lamooon commited on
Commit
227e689
·
verified ·
1 Parent(s): d8296fe

migrate batch 9/19 from lamooon/vocal-v1

Browse files
train/v1/config.json ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "encoder": "m-a-p/MERT-v1-95M",
3
+ "encoder_hidden": 768,
4
+ "unet_channels": [
5
+ 256,
6
+ 512,
7
+ 512
8
+ ],
9
+ "conformer_layers": 4,
10
+ "conformer_heads": 8,
11
+ "conformer_ff_dim": 2048,
12
+ "conformer_conv_kernel": 31,
13
+ "num_pitches": 88,
14
+ "dropout": 0.1,
15
+ "sample_rate": 24000,
16
+ "hop_length": 480,
17
+ "segment_length": 5.0,
18
+ "epochs": 20,
19
+ "batch_size": 2,
20
+ "accumulate_grad": 8,
21
+ "lr": 0.001,
22
+ "encoder_lr": 1e-05,
23
+ "weight_decay": 0.01,
24
+ "warmup_epochs": 2,
25
+ "patience": 5,
26
+ "fp16": true,
27
+ "num_workers": 4,
28
+ "seed": 42,
29
+ "w_onset": 5.0,
30
+ "w_pitch": 1.0,
31
+ "w_frame": 0.5,
32
+ "focal_alpha": 0.85,
33
+ "focal_gamma": 2.0,
34
+ "onset_threshold": 0.35,
35
+ "frame_threshold": 0.25,
36
+ "onset_tolerance_ms": 50,
37
+ "pitch_tolerance_cents": 50,
38
+ "data_dir": "./data",
39
+ "output_dir": "outputs",
40
+ "checkpoint": null,
41
+ "eval_only": false,
42
+ "per_dataset": false,
43
+ "datasets": null,
44
+ "augment": true,
45
+ "device": null
46
+ }
train/v1/training_log.jsonl ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"epoch": 1, "train_loss": 4.45803740343124, "val_onset_f1": 0.012584734016458906, "val_onset_precision": 0.0063415231909802975, "val_onset_recall": 0.8117647058717417, "val_frame_f1": 0.21186941575526047, "val_frame_precision": 0.11919323964702844, "val_frame_recall": 0.9523454257195259, "val_pitch_accuracy": 0.30429917000810935, "val_loss": 1.3725575032963682, "best_f1": 0.012584734016458906, "elapsed": 186.19270062446594}
2
+ {"epoch": 2, "train_loss": 3.282650820314329, "val_onset_f1": 0.0798215554553342, "val_onset_precision": 0.0425008483202897, "val_onset_recall": 0.6549019607757529, "val_frame_f1": 0.35189895944112315, "val_frame_precision": 0.2181557872130513, "val_frame_recall": 0.9094469831195472, "val_pitch_accuracy": 0.44017532406934606, "val_loss": 0.9902183395018106, "best_f1": 0.0798215554553342, "elapsed": 164.2449541091919}
3
+ {"epoch": 3, "train_loss": 2.776251209917126, "val_onset_f1": 0.0786046867352341, "val_onset_precision": 0.041827649924372726, "val_onset_recall": 0.6509803921483532, "val_frame_f1": 0.4029052116584129, "val_frame_precision": 0.26176914055948464, "val_frame_recall": 0.8742889116843473, "val_pitch_accuracy": 0.5154341135871346, "val_loss": 0.871544301127598, "best_f1": 0.0798215554553342, "elapsed": 147.1007046699524}
4
+ {"epoch": 4, "train_loss": 2.585367386623024, "val_onset_f1": 0.0971846494209733, "val_onset_precision": 0.052625868055498445, "val_onset_recall": 0.6339869280962878, "val_frame_f1": 0.5768562236819812, "val_frame_precision": 0.42855201159821105, "val_frame_recall": 0.882122540333039, "val_pitch_accuracy": 0.5195374428793066, "val_loss": 0.8135813919570186, "best_f1": 0.0971846494209733, "elapsed": 164.15928292274475}
5
+ {"epoch": 5, "train_loss": 2.494547103522987, "val_onset_f1": 0.09593139069258709, "val_onset_precision": 0.051927021483264675, "val_onset_recall": 0.6287581699264214, "val_frame_f1": 0.6028420839522319, "val_frame_precision": 0.4583818958816733, "val_frame_recall": 0.8802573906547791, "val_pitch_accuracy": 0.5320339457236481, "val_loss": 0.8066509140154986, "best_f1": 0.0971846494209733, "elapsed": 157.57198643684387}
6
+ {"epoch": 6, "train_loss": 2.398980882636218, "val_onset_f1": 0.10751309030124308, "val_onset_precision": 0.0585428305399691, "val_onset_recall": 0.657516339860686, "val_frame_f1": 0.6766227957465304, "val_frame_precision": 0.5604752281214033, "val_frame_recall": 0.853492492771749, "val_pitch_accuracy": 0.5495663526992917, "val_loss": 0.8013632660453667, "best_f1": 0.10751309030124308, "elapsed": 156.2379550933838}
7
+ {"epoch": 7, "train_loss": 2.328131939390296, "val_onset_f1": 0.11431423763364566, "val_onset_precision": 0.06318984547452651, "val_onset_recall": 0.5986928104496902, "val_frame_f1": 0.7023754740856444, "val_frame_precision": 0.5960850621053546, "val_frame_recall": 0.8547980975465309, "val_pitch_accuracy": 0.5595449034779822, "val_loss": 0.7300743367870464, "best_f1": 0.11431423763364566, "elapsed": 151.6288125514984}
8
+ {"epoch": 8, "train_loss": 2.2741030974372585, "val_onset_f1": 0.09709897488879997, "val_onset_precision": 0.05193975353715021, "val_onset_recall": 0.7437908496634799, "val_frame_f1": 0.6703335651927788, "val_frame_precision": 0.5410437073950043, "val_frame_recall": 0.8808169355582571, "val_pitch_accuracy": 0.5640212627058061, "val_loss": 0.757631926039762, "best_f1": 0.11431423763364566, "elapsed": 156.34912705421448}
9
+ {"epoch": 9, "train_loss": 2.2553955689003238, "val_onset_f1": 0.105232329738429, "val_onset_precision": 0.05686253824237064, "val_onset_recall": 0.7045751633894827, "val_frame_f1": 0.6973197557360968, "val_frame_precision": 0.5897717012765447, "val_frame_recall": 0.852839690384358, "val_pitch_accuracy": 0.5739065560005838, "val_loss": 0.7499113113362231, "best_f1": 0.11431423763364566, "elapsed": 147.73661541938782}
10
+ {"epoch": 10, "train_loss": 2.2617931738218338, "val_onset_f1": 0.12147134133299695, "val_onset_precision": 0.06699919115655567, "val_onset_recall": 0.6496732026058866, "val_frame_f1": 0.7073492338020017, "val_frame_precision": 0.6188811188806861, "val_frame_recall": 0.8253287326300239, "val_pitch_accuracy": 0.5762379930984087, "val_loss": 0.7551215547206465, "best_f1": 0.12147134133299695, "elapsed": 135.40274000167847}
11
+ {"epoch": 11, "train_loss": 2.2446174643711783, "val_onset_f1": 0.08792471701307533, "val_onset_precision": 0.046518698692576375, "val_onset_recall": 0.7999999999895425, "val_frame_f1": 0.6110762624700324, "val_frame_precision": 0.46033545948478133, "val_frame_recall": 0.9086076657643303, "val_pitch_accuracy": 0.5832323043918836, "val_loss": 0.7592528112225735, "best_f1": 0.12147134133299695, "elapsed": 163.4368360042572}
12
+ {"epoch": 12, "train_loss": 2.2376665972603322, "val_onset_f1": 0.10412021190815555, "val_onset_precision": 0.056230366492087715, "val_onset_recall": 0.7019607843045496, "val_frame_f1": 0.6841251058699788, "val_frame_precision": 0.5658673500515188, "val_frame_recall": 0.8648699058091346, "val_pitch_accuracy": 0.5814604121975366, "val_loss": 0.7396981518076204, "best_f1": 0.12147134133299695, "elapsed": 148.06835341453552}
13
+ {"epoch": 13, "train_loss": 2.2099307683995426, "val_onset_f1": 0.10487388448774461, "val_onset_precision": 0.05653818627946371, "val_onset_recall": 0.7228758169840147, "val_frame_f1": 0.6624010133915832, "val_frame_precision": 0.5333902647306955, "val_frame_recall": 0.8737293667808693, "val_pitch_accuracy": 0.5908794180727494, "val_loss": 0.7377091986405204, "best_f1": 0.12147134133299695, "elapsed": 148.39724278450012}
14
+ {"epoch": 14, "train_loss": 2.1925495263237638, "val_onset_f1": 0.09860769743297522, "val_onset_precision": 0.05259938837915893, "val_onset_recall": 0.7869281045648767, "val_frame_f1": 0.6458911951569469, "val_frame_precision": 0.5071546358846177, "val_frame_recall": 0.8891168516265139, "val_pitch_accuracy": 0.5792222325836247, "val_loss": 0.7403893079851934, "best_f1": 0.12147134133299695, "elapsed": 150.26107144355774}
train/v10/best.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0e908b06e4645f97020f3b1fe866e7e66e0cce21e205a14a37d9927884f24fad
3
+ size 1525491669