ai-voice-cloning/models/.template.valle.yaml

107 lines
1.9 KiB
YAML
Raw Permalink Normal View History

2023-08-23 21:42:32 +00:00
dataset:
training: [
"./training/${voice}/valle/",
]
noise: [
"./training/valle/data/Other/noise/",
]
speaker_name_getter: "lambda p: p.parts[-3]" # "lambda p: f'{p.parts[-3]}_{p.parts[-2]}'"
use_hdf5: False
hdf5_name: data.h5
hdf5_flag: r
validate: True
2023-03-14 05:02:14 +00:00
2023-08-23 21:42:32 +00:00
workers: 4
cache: False
2023-03-22 17:47:23 +00:00
2023-08-23 21:42:32 +00:00
phones_range: [4, 64]
duration_range: [1.0, 8.0]
2023-03-14 15:48:09 +00:00
2023-08-23 21:42:32 +00:00
random_utterance: 1.0
max_prompts: 3
prompt_duration: 3.0
sample_type: path
2023-08-23 21:42:32 +00:00
tasks_list: ["tts"] # ["tts", "ns", "sr", "tse", "cse", "nse", "tts"]
models:
_max_levels: 8
_models:
- name: "ar"
size: "full"
resp_levels: 1
prom_levels: 2
tasks: 8
arch_type: "retnet"
- name: "nar"
size: "full"
resp_levels: 3
prom_levels: 4
tasks: 8
arch_type: "retnet"
hyperparameters:
batch_size: ${batch_size}
gradient_accumulation_steps: ${gradient_accumulation_size}
gradient_clipping: 100
optimizer: AdamW
learning_rate: 1.0e-4
scheduler_type: ""
evaluation:
batch_size: ${batch_size}
frequency: ${validation_rate}
size: 16
steps: 300
ar_temperature: 0.95
nar_temperature: 0.25
trainer:
iterations: ${iterations}
save_tag: step
save_on_oom: True
save_on_quit: True
export_on_save: True
export_on_quit: True
save_frequency: ${save_rate}
keep_last_checkpoints: 4
aggressive_optimizations: False
load_state_dict: True
#strict_loading: False
#load_tag: "9500"
#load_states: False
#restart_step_count: True
gc_mode: None # "global_step"
weight_dtype: bfloat16
backend: deepspeed
deepspeed:
zero_optimization_level: 2
use_compression_training: True
inference:
use_vocos: True
normalize: False
weight_dtype: float32
bitsandbytes:
enabled: False
injects: True
linear: True
embedding: True