.418731 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	74.094441 Evaluation duration Included 3502.102646060085 s; excluded 1942.7834537131555 s	50.303619 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	40.145104 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	50.303619 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	40.145104 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	86.243841	68.995164	99.853948	178	0 Evaluation duration Included 627.8152749108 s; excluded 0 s	0 Evaluation duration Included 9.636145367237805 s; excluded 0 s	38.699908 Evaluation duration Included 604.004809855832 s; excluded 0 s	16.481819 Evaluation duration Included 118.59403326525016 s; excluded 0 s	0 Evaluation duration Included 16.13639543462473 s; excluded 0 s	-	0 Evaluation duration Included 3.848867791391365 s; excluded 0 s	19.041616 Evaluation duration Included 137.8963054206376 s; excluded 0 s	85.468379 Evaluation duration Included 2083.6452969301185 s; excluded 0 s	0 Evaluation duration Included 160.3285413212152 s; excluded 0 s	36.924994 Evaluation duration Included 413.7901856070912 s; excluded 0 s	0 Evaluation duration Included 21.500649381669504 s; excluded 0 s	80.845935 Evaluation duration Included 636.3085731540472 s; excluded 0 s	44.816335 Evaluation duration Included 4.363636363636363 s; excluded 0 s	0 Evaluation duration Included 20.65687772122959 s; excluded 0 s	25.338741 Evaluation duration Included 355.3293406230003 s; excluded 0 s	2.595289 Evaluation duration Included 52.26089039122396 s; excluded 0 s	0 Evaluation duration Included 2.0168067226890756 s; excluded 0 s	7.980445 Evaluation duration Included 117.54809586145142 s; excluded 0 s	5.80416 Evaluation duration Included 59.20513963649113 s; excluded 0 s
2	ChordMini 2E1D (CL full) Public weights · Inference code checkpoints/2e1d_model_best.pth Inference codeWeight license not declared	GuitarSet 1.1.0 · comping · performed	73.805395 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	79.597236 Evaluation duration Included 3631.09785836346 s; excluded 1813.7882414097799 s	53.334453 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	74.174219 Evaluation duration Included 3502.102646060085 s; excluded 1942.7834537131555 s	47.955555 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	39.150516 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	47.955555 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	39.150516 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	72.442449 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	75.692234 Evaluation duration Included 3631.09785836346 s; excluded 1813.7882414097799 s	52.052753 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	68.453329 Evaluation duration Included 3502.102646060085 s; excluded 1942.7834537131555 s	46.284043 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	37.7932 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	46.284043 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	37.7932 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	78.518746	61.758947	99.853948	178	0 Evaluation duration Included 627.8152749108 s; excluded 0 s	0 Evaluation duration Included 9.636145367237805 s; excluded 0 s	35.613936 Evaluation duration Included 604.004809855832 s; excluded 0 s	12.127208 Evaluation duration Included 118.59403326525016 s; excluded 0 s	0 Evaluation duration Included 16.13639543462473 s; excluded 0 s	-	0 Evaluation duration Included 3.848867791391365 s; excluded 0 s	0 Evaluation duration Included 137.8963054206376 s; excluded 0 s	79.745007 Evaluation duration Included 2083.6452969301185 s; excluded 0 s	0 Evaluation duration Included 160.3285413212152 s; excluded 0 s	35.322424 Evaluation duration Included 413.7901856070912 s; excluded 0 s	0 Evaluation duration Included 21.500649381669504 s; excluded 0 s	75.792358 Evaluation duration Included 636.3085731540472 s; excluded 0 s	40.559335 Evaluation duration Included 4.363636363636363 s; excluded 0 s	0 Evaluation duration Included 20.65687772122959 s; excluded 0 s	21.48945 Evaluation duration Included 355.3293406230003 s; excluded 0 s	0 Evaluation duration Included 52.26089039122396 s; excluded 0 s	0 Evaluation duration Included 2.0168067226890756 s; excluded 0 s	8.533709 Evaluation duration Included 117.54809586145142 s; excluded 0 s	5.804773 Evaluation duration Included 59.20513963649113 s; excluded 0 s
3	Omnizart ChordV1 (raw) Public weights · Inference code chord_v1 complete SavedModel Inference codeWeight license not declared	GuitarSet 1.1.0 · comping · performed	70.681852 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	78.38993 Evaluation duration Included 3631.09785836346 s; excluded 1813.7882414097794 s	52.276855 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	59.698581 Evaluation duration Included 3502.102646060085 s; excluded 1942.7834537131555 s	38.397602 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	30.586609 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	38.397602 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	30.586609 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	69.887902 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	75.061246 Evaluation duration Included 3631.09785836346 s; excluded 1813.7882414097794 s	51.23753 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	52.338084 Evaluation duration Included 3502.102646060085 s; excluded 1942.7834537131555 s	36.493738 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	28.824622 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	36.493738 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	28.824622 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	77.729493	38.652642	100	178	0 Evaluation duration Included 627.8152749108 s; excluded 0 s	0 Evaluation duration Included 9.636145367237805 s; excluded 0 s	0 Evaluation duration Included 604.004809855832 s; excluded 0 s	0 Evaluation duration Included 118.59403326525016 s; excluded 0 s	0 Evaluation duration Included 16.13639543462473 s; excluded 0 s	-	0 Evaluation duration Included 3.848867791391365 s; excluded 0 s	0 Evaluation duration Included 137.8963054206376 s; excluded 0 s	79.618967 Evaluation duration Included 2083.6452969301185 s; excluded 0 s	0 Evaluation duration Included 160.3285413212152 s; excluded 0 s	0 Evaluation duration Included 413.7901856070912 s; excluded 0 s	0 Evaluation duration Included 21.500649381669504 s; excluded 0 s	67.848955 Evaluation duration Included 636.3085731540472 s; excluded 0 s	0 Evaluation duration Included 4.363636363636363 s; excluded 0 s	0 Evaluation duration Included 20.65687772122959 s; excluded 0 s	0 Evaluation duration Included 355.3293406230003 s; excluded 0 s	0 Evaluation duration Included 52.26089039122396 s; excluded 0 s	0 Evaluation duration Included 2.0168067226890756 s; excluded 0 s	0 Evaluation duration Included 117.54809586145142 s; excluded 0 s	0 Evaluation duration Included 59.20513963649113 s; excluded 0 s
4	Omnizart ChordV1 (CSV) Public weights · Inference code chord_v1 complete SavedModel Inference codeWeight license not declared	GuitarSet 1.1.0 · comping · performed	71.032285 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	78.312621 Evaluation duration Included 3631.09785836346 s; excluded 1813.7882414097794 s	52.225299 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	59.555869 Evaluation duration Included 3502.102646060085 s; excluded 1942.7834537131555 s	38.305811 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	30.50625 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	38.305811 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	30.50625 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	70.180979 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	75.228274 Evaluation duration Included 3631.09785836346 s; excluded 1813.7882414097794 s	51.186068 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	52.503452 Evaluation duration Included 3502.102646060085 s; excluded 1942.7834537131555 s	36.40965 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	28.749232 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	36.40965 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	28.749232 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	77.615253	58.893914	89.774904	178	0 Evaluation duration Included 627.8152749108 s; excluded 0 s	0 Evaluation duration Included 9.636145367237805 s; excluded 0 s	0 Evaluation duration Included 604.004809855832 s; excluded 0 s	0 Evaluation duration Included 118.59403326525016 s; excluded 0 s	0 Evaluation duration Included 16.13639543462473 s; excluded 0 s	-	0 Evaluation duration Included 3.848867791391365 s; excluded 0 s	0 Evaluation duration Included 137.8963054206376 s; excluded 0 s	79.548417 Evaluation duration Included 2083.6452969301185 s; excluded 0 s	0 Evaluation duration Included 160.3285413212152 s; excluded 0 s	0 Evaluation duration Included 413.7901856070912 s; excluded 0 s	0 Evaluation duration Included 21.500649381669504 s; excluded 0 s	67.294524 Evaluation duration Included 636.3085731540472 s; excluded 0 s	0 Evaluation duration Included 4.363636363636363 s; excluded 0 s	0 Evaluation duration Included 20.65687772122959 s; excluded 0 s	0 Evaluation duration Included 355.3293406230003 s; excluded 0 s	0 Evaluation duration Included 52.26089039122396 s; excluded 0 s	0 Evaluation duration Included 2.0168067226890756 s; excluded 0 s	0 Evaluation duration Included 117.54809586145142 s; excluded 0 s	0 Evaluation duration Included 59.20513963649113 s; excluded 0 s

Original records (220)

TelkNet · 2026-09-13 ↗Download complete evaluation dataChecked: 2026-10-25

All evaluation sources (7)

Datasets, versions and conditions remain separate, including sources with no eligible ranking rows.

Beat This! paper, Table 1 ↗
Deployable entries 1 / source rows 2
Not ranked: 1
A complete public inference package for these experiments has not been verified. Other weights with the same model name do not qualify.
Hung
TelkNet · 2026-09-13 ↗
Deployable entries 220 / source rows 220
2602.19778 · S4.T2 ↗
Deployable entries 2 / source rows 8
Not ranked: 6
A complete public inference package for these experiments has not been verified. Other weights with the same model name do not qualify.
Bortolozzo et al. † · Prior Work (Noisy Student Framework) · Li et al. † · Prior Work (Noisy Student Framework) · 2E1D (SL) (2.2M) · Traditional supervised learning (SL) baseline (trained on full labels with data augmentation) · BTC (SL) (3.03M) · Traditional supervised learning (SL) baseline (trained on full labels with data augmentation) · 2E1D (CL) (50%) · Results of our student models after Stage 2 with Data-Incremental Continual Learning (with clean labels, α = 0.3 ) · BTC (CL) (50%) · Results of our student models after Stage 2 with Data-Incremental Continual Learning (with clean labels, α = 0.3 )
2602.19778 · S3.T1 ↗
Deployable entries 1 / source rows 11
Not ranked: 10
A complete public inference package for these experiments has not been verified. Other weights with the same model name do not qualify.
2E1D (2.2M) · Our two student models trained only with pseudo-labels (generated by the BTC teacher from unlabeled data) · FMA (short-form) · 2E1D (2.2M) · Our two student models trained only with pseudo-labels (generated by the BTC teacher from unlabeled data) · DALI (long-form) · 2E1D (2.2M) · Our two student models trained only with pseudo-labels (generated by the BTC teacher from unlabeled data) · MAESTRO+DALI · 2E1D (2.2M) · Our two student models trained only with pseudo-labels (generated by the BTC teacher from unlabeled data) · All (FMA+M+D) · 2E1D (2.2M) · Our two student models trained only with pseudo-labels (generated by the BTC teacher from unlabeled data) · All + KD ( α = 0.3 ) · BTC (3.03M) · Our two student models trained only with pseudo-labels (generated by the BTC teacher from unlabeled data) · FMA (short-form) · BTC (3.03M) · Our two student models trained only with pseudo-labels (generated by the BTC teacher from unlabeled data) · DALI (long-form) · BTC (3.03M) · Our two student models trained only with pseudo-labels (generated by the BTC teacher from unlabeled data) · MAESTRO+DALI · BTC (3.03M) · Our two student models trained only with pseudo-labels (generated by the BTC teacher from unlabeled data) · All (FMA+M+D) · BTC (3.03M) · Our two student models trained only with pseudo-labels (generated by the BTC teacher from unlabeled data) · All + KD ( α = 0.3 )
2602.19778 · S5.T3 ↗
Deployable entries 2 / source rows 8
Not ranked: 6
A complete public inference package for these experiments has not been verified. Other weights with the same model name do not qualify.
Count · Bortolozzo et al. † · 2E1D (SL) · BTC (SL) · 2E1D (CL) (50%) · BTC (CL) (50%)
2602.19778 · S5.T4 ↗
Deployable entries 0 / source rows 8
Not ranked: 8
A complete public inference package for these experiments has not been verified. Other weights with the same model name do not qualify.
2E1D · 0 · 2E1D · 0.1 · 2E1D · 0.3 · 2E1D · 0.5 · BTC · 0 · BTC · 0.1 · BTC · 0.3 · BTC · 0.4
Harmony Transformer paper, Table 1 ↗
Deployable entries 0 / source rows 4
Not ranked: 4
A complete public inference package for these experiments has not been verified. Other weights with the same model name do not qualify.
BLSTM · FK · HT · HT*

All original results and sources

Scores apply to the listed dataset, metric and model version; results from different tests are not directly comparable.

Beat This! paper, Table 1

Deployable entries 1 / source rows 2

Not ranked: 1

A complete public inference package for these experiments has not been verified. Other weights with the same model name do not qualify.

Hung

The beat grid behind the alignment is analyzed by the pinned Beat This! final0 checkpoint, so these rows are the paper's published per-dataset scores for that engine. Rubber Band R3 performs the actual stretch and publishes no comparable public table.

Rank by metric

Beat This! paper, Table 1
Metric rank	Model	ASAP		Ballroom		Beatles		Candombe		Filosax		Groove MIDI		GuitarSet		Hainsworth		Harmonix		HJDB		JAAH		RWC Classical		RWC Jazz		RWC Pop		RWC RF		Simac		SMC		TapCorrect
Metric rank	Model	Beat F1↑	Downbeat F1↑	Beat F1↑	Downbeat F1↑	Beat F1↑	Downbeat F1↑	Beat F1↑	Downbeat F1↑	Beat F1↑	Downbeat F1↑	Beat F1↑	Downbeat F1↑	Beat F1↑	Downbeat F1↑	Beat F1↑	Downbeat F1↑	Beat F1↑	Downbeat F1↑	Beat F1↑	Downbeat F1↑	Beat F1↑	Downbeat F1↑	Beat F1↑	Downbeat F1↑	Beat F1↑	Downbeat F1↑	Beat F1↑	Downbeat F1↑	Beat F1↑	Downbeat F1↑	Beat F1↑	Downbeat F1↑	Beat F1↑	Downbeat F1↑	Beat F1↑	Downbeat F1↑
-	Beat This! Public weights · Inference code fold0.ckpt through fold7.ckpt · 8-fold cross-validation Inference code Weight license (repository declaration): MIT	76.3	61.2	97.5	95.3	94.5	88.8	99.7	99.7	99.5	98.5	93.7	82.1	92.0	88.1	91.9	80.0	95.8	90.7	98.2	96.6	95.1	85.0	77.1	66.3	83.3	80.7	96.1	93.7	94.5	91.9	77.9	-	62.7	-	93.0	86.4

Original records (1)

Beat This! paper, Table 1 ↗Checked: 2026-10-25

TelkNet · 2026-09-13

Deployable entries 220 / source rows 220

Test settingRank by metric

TelkNet · 2026-09-13
Metric rank	Model	Test setting	root WCSR (%)↑	majmin WCSR (%)↑	triads WCSR (%)↑	sevenths WCSR (%)↑	tetrads WCSR (%)↑	tetrads_inv WCSR (%)↑	full_pitch_classes WCSR (%)↑	full_pitch_classes_inv WCSR (%)↑	root macro (%)↑	majmin macro (%)↑	triads macro (%)↑	sevenths macro (%)↑	tetrads macro (%)↑	tetrads_inv macro (%)↑	full_pitch_classes macro (%)↑	full_pitch_classes_inv macro (%)↑	Bar majmin macro (%)↑	Boundary F1 (±0.25 s, %)↑	Timeline coverage (%)	N	Explicit interval set (%)↑	11 (%)↑	7 (%)↑	9 (%)↑	aug (%)↑	dim (%)↑	dim7 (%)↑	hdim7 (%)↑	maj (%)↑	maj6 (%)↑	maj7 (%)↑	maj9 (%)↑	min (%)↑	min11 (%)↑	min6 (%)↑	min7 (%)↑	min9 (%)↑	minmaj7 (%)↑	sus2 (%)↑	sus4 (%)↑
1	ChordMini BTC (CL full) Public weights · Inference code checkpoints/btc_model_best.pth Inference codeWeight license not declared	GuitarSet 1.1.0 · comping · performed	81.707611 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	86.753748 Evaluation duration Included 3631.09785836346 s; excluded 1813.7882414097794 s	58.609709 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	79.801292 Evaluation duration Included 3502.102646060085 s; excluded 1942.7834537131555 s	52.045132 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	41.559627 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	52.045132 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	41.559627 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	80.737427 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	83.252875 Evaluation duration Included 3631.09785836346 s; excluded 1813.7882414097794 s	57.418731 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	74.094441 Evaluation duration Included 3502.102646060085 s; excluded 1942.7834537131555 s	50.303619 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	40.145104 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	50.303619 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	40.145104 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	86.243841	68.995164	99.853948	178	0 Evaluation duration Included 627.8152749108 s; excluded 0 s	0 Evaluation duration Included 9.636145367237805 s; excluded 0 s	38.699908 Evaluation duration Included 604.004809855832 s; excluded 0 s	16.481819 Evaluation duration Included 118.59403326525016 s; excluded 0 s	0 Evaluation duration Included 16.13639543462473 s; excluded 0 s	-	0 Evaluation duration Included 3.848867791391365 s; excluded 0 s	19.041616 Evaluation duration Included 137.8963054206376 s; excluded 0 s	85.468379 Evaluation duration Included 2083.6452969301185 s; excluded 0 s	0 Evaluation duration Included 160.3285413212152 s; excluded 0 s	36.924994 Evaluation duration Included 413.7901856070912 s; excluded 0 s	0 Evaluation duration Included 21.500649381669504 s; excluded 0 s	80.845935 Evaluation duration Included 636.3085731540472 s; excluded 0 s	44.816335 Evaluation duration Included 4.363636363636363 s; excluded 0 s	0 Evaluation duration Included 20.65687772122959 s; excluded 0 s	25.338741 Evaluation duration Included 355.3293406230003 s; excluded 0 s	2.595289 Evaluation duration Included 52.26089039122396 s; excluded 0 s	0 Evaluation duration Included 2.0168067226890756 s; excluded 0 s	7.980445 Evaluation duration Included 117.54809586145142 s; excluded 0 s	5.80416 Evaluation duration Included 59.20513963649113 s; excluded 0 s
2	ChordMini 2E1D (CL full) Public weights · Inference code checkpoints/2e1d_model_best.pth Inference codeWeight license not declared	GuitarSet 1.1.0 · comping · performed	73.805395 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	79.597236 Evaluation duration Included 3631.09785836346 s; excluded 1813.7882414097799 s	53.334453 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	74.174219 Evaluation duration Included 3502.102646060085 s; excluded 1942.7834537131555 s	47.955555 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	39.150516 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	47.955555 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	39.150516 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	72.442449 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	75.692234 Evaluation duration Included 3631.09785836346 s; excluded 1813.7882414097799 s	52.052753 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	68.453329 Evaluation duration Included 3502.102646060085 s; excluded 1942.7834537131555 s	46.284043 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	37.7932 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	46.284043 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	37.7932 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	78.518746	61.758947	99.853948	178	0 Evaluation duration Included 627.8152749108 s; excluded 0 s	0 Evaluation duration Included 9.636145367237805 s; excluded 0 s	35.613936 Evaluation duration Included 604.004809855832 s; excluded 0 s	12.127208 Evaluation duration Included 118.59403326525016 s; excluded 0 s	0 Evaluation duration Included 16.13639543462473 s; excluded 0 s	-	0 Evaluation duration Included 3.848867791391365 s; excluded 0 s	0 Evaluation duration Included 137.8963054206376 s; excluded 0 s	79.745007 Evaluation duration Included 2083.6452969301185 s; excluded 0 s	0 Evaluation duration Included 160.3285413212152 s; excluded 0 s	35.322424 Evaluation duration Included 413.7901856070912 s; excluded 0 s	0 Evaluation duration Included 21.500649381669504 s; excluded 0 s	75.792358 Evaluation duration Included 636.3085731540472 s; excluded 0 s	40.559335 Evaluation duration Included 4.363636363636363 s; excluded 0 s	0 Evaluation duration Included 20.65687772122959 s; excluded 0 s	21.48945 Evaluation duration Included 355.3293406230003 s; excluded 0 s	0 Evaluation duration Included 52.26089039122396 s; excluded 0 s	0 Evaluation duration Included 2.0168067226890756 s; excluded 0 s	8.533709 Evaluation duration Included 117.54809586145142 s; excluded 0 s	5.804773 Evaluation duration Included 59.20513963649113 s; excluded 0 s
3	Omnizart ChordV1 (raw) Public weights · Inference code chord_v1 complete SavedModel Inference codeWeight license not declared	GuitarSet 1.1.0 · comping · performed	70.681852 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	78.38993 Evaluation duration Included 3631.09785836346 s; excluded 1813.7882414097794 s	52.276855 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	59.698581 Evaluation duration Included 3502.102646060085 s; excluded 1942.7834537131555 s	38.397602 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	30.586609 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	38.397602 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	30.586609 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	69.887902 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	75.061246 Evaluation duration Included 3631.09785836346 s; excluded 1813.7882414097794 s	51.23753 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	52.338084 Evaluation duration Included 3502.102646060085 s; excluded 1942.7834537131555 s	36.493738 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	28.824622 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	36.493738 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	28.824622 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	77.729493	38.652642	100	178	0 Evaluation duration Included 627.8152749108 s; excluded 0 s	0 Evaluation duration Included 9.636145367237805 s; excluded 0 s	0 Evaluation duration Included 604.004809855832 s; excluded 0 s	0 Evaluation duration Included 118.59403326525016 s; excluded 0 s	0 Evaluation duration Included 16.13639543462473 s; excluded 0 s	-	0 Evaluation duration Included 3.848867791391365 s; excluded 0 s	0 Evaluation duration Included 137.8963054206376 s; excluded 0 s	79.618967 Evaluation duration Included 2083.6452969301185 s; excluded 0 s	0 Evaluation duration Included 160.3285413212152 s; excluded 0 s	0 Evaluation duration Included 413.7901856070912 s; excluded 0 s	0 Evaluation duration Included 21.500649381669504 s; excluded 0 s	67.848955 Evaluation duration Included 636.3085731540472 s; excluded 0 s	0 Evaluation duration Included 4.363636363636363 s; excluded 0 s	0 Evaluation duration Included 20.65687772122959 s; excluded 0 s	0 Evaluation duration Included 355.3293406230003 s; excluded 0 s	0 Evaluation duration Included 52.26089039122396 s; excluded 0 s	0 Evaluation duration Included 2.0168067226890756 s; excluded 0 s	0 Evaluation duration Included 117.54809586145142 s; excluded 0 s	0 Evaluation duration Included 59.20513963649113 s; excluded 0 s
4	Omnizart ChordV1 (CSV) Public weights · Inference code chord_v1 complete SavedModel Inference codeWeight license not declared	GuitarSet 1.1.0 · comping · performed	71.032285 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	78.312621 Evaluation duration Included 3631.09785836346 s; excluded 1813.7882414097794 s	52.225299 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	59.555869 Evaluation duration Included 3502.102646060085 s; excluded 1942.7834537131555 s	38.305811 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	30.50625 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	38.305811 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	30.50625 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	70.180979 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	75.228274 Evaluation duration Included 3631.09785836346 s; excluded 1813.7882414097794 s	51.186068 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	52.503452 Evaluation duration Included 3502.102646060085 s; excluded 1942.7834537131555 s	36.40965 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	28.749232 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	36.40965 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	28.749232 Evaluation duration Included 5444.885865759639 s; excluded 0.00023401360525099335 s	77.615253	58.893914	89.774904	178	0 Evaluation duration Included 627.8152749108 s; excluded 0 s	0 Evaluation duration Included 9.636145367237805 s; excluded 0 s	0 Evaluation duration Included 604.004809855832 s; excluded 0 s	0 Evaluation duration Included 118.59403326525016 s; excluded 0 s	0 Evaluation duration Included 16.13639543462473 s; excluded 0 s	-	0 Evaluation duration Included 3.848867791391365 s; excluded 0 s	0 Evaluation duration Included 137.8963054206376 s; excluded 0 s	79.548417 Evaluation duration Included 2083.6452969301185 s; excluded 0 s	0 Evaluation duration Included 160.3285413212152 s; excluded 0 s	0 Evaluation duration Included 413.7901856070912 s; excluded 0 s	0 Evaluation duration Included 21.500649381669504 s; excluded 0 s	67.294524 Evaluation duration Included 636.3085731540472 s; excluded 0 s	0 Evaluation duration Included 4.363636363636363 s; excluded 0 s	0 Evaluation duration Included 20.65687772122959 s; excluded 0 s	0 Evaluation duration Included 355.3293406230003 s; excluded 0 s	0 Evaluation duration Included 52.26089039122396 s; excluded 0 s	0 Evaluation duration Included 2.0168067226890756 s; excluded 0 s	0 Evaluation duration Included 117.54809586145142 s; excluded 0 s	0 Evaluation duration Included 59.20513963649113 s; excluded 0 s

Original records (220)

TelkNet · 2026-09-13 ↗Download complete evaluation dataChecked: 2026-10-25

2602.19778 · S4.T2

Deployable entries 2 / source rows 8

Not ranked: 6

A complete public inference package for these experiments has not been verified. Other weights with the same model name do not qualify.

Bortolozzo et al. † · Prior Work (Noisy Student Framework) · Li et al. † · Prior Work (Noisy Student Framework) · 2E1D (SL) (2.2M) · Traditional supervised learning (SL) baseline (trained on full labels with data augmentation) · BTC (SL) (3.03M) · Traditional supervised learning (SL) baseline (trained on full labels with data augmentation) · 2E1D (CL) (50%) · Results of our student models after Stage 2 with Data-Incremental Continual Learning (with clean labels, α = 0.3 ) · BTC (CL) (50%) · Results of our student models after Stage 2 with Data-Incremental Continual Learning (with clean labels, α = 0.3 )

ChordMini uses a fixed paper split of 420 training, 60 validation and 120 test songs. Only experiments matched to published weights are shown; ± denotes standard deviation. Paper seed averages and this site’s fixed CL full checkpoint are separate.

Rank by metric

2602.19778 · S4.T2
Metric rank	Model	Test setting	Root↑	Thirds↑	Triads↑	7ths↑	Tetrads↑	Majmin↑	MIREX↑	Seg↑	WCSR↑	ACQA↑
1	2E1D (CL) (full) Public weights · Inference code checkpoints/2e1d_model_best.pth Inference codeWeight license not declared	Results of our student models after Stage 2 with Data-Incremental Continual Learning (with clean labels, α = 0.3 )	81.55 ± 0.05	78.52 ± 0.12	76.85 ± 0.13	68.19 ± 0.12	65.49 ± 0.15	78.98 ± 0.11	78.52 ± 0.06	80.73 ± 0.14	71.9 ± 0.2	36.1 ± 0.4
2	BTC (CL) (full) Public weights · Inference code checkpoints/btc_model_best.pth Inference codeWeight license not declared	Results of our student models after Stage 2 with Data-Incremental Continual Learning (with clean labels, α = 0.3 )	83.03 ± 0.04	80.17 ± 0.11	78.33 ± 0.12	69.38 ± 0.10	67.00 ± 0.13	80.24 ± 0.10	80.16 ± 0.05	81.71 ± 0.12	71.6 ± 0.1	39.5 ± 0.4

Original records (2)

2602.19778 · S4.T2 ↗Checked: 2026-10-25

2602.19778 · S3.T1

Deployable entries 1 / source rows 11

Not ranked: 10

A complete public inference package for these experiments has not been verified. Other weights with the same model name do not qualify.

2E1D (2.2M) · Our two student models trained only with pseudo-labels (generated by the BTC teacher from unlabeled data) · FMA (short-form) · 2E1D (2.2M) · Our two student models trained only with pseudo-labels (generated by the BTC teacher from unlabeled data) · DALI (long-form) · 2E1D (2.2M) · Our two student models trained only with pseudo-labels (generated by the BTC teacher from unlabeled data) · MAESTRO+DALI · 2E1D (2.2M) · Our two student models trained only with pseudo-labels (generated by the BTC teacher from unlabeled data) · All (FMA+M+D) · 2E1D (2.2M) · Our two student models trained only with pseudo-labels (generated by the BTC teacher from unlabeled data) · All + KD ( α = 0.3 ) · BTC (3.03M) · Our two student models trained only with pseudo-labels (generated by the BTC teacher from unlabeled data) · FMA (short-form) · BTC (3.03M) · Our two student models trained only with pseudo-labels (generated by the BTC teacher from unlabeled data) · DALI (long-form) · BTC (3.03M) · Our two student models trained only with pseudo-labels (generated by the BTC teacher from unlabeled data) · MAESTRO+DALI · BTC (3.03M) · Our two student models trained only with pseudo-labels (generated by the BTC teacher from unlabeled data) · All (FMA+M+D) · BTC (3.03M) · Our two student models trained only with pseudo-labels (generated by the BTC teacher from unlabeled data) · All + KD ( α = 0.3 )

Only training configurations with verified matching weights and their recognition/segmentation metrics are shown. Root–MIREX and segmentation use human references; Teacher Acc/Prec/Rec/F1 measure teacher agreement, not true chord accuracy.

Rank by metric

2602.19778 · S3.T1
Metric rank	Model	Test setting	mir_eval Metrics							Segmentation			Frame-wise vs. teacher
Metric rank	Model	Test setting	Root↑	Thirds↑	Triads↑	7ths↑	Tetrads↑	Majmin↑	MIREX↑	Over↑	Under↑	Seg↑	Acc↑	Prec↑	Rec↑	F1↑
-	BTC Public weights · Inference code checkpoints/btc_model_large_voca.pt Inference codeWeight license not declared	Pre-trained weights	81.89	78.49	76.85	66.29	63.72	79.00	78.65	82.16	90.33	80.85	-	-	-	-

Original records (1)

2602.19778 · S3.T1 ↗Checked: 2026-10-25

2602.19778 · S5.T3

Deployable entries 2 / source rows 8

Not ranked: 6

A complete public inference package for these experiments has not been verified. Other weights with the same model name do not qualify.

Count · Bortolozzo et al. † · 2E1D (SL) · BTC (SL) · 2E1D (CL) (50%) · BTC (CL) (50%)

All ten chord classes remain, including rare classes and zeros. Source frame counts: N 14.7k, Maj 137k, Min 36.4k, Dom7 32.6k, Maj7 11.4k, Min7 25.3k, Dim 540, Dim7 212, Aug 517, Sus 5.6k.

Rank by metric

2602.19778 · S5.T3
Metric rank	Model	N↑	Maj↑	Min↑	Dom7↑	Maj7↑	Min7↑	Dim↑	Dim7↑	Aug↑	Sus↑
1	BTC (CL) (full) Public weights · Inference code checkpoints/btc_model_best.pth Inference codeWeight license not declared	71.2	87.1	65.6	51.6	42.3	55.7	23.5	45.6	18.5	25.5
2	2E1D (CL) (full) Public weights · Inference code checkpoints/2e1d_model_best.pth Inference codeWeight license not declared	69.2	90.0	68.9	39.4	35.3	51.9	14.8	46.4	8.4	18.8

Original records (2)

2602.19778 · S5.T3 ↗Checked: 2026-10-25

Song BPM Alignment, Stems & DAW Export

Quick feedback

Models, parameters, and sources

Rubber Band R3 + Beat This! + ChordMini BTC + Leap XE 90 + Leap Instrumental

Paper and benchmark notes

Rubber Band Library: high-quality audio time-stretching and pitch-shifting

Beat this! Accurate beat tracking without DBN postprocessing

Omnizart: A General Toolbox for Automatic Music Transcription

Official / repositories

Papers / technical notes

Parameter guide

Workflow

What to check

Models in this field

TelkNet · 2026-09-13

Beat This! paper, Table 1

TelkNet · 2026-09-13

2602.19778 · S4.T2

2602.19778 · S3.T1

2602.19778 · S5.T3

Related use-case guides