File size: 13,521 Bytes
cee3305
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "831245a1",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "2ac8a30f",
   "metadata": {},
   "outputs": [],
   "source": [
    "target_lang=\"ga-IE\"  # change to your target lang"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "id": "15710167",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using custom data configuration ga-pl-lang1=ga,lang2=pl\n",
      "Reusing dataset opus_dgt (/workspace/cache/hf/datasets/opus_dgt/ga-pl-lang1=ga,lang2=pl/0.0.0/a4db75cea3712eb5d4384f0539db82abf897c6b6da5e5e81693e8fd201efc346)\n"
     ]
    }
   ],
   "source": [
    "from datasets import load_dataset\n",
    "\n",
    "# dataset = load_dataset(\"mozilla-foundation/common_voice_8_0\", \n",
    "#                        \"ga-IE\", \n",
    "#                        split=\"train\", \n",
    "#                        use_auth_token = True)\n",
    "\n",
    "dataset = load_dataset(\"opus_dgt\", lang1=\"ga\", lang2=\"pl\", split = 'train')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "id": "fb20d4de",
   "metadata": {},
   "outputs": [],
   "source": [
    "# ga_txt = [i['ga'] for i in dataset['translation']]\n",
    "# ga_txt = pd.Series(ga_txt)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "id": "eeca1851",
   "metadata": {},
   "outputs": [],
   "source": [
    "chars_to_ignore_regex = '[,?.!\\-\\;\\:\"“%‘”�—’…–]'  # change to the ignored characters of your fine-tuned model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "id": "4df93c9c",
   "metadata": {},
   "outputs": [],
   "source": [
    "import re\n",
    "\n",
    "def extract_text(batch):\n",
    "  text = batch[\"translation\"]\n",
    "  ga_text = text['ga']\n",
    "  batch[\"text\"] = re.sub(chars_to_ignore_regex, \"\", ga_text.lower())\n",
    "  return batch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "id": "84bedd13",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d9a11f167bb94faa8e9f6a511407acb4",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "0ex [00:00, ?ex/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "dataset = dataset.map(extract_text, remove_columns=dataset.column_names)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "id": "31cb3c6b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "342d92a5d9c44c59bcb5dca143ced3b6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Pushing dataset shards to the dataset hub:   0%|          | 0/1 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "dataset.push_to_hub(f\"{target_lang}_opus_dgt_train\", split=\"train\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "70952673",
   "metadata": {},
   "source": [
    "## N-gram KenLM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "id": "51756959",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "38d3c229117f4e60a7778f974ac609de",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Downloading:   0%|          | 0.00/1.60k [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using custom data configuration jcmc--ga-IE_opus_dgt_train-aa318da91f5f84f6\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading and preparing dataset opus_dgt/ga-pl (download: 12.11 MiB, generated: 28.99 MiB, post-processed: Unknown size, total: 41.11 MiB) to /workspace/cache/hf/datasets/parquet/jcmc--ga-IE_opus_dgt_train-aa318da91f5f84f6/0.0.0/1638526fd0e8d960534e2155dc54fdff8dce73851f21f031d2fb9c2cf757c121...\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e5e07f18549b443ead74991a9b338593",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/1 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0e83c78fa1bc43f19a56b623c92a64a4",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Downloading:   0%|          | 0.00/12.7M [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "06649f5cd3324eb49a1bd09b68aa23b6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/1 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataset parquet downloaded and prepared to /workspace/cache/hf/datasets/parquet/jcmc--ga-IE_opus_dgt_train-aa318da91f5f84f6/0.0.0/1638526fd0e8d960534e2155dc54fdff8dce73851f21f031d2fb9c2cf757c121. Subsequent calls will reuse this data.\n"
     ]
    }
   ],
   "source": [
    "from datasets import load_dataset\n",
    "\n",
    "dataset = load_dataset(\"jcmc/ga-IE_opus_dgt_train\", split=\"train\")\n",
    "\n",
    "with open(\"text.txt\", \"w\") as file:\n",
    "  file.write(\" \".join(dataset[\"text\"]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "id": "77eb3a41",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "=== 1/5 Counting and sorting n-grams ===\n",
      "Reading /workspace/wav2vec-1b-cv8-ir/text.txt\n",
      "----5---10---15---20---25---30---35---40---45---50---55---60---65---70---75---80---85---90---95--100\n",
      "****************************************************************************************************\n",
      "Unigram tokens 4378228 types 70781\n",
      "=== 2/5 Calculating and sorting adjusted counts ===\n",
      "Chain sizes: 1:849372 2:14475680768 3:27141902336 4:43427041280 5:63331106816\n",
      "Statistics:\n",
      "1 70780 D1=0.684187 D2=1.0538 D3+=1.37643\n",
      "2 652306 D1=0.766205 D2=1.12085 D3+=1.39031\n",
      "3 1669326 D1=0.84217 D2=1.20654 D3+=1.39941\n",
      "4 2514789 D1=0.896214 D2=1.29731 D3+=1.47431\n",
      "5 3053088 D1=0.794858 D2=1.47897 D3+=1.5117\n",
      "Memory estimate for binary LM:\n",
      "type     MB\n",
      "probing 164 assuming -p 1.5\n",
      "probing 192 assuming -r models -p 1.5\n",
      "trie     77 without quantization\n",
      "trie     42 assuming -q 8 -b 8 quantization \n",
      "trie     69 assuming -a 22 array pointer compression\n",
      "trie     34 assuming -a 22 -q 8 -b 8 array pointer compression and quantization\n",
      "=== 3/5 Calculating and sorting initial probabilities ===\n",
      "Chain sizes: 1:849360 2:10436896 3:33386520 4:60354936 5:85486464\n",
      "----5---10---15---20---25---30---35---40---45---50---55---60---65---70---75---80---85---90---95--100\n",
      "####################################################################################################\n",
      "=== 4/5 Calculating and writing order-interpolated probabilities ===\n",
      "Chain sizes: 1:849360 2:10436896 3:33386520 4:60354936 5:85486464\n",
      "----5---10---15---20---25---30---35---40---45---50---55---60---65---70---75---80---85---90---95--100\n",
      "####################################################################################################\n",
      "=== 5/5 Writing ARPA model ===\n",
      "----5---10---15---20---25---30---35---40---45---50---55---60---65---70---75---80---85---90---95--100\n",
      "****************************************************************************************************\n",
      "Name:lmplz\tVmPeak:145097728 kB\tVmRSS:51788 kB\tRSSMax:25679020 kB\tuser:9.15304\tsys:14.1178\tCPU:23.2708\treal:20.9339\n"
     ]
    }
   ],
   "source": [
    "!../kenlm/build/bin/lmplz -o 5 <\"text.txt\" > \"5gram.arpa\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "id": "0e043b87",
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"5gram.arpa\", \"r\") as read_file, open(\"5gram_correct.arpa\", \"w\") as write_file:\n",
    "  has_added_eos = False\n",
    "  for line in read_file:\n",
    "    if not has_added_eos and \"ngram 1=\" in line:\n",
    "      count=line.strip().split(\"=\")[-1]\n",
    "      write_file.write(line.replace(f\"{count}\", f\"{int(count)+1}\"))\n",
    "    elif not has_added_eos and \"<s>\" in line:\n",
    "      write_file.write(line)\n",
    "      write_file.write(line.replace(\"<s>\", \"</s>\"))\n",
    "      has_added_eos = True\n",
    "    else:\n",
    "      write_file.write(line)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "id": "d106c7d1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\\data\\\n",
      "ngram 1=70781\n",
      "ngram 2=652306\n",
      "ngram 3=1669326\n",
      "ngram 4=2514789\n",
      "ngram 5=3053088\n",
      "\n",
      "\\1-grams:\n",
      "-5.8501472\t<unk>\t0\n",
      "0\t<s>\t-0.11565505\n",
      "0\t</s>\t-0.11565505\n",
      "-5.4088216\tmiontuairisc\t-0.20133564\n",
      "-4.6517477\tcheartaitheach\t-0.24842946\n",
      "-2.1893916\tmaidir\t-1.7147961\n",
      "-2.1071756\tle\t-0.7007309\n",
      "-4.156014\tcoinbhinsiún\t-0.31064242\n",
      "-1.8876181\tar\t-0.9045828\n",
      "-4.62287\tdhlínse\t-0.24268326\n",
      "-1.6051095\tagus\t-0.8729715\n",
      "-4.1465816\taithint\t-0.21693327\n"
     ]
    }
   ],
   "source": [
    "!head -20 5gram_correct.arpa"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "id": "85ef4c43",
   "metadata": {},
   "outputs": [],
   "source": [
    "from transformers import AutoProcessor\n",
    "\n",
    "processor = AutoProcessor.from_pretrained(\"./\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "id": "cb2a2768",
   "metadata": {},
   "outputs": [],
   "source": [
    "vocab_dict = processor.tokenizer.get_vocab()\n",
    "sorted_vocab_dict = {k.lower(): v for k, v in sorted(vocab_dict.items(), key=lambda item: item[1])}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "id": "d19eee6f",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Found entries of length > 1 in alphabet. This is unusual unless style is BPE, but the alphabet was not recognized as BPE type. Is this correct?\n",
      "Unigrams and labels don't seem to agree.\n"
     ]
    }
   ],
   "source": [
    "from pyctcdecode import build_ctcdecoder\n",
    "\n",
    "decoder = build_ctcdecoder(\n",
    "    labels=list(sorted_vocab_dict.keys()),\n",
    "    kenlm_model_path=\"5gram_correct.arpa\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "id": "4e8031a9",
   "metadata": {},
   "outputs": [],
   "source": [
    "from transformers import Wav2Vec2ProcessorWithLM\n",
    "\n",
    "processor_with_lm = Wav2Vec2ProcessorWithLM(\n",
    "    feature_extractor=processor.feature_extractor,\n",
    "    tokenizer=processor.tokenizer,\n",
    "    decoder=decoder\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "id": "6f32faf4",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/workspace/wav2vec-1b-cv8-ir/./ is already a clone of https://huggingface.co/jcmc/wav2vec-1b-cv8-ir. Make sure you pull the latest changes with `repo.git_pull()`.\n"
     ]
    }
   ],
   "source": [
    "from huggingface_hub import Repository\n",
    "\n",
    "repo = Repository(local_dir=\"./\", clone_from=\"jcmc/wav2vec-1b-cv8-ir\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "id": "a7e91068",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'/workspace/wav2vec-1b-cv8-ir'"
      ]
     },
     "execution_count": 129,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pwd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0a1de336",
   "metadata": {},
   "outputs": [],
   "source": [
    "processor_with_lm.save_pretrained(\"xls-r-300m-sv\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}