rufaelfekadu's picture
Upload hparams.yaml with huggingface_hub
9c444b1 verified
Raw
History Blame
6.94 kB
config: !!python/object/new:yacs.config.CfgNode
dictitems:
CONSTANTS_PATH: constants/
DATA: !!python/object/new:yacs.config.CfgNode
dictitems:
MAX_LENGTH: 270
TEST_PATH: data/kssa_data_1/dev+asr_nobom.txt
TRAIN_PATH: data/kssa_data_1/train+asr_nobom.txt
VAL_PATH: data/kssa_data_1/dev+asr_nobom.txt
state:
__deprecated_keys__: !!set {}
__immutable__: true
__new_allowed__: false
__renamed_keys__: {}
INFERENCE: !!python/object/new:yacs.config.CfgNode
dictitems:
ASR_MODEL_NAME: sashat/whisper-medium-ClassicalAr
BATCH_SIZE: 16
BUFFER_SIZE: 25
DEVICE: cpu
FORCED_IDS: null
MAX_LENGTH: 270
MODEL_PATH: /home/rufael/Projects/forced_alignment/Diac/outputs/kssa/transformer-text+asr/tashkeela+clartts+kssa/tensorboard/version_13/checkpoints/best_model.ckpt
OUTPUT_PATH: outputs/kssa/transformer-text+asr/tashkeela+clartts+kssa/results.txt
USE_ASR: true
WINDOW_SIZE: 50
state:
__deprecated_keys__: !!set {}
__immutable__: true
__new_allowed__: false
__renamed_keys__: {}
MODEL: !!python/object/new:yacs.config.CfgNode
dictitems:
ASR_VOCAB_SIZE: 91
DFF: 128
DROPOUT_RATE: 0.2
D_MODEL: 256
LOAD_TEXT_BRANCH_ONLY: false
MAXLEN: 272
NUM_BLOCKS: 2
NUM_HEADS: 4
OUTPUT_SIZE: 19
PRETRAINED_PATH: /home/rufael/Projects/forced_alignment/Diac/outputs/old/results/transformer-text+asr/tashkeela+clartts/tensorboard/version_0/checkpoints/best_model.ckpt
TYPE: Transformer
USE_ASR: true
VOCAB_SIZE: 77
WITH_CONN: false
state:
__deprecated_keys__: !!set {}
__immutable__: true
__new_allowed__: false
__renamed_keys__: {}
TRAIN: !!python/object/new:yacs.config.CfgNode
dictitems:
ACCUMULATE_GRAD_BATCHES: 1
BATCH_SIZE: 128
DEVICE: cuda
EARLY_STOPPING_PATIENCE: 50
EVAL_FREQ: 1
GRAD_CLIP_NORM: 0.5
LEARNING_RATE: 0.001
LR_SCHEDULER_FACTOR: 0.5
LR_SCHEDULER_PATIENCE: 5
NUM_EPOCHS: 1000
SAVE_DIR: outputs/kssa/transformer-text+asr/tashkeela+clartts+kssa
SAVE_FREQ: 30
VAL_CHECK_INTERVAL: 1.0
state:
__deprecated_keys__: !!set {}
__immutable__: true
__new_allowed__: false
__renamed_keys__: {}
state:
__deprecated_keys__: !!set {}
__immutable__: true
__new_allowed__: false
__renamed_keys__: {}
tokenizer: !!python/object:diac.tokenizer.ArabicDiacritizationTokenizer
constants: !!python/object:diac.tokenizer.TokenizerConstants
arabic_letters_list: "\u0621\u0622\u0623\u0624\u0625\u0626\u0627\u0628\u0629\u062A\
\u062B\u062C\u062D\u062E\u062F\u0630\u0631\u0632\u0633\u0634\u0635\u0636\u0637\
\u0638\u0639\u063A\u0641\u0642\u0643\u0644\u0645\u0646\u0647\u0648\u0649\u064A"
characters_mapping:
? '
'
: 4
' ': 5
'!': 6
'"': 7
'''': 8
(: 9
): 10
'*': 11
',': 12
'-': 13
.: 14
/: 15
'0': 16
'1': 17
'2': 18
'3': 19
'4': 20
'5': 21
'6': 22
'7': 23
'8': 24
'9': 25
':': 26
;: 27
<EOS>: 2
<PAD>: 0
<SOS>: 1
<UNK>: 3
'[': 28
']': 29
'`': 30
'{': 31
'}': 32
'~': 33
"\xAB": 34
"\xBB": 35
"\u060C": 36
"\u061B": 37
"\u061F": 38
"\u0621": 39
"\u0622": 40
"\u0623": 41
"\u0624": 42
"\u0625": 43
"\u0626": 44
"\u0627": 45
"\u0628": 46
"\u0629": 47
"\u062A": 48
"\u062B": 49
"\u062C": 50
"\u062D": 51
"\u062E": 52
"\u062F": 53
"\u0630": 54
"\u0631": 55
"\u0632": 56
"\u0633": 57
"\u0634": 58
"\u0635": 59
"\u0636": 60
"\u0637": 61
"\u0638": 62
"\u0639": 63
"\u063A": 64
"\u0641": 65
"\u0642": 66
"\u0643": 67
"\u0644": 68
"\u0645": 69
"\u0646": 70
"\u0647": 71
"\u0648": 72
"\u0649": 73
"\u064A": 74
"\u200F": 75
"\u2013": 76
classes_mapping:
? ''
: 0
<EOS>: 17
<N/A>: 18
<PAD>: 15
<SOS>: 16
"\u064B": 2
"\u064C": 4
"\u064D": 6
"\u064E": 1
"\u064F": 3
"\u0650": 5
"\u0651": 8
"\u0651\u064B": 10
"\u0651\u064C": 12
"\u0651\u064D": 14
"\u0651\u064E": 9
"\u0651\u064F": 11
"\u0651\u0650": 13
"\u0652": 7
diacritics_list:
- "\u064E"
- "\u064B"
- "\u0650"
- "\u064D"
- "\u064F"
- "\u064C"
- "\u0652"
- "\u0651"
expanded_vocabulary:
? '
'
: 4
' ': 5
'!': 6
'"': 7
'''': 8
(: 9
): 10
'*': 11
',': 12
'-': 13
.: 14
/: 15
'0': 16
'1': 17
'2': 18
'3': 19
'4': 20
'5': 21
'6': 22
'7': 23
'8': 24
'9': 25
':': 26
;: 27
<EOS>: 2
<PAD>: 0
<SOS>: 1
<UNK>: 3
'[': 28
']': 29
'`': 30
'{': 31
'}': 32
'~': 33
"\xAB": 34
"\xBB": 35
"\u060C": 36
"\u061B": 37
"\u061F": 38
"\u0621": 39
"\u0622": 40
"\u0623": 41
"\u0624": 42
"\u0625": 43
"\u0626": 44
"\u0627": 45
"\u0628": 46
"\u0629": 47
"\u062A": 48
"\u062B": 49
"\u062C": 50
"\u062D": 51
"\u062E": 52
"\u062F": 53
"\u0630": 54
"\u0631": 55
"\u0632": 56
"\u0633": 57
"\u0634": 58
"\u0635": 59
"\u0636": 60
"\u0637": 61
"\u0638": 62
"\u0639": 63
"\u063A": 64
"\u0641": 65
"\u0642": 66
"\u0643": 67
"\u0644": 68
"\u0645": 69
"\u0646": 70
"\u0647": 71
"\u0648": 72
"\u0649": 73
"\u064A": 74
"\u064B": 78
"\u064C": 80
"\u064D": 82
"\u064E": 77
"\u064F": 79
"\u0650": 81
"\u0651": 84
"\u0651\u064B": 86
"\u0651\u064C": 88
"\u0651\u064D": 90
"\u0651\u064E": 85
"\u0651\u064F": 87
"\u0651\u0650": 89
"\u0652": 83
"\u200F": 75
"\u2013": 76
rev_classes_mapping:
0: ''
1: "\u064E"
2: "\u064B"
3: "\u064F"
4: "\u064C"
5: "\u0650"
6: "\u064D"
7: "\u0652"
8: "\u0651"
9: "\u0651\u064E"
10: "\u0651\u064B"
11: "\u0651\u064F"
12: "\u0651\u064C"
13: "\u0651\u0650"
14: "\u0651\u064D"
15: <PAD>
16: <SOS>
17: <EOS>
18: <N/A>
constants_path: constants/
with_extra_train: false