File size: 7,518 Bytes
4281bc1 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 |
{
"cells": [
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"!pip install sounddevice scipy torch transformers lang_trans nltk tqdm pyquran"
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"from os import path\n",
"import sounddevice as sd\n",
"import scipy.io.wavfile as wav\n",
"import torch\n",
"from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor\n",
"from lang_trans.arabic import buckwalter\n",
"from nltk import edit_distance\n",
"from tqdm import tqdm\n",
"import pyquran as q"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [],
"source": [
"def record():\n",
" fs = 16000 # Sample rate\n",
" seconds = 5 # Duration of recording\n",
" print(\"Recording...\")\n",
" myrecording = sd.rec(int(seconds * fs), samplerate=fs, channels=1)\n",
" sd.wait() # Wait until recording is finished\n",
" print(\"Finished recording.\")\n",
" return fs , myrecording[:,0]"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [],
"source": [
"def load_Quran_fine_tuned_elgeish_xlsr_53_model_and_processor():\n",
" global loaded_model, loaded_processor\n",
" loaded_model = Wav2Vec2ForCTC.from_pretrained(\"Nuwaisir/Quran_speech_recognizer\").eval()\n",
" loaded_processor = Wav2Vec2Processor.from_pretrained(\"Nuwaisir/Quran_speech_recognizer\")"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [],
"source": [
"def load_elgeish_xlsr_53_model_and_processor():\n",
" global loaded_model, loaded_processor\n",
" loaded_model = Wav2Vec2ForCTC.from_pretrained(\"elgeish/wav2vec2-large-xlsr-53-arabic\").eval()\n",
" loaded_processor = Wav2Vec2Processor.from_pretrained(\"elgeish/wav2vec2-large-xlsr-53-arabic\")"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [],
"source": [
"def predict(single):\n",
" inputs = loaded_processor(single[\"speech\"], sampling_rate=16000, return_tensors=\"pt\", padding=True)\n",
" with torch.no_grad():\n",
" predicted = torch.argmax(loaded_model(inputs.input_values).logits, dim=-1)\n",
" predicted[predicted == -100] = loaded_processor.tokenizer.pad_token_id # see fine-tuning script\n",
" pred_1 = loaded_processor.tokenizer.batch_decode(predicted)[0]\n",
" single[\"predicted\"] = buckwalter.untrans(pred_1)\n",
" return single"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [],
"source": [
"def last_para_str(taskeel=False):\n",
" quran_string = ''\n",
" for i in range (78, 115):\n",
" quran_string += ' '.join(q.quran.get_sura(i, with_tashkeel=taskeel,basmalah=False))\n",
" quran_string += ' '\n",
" return quran_string\n",
"\n",
"def find_match_2(q_str, s, spaces, threshhold = 10):\n",
" len_q = len(q_str)\n",
" len_s = len(s)\n",
" min_dist = 1000000000\n",
" min_dist_pos = []\n",
" for i in tqdm(spaces):\n",
" j = i+1\n",
" k = j + len_s + len_s // 3\n",
" if k > len_q:\n",
" break\n",
" dist = edit_distance(q_str[j:k],s)\n",
" if dist < min_dist:\n",
" min_dist = dist\n",
" min_dist_pos = [j]\n",
" elif dist == min_dist:\n",
" min_dist_pos.append(j)\n",
" return min_dist, min_dist_pos\n",
"\n",
"def find_all_index(s, ch):\n",
" return [i for i, ltr in enumerate(s) if ltr == ch]"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [],
"source": [
"last_para = last_para_str(taskeel=True)\n",
"last_para_spaces = find_all_index(last_para,' ')\n",
"last_para_spaces.insert(0, -1)"
]
},
{
"cell_type": "code",
"execution_count": 13,
"metadata": {},
"outputs": [],
"source": [
"def pipeline():\n",
" fs, myrecording = record()\n",
" single_example = {\n",
" \"speech\": myrecording,\n",
" \"sampling_rate\": fs,\n",
" }\n",
" predicted = predict(single_example)\n",
" print(predicted[\"predicted\"])\n",
" dist,poses = find_match_2(last_para, predicted['predicted'], spaces=last_para_spaces)\n",
" print(\"distance:\",dist)\n",
" print(\"number of matches:\", len(poses))\n",
" for i in poses:\n",
" print(last_para[i:i+200],'\\n')\n"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Load the elgeish_xlsr_53 model"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [],
"source": [
"# load_elgeish_xlsr_53_model_and_processor()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Load Quran fine-tuned elgeish_xlsr_53 model"
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {},
"outputs": [],
"source": [
"load_Quran_fine_tuned_elgeish_xlsr_53_model_and_processor()"
]
},
{
"cell_type": "code",
"execution_count": 14,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Recording...\n",
"Finished recording.\n",
"ููุฅูููุง ูู ููุฑูุงูุดู ุฅูููุง ููููู\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"100%|โโโโโโโโโโ| 2304/2309 [00:03<00:00, 587.76it/s]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"distance: 23\n",
"number of matches: 1\n",
"ููุฅูููููู ููุฑูููุดู ุฅูููููููู
ู ุฑูุญูููุฉู ุงูุดููุชูุงุกู ููุงูุตูููููู ููููููุนูุจูุฏููุง ุฑูุจูู ููุฐูุง ุงููุจูููุชู ุงูููุฐูู ุฃูุทูุนูู
ูููู
ู
ููู ุฌููุนู ููุกูุงู
ูููููู
ู
ูููู ุฎููููู ุฃูุฑูุกูููุชู ุงูููุฐูู ููููุฐููุจู ุจูุงูุฏููููู ููุฐู \n",
"\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"# Recite after running this cell. The first 5 seconds will capture your audio\n",
"pipeline()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": []
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "markdown",
"metadata": {},
"source": []
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"interpreter": {
"hash": "35541def04ad193058c9b5b3afd24560c7277f209ee76d36789dee7d6c5bcde6"
},
"kernelspec": {
"display_name": "Python 3.10.2 64-bit",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.8.10"
},
"orig_nbformat": 4
},
"nbformat": 4,
"nbformat_minor": 2
}
|