voice2motion/configs/example.json at main · ksadov/voice2motion · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
{
  "log_dir": "runs/fp16_0.2_constant",
  "log_tb_every": 500,
  "data_config": {
    "train": {
      "data_dir": "/mnt/storage_4tb/train_preprocessed/",
      "random_start_idx": true,
      "dataloader": {
        "batch_size": 64,
        "shuffle": true,
        "num_workers": 4,
        "pin_memory": true
      }
    },
    "val": {
      "data_dir": "data/TalkingHead-1KH/val/cropped_clips_info",
      "dataloader": {
        "shuffle": false,
        "num_workers": 0
      }
    },
    "shapekey_smoothing": 0,
    "head_angle_smoothing": 0
  },
  "model_config": {
    "type": "Simple",
    "decoder_dimensions": {
      "d_model": 256,
      "nhead": 4,
      "dim_feedforward": 1024,
      "dropout": 0.1,
      "activation": "gelu",
      "num_layers": 4,
      "n_ctx": 90
    },
    "hubert_params": {
      "model_name": "facebook/hubert-base-ls960",
      "return_attention_mask": false,
      "feature_dim": 768
    }
  },
  "device": "cuda",
  "lr": 0.0001,
  "warmup_steps": 1000,
  "lr_patience": 500,
  "max_epochs": 100,
  "val_per_epoch": 1,
  "val_steps": null,
  "render_params": {
    "shape_keys_path": "assets/reference_mesh/shape_keys",
    "min_audio_samples_per_step": 48000,
    "max_audio_samples_per_step": 48000,
    "pipeline_crossfade": 5,
    "max_audio_duration": 6.0
  },
  "checkpoint_every": 5000,
  "start_checkpoint": null,
  "weight_decay": 0.0,
  "max_grad_norm": 1.0,
  "mask_audio_padding": false,
  "teacher_forcing": {
    "start_step": 0,
    "end_step": 0,
    "start_prob": 0.2,
    "end_prob": 0.2
  },
  "zero_padding_prob": 0.5,
  "loss_metric": "l1_smooth",
  "save_ema": true,
  "fp16": true,
  "val_params": {
    "num_video_examples": 3,
    "num_pipeline_examples": 2
  }
}