File size: 1,730 Bytes
a4005ad
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f0fb314
a4005ad
f0fb314
a4005ad
f0fb314
 
 
 
 
 
 
a4005ad
f0fb314
 
 
 
 
 
a4005ad
f0fb314
 
 
a4005ad
 
f0fb314
a4005ad
f0fb314
a4005ad
 
 
f0fb314
 
a4005ad
f0fb314
a4005ad
 
f0fb314
 
 
a4005ad
f0fb314
 
 
a4005ad
f0fb314
 
 
a4005ad
 
f0fb314
a4005ad
f0fb314
 
a4005ad
f0fb314
 
a4005ad
f0fb314
 
1673dfc
a4005ad
f0fb314
1673dfc
f0fb314
1673dfc
f0fb314
 
 
a4005ad
 
 
f0fb314
 
 
a4005ad
 
f0fb314
 
 
 
 
 
f0a9fb7
a4005ad
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
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
dataset:
  training: [
  ]
  validation: [
  ]
  noise: [
  ]
  
  speaker_name_getter: "lambda p: f'{p.parts[-3]}_{p.parts[-2]}'"
  
  use_hdf5: True
  use_metadata: True
  hdf5_flag: r
  validate: True

  workers: 4
  cache: True

  phones_range: [4, 256]
  duration_range: [1.0, 16.0]

  random_utterance: 1.0
  max_prompts: 3
  prompt_duration: 3.0

  sample_type: speaker

  tasks_list: ["tts"] # , "ns", "sr", "tse", "cse", "nse", "tts"]

models:
  _prom_levels: 4
  _max_levels: 8

  _models:
  - name: "ar"
    size: "full"
    resp_levels: 1
    prom_levels: 2
    tasks: 8
    arch_type: "retnet"
    training: True
  - name: "nar"
    size: "full"
    resp_levels: 3
    prom_levels: 4
    tasks: 8
    arch_type: "retnet"
    training: True


hyperparameters:
  batch_size: 8
  gradient_accumulation_steps: 1
  gradient_clipping: 100
  
  optimizer: AdamW
  learning_rate: 1.0e-5
  
  scheduler_type: ""

evaluation:
  batch_size: 16
  frequency: 500
  size: 16
  
  steps: 300
  ar_temperature: 0.95
  nar_temperature: 0.25
  load_disabled_engines: True

trainer:
  iterations: 1_000_000
  
  save_tag: step
  save_on_oom: True
  save_on_quit: True
  save_frequency: 500
  export_on_save: True

  keep_last_checkpoints: 4

  aggressive_optimizations: False
  load_disabled_engines: False

  load_state_dict: True
  
  gc_mode: None # "global_step"

  weight_dtype: float32
  amp: False

  backend: local
  deepspeed:
    zero_optimization_level: 0
    use_compression_training: True

inference:
  weight_dtype: float32
  amp: False

  use_vocos: True
  normalize: False

  recurrent_chunk_size: 0
  recurrent_forward: False

bitsandbytes:
  enabled: False
  injects: True
  linear: True
  embedding: True

device: cpu