Upload 5 files
Browse files- .gitattributes +2 -0
- asr/config.yml +29 -0
- asr/epoch_00080.pth +3 -0
- f0/bst.t7 +3 -0
- plbert/config.yml +30 -0
- plbert/step_1000000.t7 +3 -0
.gitattributes
CHANGED
@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
36 |
+
f0/bst.t7 filter=lfs diff=lfs merge=lfs -text
|
37 |
+
plbert/step_1000000.t7 filter=lfs diff=lfs merge=lfs -text
|
asr/config.yml
ADDED
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
log_dir: "logs/20201006"
|
2 |
+
save_freq: 5
|
3 |
+
device: "cuda"
|
4 |
+
epochs: 180
|
5 |
+
batch_size: 64
|
6 |
+
pretrained_model: ""
|
7 |
+
train_data: "ASRDataset/train_list.txt"
|
8 |
+
val_data: "ASRDataset/val_list.txt"
|
9 |
+
|
10 |
+
dataset_params:
|
11 |
+
data_augmentation: false
|
12 |
+
|
13 |
+
preprocess_parasm:
|
14 |
+
sr: 24000
|
15 |
+
spect_params:
|
16 |
+
n_fft: 2048
|
17 |
+
win_length: 1200
|
18 |
+
hop_length: 300
|
19 |
+
mel_params:
|
20 |
+
n_mels: 80
|
21 |
+
|
22 |
+
model_params:
|
23 |
+
input_dim: 80
|
24 |
+
hidden_dim: 256
|
25 |
+
n_token: 178
|
26 |
+
token_embedding_dim: 512
|
27 |
+
|
28 |
+
optimizer_params:
|
29 |
+
lr: 0.0005
|
asr/epoch_00080.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:fedd55a1234b0c56e1e8b509c74edf3a5e2f27106a66038a4a946047a775bd6c
|
3 |
+
size 94552811
|
f0/bst.t7
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:54dc94364b97e18ac1dfa6287714ed121248cfaac4cfd39d061c6e0a089ef169
|
3 |
+
size 21029926
|
plbert/config.yml
ADDED
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
log_dir: "Checkpoint"
|
2 |
+
mixed_precision: "fp16"
|
3 |
+
data_folder: "wikipedia_20220301.en.processed"
|
4 |
+
batch_size: 192
|
5 |
+
save_interval: 5000
|
6 |
+
log_interval: 10
|
7 |
+
num_process: 1 # number of GPUs
|
8 |
+
num_steps: 1000000
|
9 |
+
|
10 |
+
dataset_params:
|
11 |
+
tokenizer: "transfo-xl-wt103"
|
12 |
+
token_separator: " " # token used for phoneme separator (space)
|
13 |
+
token_mask: "M" # token used for phoneme mask (M)
|
14 |
+
word_separator: 3039 # token used for word separator (<formula>)
|
15 |
+
token_maps: "token_maps.pkl" # token map path
|
16 |
+
|
17 |
+
max_mel_length: 512 # max phoneme length
|
18 |
+
|
19 |
+
word_mask_prob: 0.15 # probability to mask the entire word
|
20 |
+
phoneme_mask_prob: 0.1 # probability to mask each phoneme
|
21 |
+
replace_prob: 0.2 # probablity to replace phonemes
|
22 |
+
|
23 |
+
model_params:
|
24 |
+
vocab_size: 178
|
25 |
+
hidden_size: 768
|
26 |
+
num_attention_heads: 12
|
27 |
+
intermediate_size: 2048
|
28 |
+
max_position_embeddings: 512
|
29 |
+
num_hidden_layers: 12
|
30 |
+
dropout: 0.1
|
plbert/step_1000000.t7
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:0714ff85804db43e06b3b0ac5749bf90cf206257c6c5916e8a98c5933b4c21e0
|
3 |
+
size 25185187
|