Upload hparams.yaml with huggingface_hub
Browse files- hparams.yaml +310 -0
hparams.yaml
ADDED
|
@@ -0,0 +1,310 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
config: !!python/object/new:yacs.config.CfgNode
|
| 2 |
+
dictitems:
|
| 3 |
+
CONSTANTS_PATH: constants/
|
| 4 |
+
DATA: !!python/object/new:yacs.config.CfgNode
|
| 5 |
+
dictitems:
|
| 6 |
+
MAX_LENGTH: 270
|
| 7 |
+
TEST_PATH: data/kssa_data_1/dev+asr_nobom.txt
|
| 8 |
+
TRAIN_PATH: data/kssa_data_1/train+asr_nobom.txt
|
| 9 |
+
VAL_PATH: data/kssa_data_1/dev+asr_nobom.txt
|
| 10 |
+
state:
|
| 11 |
+
__deprecated_keys__: !!set {}
|
| 12 |
+
__immutable__: true
|
| 13 |
+
__new_allowed__: false
|
| 14 |
+
__renamed_keys__: {}
|
| 15 |
+
INFERENCE: !!python/object/new:yacs.config.CfgNode
|
| 16 |
+
dictitems:
|
| 17 |
+
ASR_MODEL_NAME: sashat/whisper-medium-ClassicalAr
|
| 18 |
+
BATCH_SIZE: 16
|
| 19 |
+
BUFFER_SIZE: 25
|
| 20 |
+
DEVICE: cpu
|
| 21 |
+
FORCED_IDS: null
|
| 22 |
+
MAX_LENGTH: 270
|
| 23 |
+
MODEL_PATH: /home/rufael/Projects/forced_alignment/Diac/outputs/kssa/transformer-text+asr/tashkeela+clartts+kssa/tensorboard/version_13/checkpoints/best_model.ckpt
|
| 24 |
+
OUTPUT_PATH: outputs/kssa/transformer-text+asr/tashkeela+clartts+kssa/results.txt
|
| 25 |
+
USE_ASR: true
|
| 26 |
+
WINDOW_SIZE: 50
|
| 27 |
+
state:
|
| 28 |
+
__deprecated_keys__: !!set {}
|
| 29 |
+
__immutable__: true
|
| 30 |
+
__new_allowed__: false
|
| 31 |
+
__renamed_keys__: {}
|
| 32 |
+
MODEL: !!python/object/new:yacs.config.CfgNode
|
| 33 |
+
dictitems:
|
| 34 |
+
ASR_VOCAB_SIZE: 91
|
| 35 |
+
DFF: 128
|
| 36 |
+
DROPOUT_RATE: 0.2
|
| 37 |
+
D_MODEL: 256
|
| 38 |
+
LOAD_TEXT_BRANCH_ONLY: false
|
| 39 |
+
MAXLEN: 272
|
| 40 |
+
NUM_BLOCKS: 2
|
| 41 |
+
NUM_HEADS: 4
|
| 42 |
+
OUTPUT_SIZE: 19
|
| 43 |
+
PRETRAINED_PATH: /home/rufael/Projects/forced_alignment/Diac/outputs/old/results/transformer-text+asr/tashkeela+clartts/tensorboard/version_0/checkpoints/best_model.ckpt
|
| 44 |
+
TYPE: Transformer
|
| 45 |
+
USE_ASR: true
|
| 46 |
+
VOCAB_SIZE: 77
|
| 47 |
+
WITH_CONN: false
|
| 48 |
+
state:
|
| 49 |
+
__deprecated_keys__: !!set {}
|
| 50 |
+
__immutable__: true
|
| 51 |
+
__new_allowed__: false
|
| 52 |
+
__renamed_keys__: {}
|
| 53 |
+
TRAIN: !!python/object/new:yacs.config.CfgNode
|
| 54 |
+
dictitems:
|
| 55 |
+
ACCUMULATE_GRAD_BATCHES: 1
|
| 56 |
+
BATCH_SIZE: 128
|
| 57 |
+
DEVICE: cuda
|
| 58 |
+
EARLY_STOPPING_PATIENCE: 50
|
| 59 |
+
EVAL_FREQ: 1
|
| 60 |
+
GRAD_CLIP_NORM: 0.5
|
| 61 |
+
LEARNING_RATE: 0.001
|
| 62 |
+
LR_SCHEDULER_FACTOR: 0.5
|
| 63 |
+
LR_SCHEDULER_PATIENCE: 5
|
| 64 |
+
NUM_EPOCHS: 1000
|
| 65 |
+
SAVE_DIR: outputs/kssa/transformer-text+asr/tashkeela+clartts+kssa
|
| 66 |
+
SAVE_FREQ: 30
|
| 67 |
+
VAL_CHECK_INTERVAL: 1.0
|
| 68 |
+
state:
|
| 69 |
+
__deprecated_keys__: !!set {}
|
| 70 |
+
__immutable__: true
|
| 71 |
+
__new_allowed__: false
|
| 72 |
+
__renamed_keys__: {}
|
| 73 |
+
state:
|
| 74 |
+
__deprecated_keys__: !!set {}
|
| 75 |
+
__immutable__: true
|
| 76 |
+
__new_allowed__: false
|
| 77 |
+
__renamed_keys__: {}
|
| 78 |
+
tokenizer: !!python/object:diac.tokenizer.ArabicDiacritizationTokenizer
|
| 79 |
+
constants: !!python/object:diac.tokenizer.TokenizerConstants
|
| 80 |
+
arabic_letters_list: "\u0621\u0622\u0623\u0624\u0625\u0626\u0627\u0628\u0629\u062A\
|
| 81 |
+
\u062B\u062C\u062D\u062E\u062F\u0630\u0631\u0632\u0633\u0634\u0635\u0636\u0637\
|
| 82 |
+
\u0638\u0639\u063A\u0641\u0642\u0643\u0644\u0645\u0646\u0647\u0648\u0649\u064A"
|
| 83 |
+
characters_mapping:
|
| 84 |
+
? '
|
| 85 |
+
|
| 86 |
+
'
|
| 87 |
+
: 4
|
| 88 |
+
' ': 5
|
| 89 |
+
'!': 6
|
| 90 |
+
'"': 7
|
| 91 |
+
'''': 8
|
| 92 |
+
(: 9
|
| 93 |
+
): 10
|
| 94 |
+
'*': 11
|
| 95 |
+
',': 12
|
| 96 |
+
'-': 13
|
| 97 |
+
.: 14
|
| 98 |
+
/: 15
|
| 99 |
+
'0': 16
|
| 100 |
+
'1': 17
|
| 101 |
+
'2': 18
|
| 102 |
+
'3': 19
|
| 103 |
+
'4': 20
|
| 104 |
+
'5': 21
|
| 105 |
+
'6': 22
|
| 106 |
+
'7': 23
|
| 107 |
+
'8': 24
|
| 108 |
+
'9': 25
|
| 109 |
+
':': 26
|
| 110 |
+
;: 27
|
| 111 |
+
<EOS>: 2
|
| 112 |
+
<PAD>: 0
|
| 113 |
+
<SOS>: 1
|
| 114 |
+
<UNK>: 3
|
| 115 |
+
'[': 28
|
| 116 |
+
']': 29
|
| 117 |
+
'`': 30
|
| 118 |
+
'{': 31
|
| 119 |
+
'}': 32
|
| 120 |
+
'~': 33
|
| 121 |
+
"\xAB": 34
|
| 122 |
+
"\xBB": 35
|
| 123 |
+
"\u060C": 36
|
| 124 |
+
"\u061B": 37
|
| 125 |
+
"\u061F": 38
|
| 126 |
+
"\u0621": 39
|
| 127 |
+
"\u0622": 40
|
| 128 |
+
"\u0623": 41
|
| 129 |
+
"\u0624": 42
|
| 130 |
+
"\u0625": 43
|
| 131 |
+
"\u0626": 44
|
| 132 |
+
"\u0627": 45
|
| 133 |
+
"\u0628": 46
|
| 134 |
+
"\u0629": 47
|
| 135 |
+
"\u062A": 48
|
| 136 |
+
"\u062B": 49
|
| 137 |
+
"\u062C": 50
|
| 138 |
+
"\u062D": 51
|
| 139 |
+
"\u062E": 52
|
| 140 |
+
"\u062F": 53
|
| 141 |
+
"\u0630": 54
|
| 142 |
+
"\u0631": 55
|
| 143 |
+
"\u0632": 56
|
| 144 |
+
"\u0633": 57
|
| 145 |
+
"\u0634": 58
|
| 146 |
+
"\u0635": 59
|
| 147 |
+
"\u0636": 60
|
| 148 |
+
"\u0637": 61
|
| 149 |
+
"\u0638": 62
|
| 150 |
+
"\u0639": 63
|
| 151 |
+
"\u063A": 64
|
| 152 |
+
"\u0641": 65
|
| 153 |
+
"\u0642": 66
|
| 154 |
+
"\u0643": 67
|
| 155 |
+
"\u0644": 68
|
| 156 |
+
"\u0645": 69
|
| 157 |
+
"\u0646": 70
|
| 158 |
+
"\u0647": 71
|
| 159 |
+
"\u0648": 72
|
| 160 |
+
"\u0649": 73
|
| 161 |
+
"\u064A": 74
|
| 162 |
+
"\u200F": 75
|
| 163 |
+
"\u2013": 76
|
| 164 |
+
classes_mapping:
|
| 165 |
+
? ''
|
| 166 |
+
: 0
|
| 167 |
+
<EOS>: 17
|
| 168 |
+
<N/A>: 18
|
| 169 |
+
<PAD>: 15
|
| 170 |
+
<SOS>: 16
|
| 171 |
+
"\u064B": 2
|
| 172 |
+
"\u064C": 4
|
| 173 |
+
"\u064D": 6
|
| 174 |
+
"\u064E": 1
|
| 175 |
+
"\u064F": 3
|
| 176 |
+
"\u0650": 5
|
| 177 |
+
"\u0651": 8
|
| 178 |
+
"\u0651\u064B": 10
|
| 179 |
+
"\u0651\u064C": 12
|
| 180 |
+
"\u0651\u064D": 14
|
| 181 |
+
"\u0651\u064E": 9
|
| 182 |
+
"\u0651\u064F": 11
|
| 183 |
+
"\u0651\u0650": 13
|
| 184 |
+
"\u0652": 7
|
| 185 |
+
diacritics_list:
|
| 186 |
+
- "\u064E"
|
| 187 |
+
- "\u064B"
|
| 188 |
+
- "\u0650"
|
| 189 |
+
- "\u064D"
|
| 190 |
+
- "\u064F"
|
| 191 |
+
- "\u064C"
|
| 192 |
+
- "\u0652"
|
| 193 |
+
- "\u0651"
|
| 194 |
+
expanded_vocabulary:
|
| 195 |
+
? '
|
| 196 |
+
|
| 197 |
+
'
|
| 198 |
+
: 4
|
| 199 |
+
' ': 5
|
| 200 |
+
'!': 6
|
| 201 |
+
'"': 7
|
| 202 |
+
'''': 8
|
| 203 |
+
(: 9
|
| 204 |
+
): 10
|
| 205 |
+
'*': 11
|
| 206 |
+
',': 12
|
| 207 |
+
'-': 13
|
| 208 |
+
.: 14
|
| 209 |
+
/: 15
|
| 210 |
+
'0': 16
|
| 211 |
+
'1': 17
|
| 212 |
+
'2': 18
|
| 213 |
+
'3': 19
|
| 214 |
+
'4': 20
|
| 215 |
+
'5': 21
|
| 216 |
+
'6': 22
|
| 217 |
+
'7': 23
|
| 218 |
+
'8': 24
|
| 219 |
+
'9': 25
|
| 220 |
+
':': 26
|
| 221 |
+
;: 27
|
| 222 |
+
<EOS>: 2
|
| 223 |
+
<PAD>: 0
|
| 224 |
+
<SOS>: 1
|
| 225 |
+
<UNK>: 3
|
| 226 |
+
'[': 28
|
| 227 |
+
']': 29
|
| 228 |
+
'`': 30
|
| 229 |
+
'{': 31
|
| 230 |
+
'}': 32
|
| 231 |
+
'~': 33
|
| 232 |
+
"\xAB": 34
|
| 233 |
+
"\xBB": 35
|
| 234 |
+
"\u060C": 36
|
| 235 |
+
"\u061B": 37
|
| 236 |
+
"\u061F": 38
|
| 237 |
+
"\u0621": 39
|
| 238 |
+
"\u0622": 40
|
| 239 |
+
"\u0623": 41
|
| 240 |
+
"\u0624": 42
|
| 241 |
+
"\u0625": 43
|
| 242 |
+
"\u0626": 44
|
| 243 |
+
"\u0627": 45
|
| 244 |
+
"\u0628": 46
|
| 245 |
+
"\u0629": 47
|
| 246 |
+
"\u062A": 48
|
| 247 |
+
"\u062B": 49
|
| 248 |
+
"\u062C": 50
|
| 249 |
+
"\u062D": 51
|
| 250 |
+
"\u062E": 52
|
| 251 |
+
"\u062F": 53
|
| 252 |
+
"\u0630": 54
|
| 253 |
+
"\u0631": 55
|
| 254 |
+
"\u0632": 56
|
| 255 |
+
"\u0633": 57
|
| 256 |
+
"\u0634": 58
|
| 257 |
+
"\u0635": 59
|
| 258 |
+
"\u0636": 60
|
| 259 |
+
"\u0637": 61
|
| 260 |
+
"\u0638": 62
|
| 261 |
+
"\u0639": 63
|
| 262 |
+
"\u063A": 64
|
| 263 |
+
"\u0641": 65
|
| 264 |
+
"\u0642": 66
|
| 265 |
+
"\u0643": 67
|
| 266 |
+
"\u0644": 68
|
| 267 |
+
"\u0645": 69
|
| 268 |
+
"\u0646": 70
|
| 269 |
+
"\u0647": 71
|
| 270 |
+
"\u0648": 72
|
| 271 |
+
"\u0649": 73
|
| 272 |
+
"\u064A": 74
|
| 273 |
+
"\u064B": 78
|
| 274 |
+
"\u064C": 80
|
| 275 |
+
"\u064D": 82
|
| 276 |
+
"\u064E": 77
|
| 277 |
+
"\u064F": 79
|
| 278 |
+
"\u0650": 81
|
| 279 |
+
"\u0651": 84
|
| 280 |
+
"\u0651\u064B": 86
|
| 281 |
+
"\u0651\u064C": 88
|
| 282 |
+
"\u0651\u064D": 90
|
| 283 |
+
"\u0651\u064E": 85
|
| 284 |
+
"\u0651\u064F": 87
|
| 285 |
+
"\u0651\u0650": 89
|
| 286 |
+
"\u0652": 83
|
| 287 |
+
"\u200F": 75
|
| 288 |
+
"\u2013": 76
|
| 289 |
+
rev_classes_mapping:
|
| 290 |
+
0: ''
|
| 291 |
+
1: "\u064E"
|
| 292 |
+
2: "\u064B"
|
| 293 |
+
3: "\u064F"
|
| 294 |
+
4: "\u064C"
|
| 295 |
+
5: "\u0650"
|
| 296 |
+
6: "\u064D"
|
| 297 |
+
7: "\u0652"
|
| 298 |
+
8: "\u0651"
|
| 299 |
+
9: "\u0651\u064E"
|
| 300 |
+
10: "\u0651\u064B"
|
| 301 |
+
11: "\u0651\u064F"
|
| 302 |
+
12: "\u0651\u064C"
|
| 303 |
+
13: "\u0651\u0650"
|
| 304 |
+
14: "\u0651\u064D"
|
| 305 |
+
15: <PAD>
|
| 306 |
+
16: <SOS>
|
| 307 |
+
17: <EOS>
|
| 308 |
+
18: <N/A>
|
| 309 |
+
constants_path: constants/
|
| 310 |
+
with_extra_train: false
|