chore: add model and gradio app

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+src/model/*.pt filter=lfs diff=lfs merge=lfs -text

app.py

@@ -0,0 +1,21 @@
+import gradio as gr
+from prediction import main
+
+def infer(word1, word2, word3):
+    inputs = [word1, word2, word3]
+    result = main(inputs)
+    return result
+
+demo = gr.Interface(
+    fn=infer,
+    inputs=[
+        gr.Textbox(label="carro"),
+        gr.Textbox(label="asadelta"),
+        gr.Textbox(label="assoviar"),
+    ],
+    outputs=gr.Textbox(label="Generated Word + Definition"),
+    title="Rellow: Imaginary Word Generator",
+    description="Enter three words — the model invents a new word and its definition."
+)
+
+demo.launch()

prediction.py

@@ -0,0 +1,53 @@
+import torch
+import tiktoken
+from constants.tokens import special_tokens
+from services.model import load_model, get_device
+
+# Initialize tokenizer
+_tokenizer = tiktoken.get_encoding("cl100k_base")
+
+# Generate an imaginary word and its definition from three input words.
+def generate_word(words, model, vocab, inv_vocab, max_length=64):
+  device = get_device()
+  
+  # Tokenize input words
+  input_text = ",".join(words)
+  input_tokens = _tokenizer.encode(input_text)
+  input_tensor = torch.tensor([vocab.get(str(tok), vocab["<pad>"]) for tok in input_tokens]).unsqueeze(0).to(device)
+  
+  # Initialize target with SOS token
+  target = torch.tensor([[vocab["<sos>"]]]).to(device)
+  
+  # Generate output
+  with torch.no_grad():
+    for _ in range(max_length):
+      output = model(input_tensor, target)
+      next_token = output[:, -1, :].argmax(dim=-1, keepdim=True)
+      
+      # Stop if we predict EOS token
+      if next_token.item() == vocab["<eos>"]:
+        break
+        
+      target = torch.cat([target, next_token], dim=1)
+  
+  # Convert output tokens to text
+  output_tokens = target[0].cpu().numpy()
+  output_text = _tokenizer.decode([int(inv_vocab[tok]) for tok in output_tokens if tok not in special_tokens.values()])
+  
+  return output_text
+
+def main(words=None):
+  # Load model and vocabulary
+  model, vocab, inv_vocab = load_model()
+  
+  # Use provided words or default example
+  if words is None:
+    words = ["muito", "grande", "imenso"]
+  
+  result = generate_word(words, model, vocab, inv_vocab)
+  print(f"Input words: {', '.join(words)}")
+  print(f"Generated: {result}")
+  return result
+
+if __name__ == "__main__":
+  main()

requirements.txt

@@ -0,0 +1,4 @@
+torch
+gradio
+tiktoken==0.7.0
+numpy

src/constants/tokens.py

@@ -0,0 +1,9 @@
+special_tokens = {
+  '<pad>': 0,
+  '<sos>': 1,
+  '<eos>': 2,
+}
+
+PAD_ID = special_tokens["<pad>"]
+SOS_ID = special_tokens["<sos>"]
+EOS_ID = special_tokens["<eos>"]

src/data/vocab.json

@@ -0,0 +1,1531 @@
+{
+  "11": 3,
+  "12": 4,
+  "13": 5,
+  "64": 6,
+  "65": 7,
+  "66": 8,
+  "67": 9,
+  "68": 10,
+  "69": 11,
+  "70": 12,
+  "71": 13,
+  "72": 14,
+  "73": 15,
+  "75": 16,
+  "76": 17,
+  "77": 18,
+  "78": 19,
+  "79": 20,
+  "81": 21,
+  "82": 22,
+  "83": 23,
+  "84": 24,
+  "85": 25,
+  "87": 26,
+  "89": 27,
+  "258": 28,
+  "259": 29,
+  "263": 30,
+  "264": 31,
+  "265": 32,
+  "266": 33,
+  "267": 34,
+  "268": 35,
+  "269": 36,
+  "272": 37,
+  "273": 38,
+  "274": 39,
+  "275": 40,
+  "276": 41,
+  "277": 42,
+  "278": 43,
+  "281": 44,
+  "282": 45,
+  "285": 46,
+  "288": 47,
+  "291": 48,
+  "292": 49,
+  "293": 50,
+  "294": 51,
+  "295": 52,
+  "296": 53,
+  "297": 54,
+  "299": 55,
+  "300": 56,
+  "301": 57,
+  "303": 58,
+  "304": 59,
+  "305": 60,
+  "306": 61,
+  "308": 62,
+  "309": 63,
+  "311": 64,
+  "312": 65,
+  "316": 66,
+  "318": 67,
+  "320": 68,
+  "321": 69,
+  "323": 70,
+  "324": 71,
+  "325": 72,
+  "326": 73,
+  "327": 74,
+  "329": 75,
+  "331": 76,
+  "332": 77,
+  "336": 78,
+  "337": 79,
+  "342": 80,
+  "343": 81,
+  "344": 82,
+  "346": 83,
+  "347": 84,
+  "348": 85,
+  "349": 86,
+  "351": 87,
+  "354": 88,
+  "355": 89,
+  "359": 90,
+  "360": 91,
+  "361": 92,
+  "369": 93,
+  "370": 94,
+  "372": 95,
+  "373": 96,
+  "374": 97,
+  "375": 98,
+  "376": 99,
+  "380": 100,
+  "383": 101,
+  "384": 102,
+  "385": 103,
+  "387": 104,
+  "388": 105,
+  "390": 106,
+  "391": 107,
+  "395": 108,
+  "396": 109,
+  "402": 110,
+  "403": 111,
+  "404": 112,
+  "408": 113,
+  "409": 114,
+  "414": 115,
+  "417": 116,
+  "424": 117,
+  "428": 118,
+  "436": 119,
+  "437": 120,
+  "438": 121,
+  "439": 122,
+  "444": 123,
+  "447": 124,
+  "450": 125,
+  "451": 126,
+  "453": 127,
+  "454": 128,
+  "459": 129,
+  "460": 130,
+  "461": 131,
+  "462": 132,
+  "463": 133,
+  "466": 134,
+  "470": 135,
+  "472": 136,
+  "478": 137,
+  "483": 138,
+  "485": 139,
+  "490": 140,
+  "496": 141,
+  "501": 142,
+  "503": 143,
+  "506": 144,
+  "511": 145,
+  "513": 146,
+  "514": 147,
+  "519": 148,
+  "520": 149,
+  "521": 150,
+  "523": 151,
+  "525": 152,
+  "531": 153,
+  "537": 154,
+  "539": 155,
+  "540": 156,
+  "541": 157,
+  "546": 158,
+  "548": 159,
+  "550": 160,
+  "554": 161,
+  "566": 162,
+  "569": 163,
+  "575": 164,
+  "579": 165,
+  "580": 166,
+  "581": 167,
+  "582": 168,
+  "588": 169,
+  "592": 170,
+  "593": 171,
+  "594": 172,
+  "603": 173,
+  "613": 174,
+  "614": 175,
+  "628": 176,
+  "630": 177,
+  "635": 178,
+  "638": 179,
+  "648": 180,
+  "649": 181,
+  "651": 182,
+  "653": 183,
+  "655": 184,
+  "656": 185,
+  "664": 186,
+  "665": 187,
+  "668": 188,
+  "671": 189,
+  "676": 190,
+  "677": 191,
+  "687": 192,
+  "688": 193,
+  "689": 194,
+  "708": 195,
+  "710": 196,
+  "712": 197,
+  "713": 198,
+  "716": 199,
+  "729": 200,
+  "737": 201,
+  "738": 202,
+  "743": 203,
+  "749": 204,
+  "751": 205,
+  "754": 206,
+  "757": 207,
+  "762": 208,
+  "778": 209,
+  "779": 210,
+  "782": 211,
+  "784": 212,
+  "797": 213,
+  "799": 214,
+  "802": 215,
+  "805": 216,
+  "809": 217,
+  "818": 218,
+  "822": 219,
+  "830": 220,
+  "831": 221,
+  "834": 222,
+  "840": 223,
+  "846": 224,
+  "857": 225,
+  "864": 226,
+  "865": 227,
+  "867": 228,
+  "880": 229,
+  "884": 230,
+  "893": 231,
+  "896": 232,
+  "899": 233,
+  "911": 234,
+  "912": 235,
+  "924": 236,
+  "934": 237,
+  "936": 238,
+  "938": 239,
+  "940": 240,
+  "944": 241,
+  "945": 242,
+  "951": 243,
+  "953": 244,
+  "958": 245,
+  "959": 246,
+  "967": 247,
+  "969": 248,
+  "971": 249,
+  "978": 250,
+  "981": 251,
+  "983": 252,
+  "991": 253,
+  "992": 254,
+  "998": 255,
+  "1007": 256,
+  "1008": 257,
+  "1022": 258,
+  "1028": 259,
+  "1030": 260,
+  "1031": 261,
+  "1037": 262,
+  "1040": 263,
+  "1048": 264,
+  "1052": 265,
+  "1069": 266,
+  "1072": 267,
+  "1080": 268,
+  "1082": 269,
+  "1083": 270,
+  "1089": 271,
+  "1097": 272,
+  "1098": 273,
+  "1099": 274,
+  "1113": 275,
+  "1116": 276,
+  "1121": 277,
+  "1126": 278,
+  "1138": 279,
+  "1149": 280,
+  "1157": 281,
+  "1166": 282,
+  "1169": 283,
+  "1172": 284,
+  "1189": 285,
+  "1195": 286,
+  "1206": 287,
+  "1213": 288,
+  "1218": 289,
+  "1224": 290,
+  "1237": 291,
+  "1239": 292,
+  "1263": 293,
+  "1279": 294,
+  "1280": 295,
+  "1295": 296,
+  "1296": 297,
+  "1299": 298,
+  "1317": 299,
+  "1321": 300,
+  "1323": 301,
+  "1327": 302,
+  "1331": 303,
+  "1332": 304,
+  "1342": 305,
+  "1347": 306,
+  "1355": 307,
+  "1367": 308,
+  "1370": 309,
+  "1376": 310,
+  "1380": 311,
+  "1391": 312,
+  "1394": 313,
+  "1395": 314,
+  "1400": 315,
+  "1421": 316,
+  "1437": 317,
+  "1446": 318,
+  "1448": 319,
+  "1466": 320,
+  "1469": 321,
+  "1474": 322,
+  "1485": 323,
+  "1499": 324,
+  "1501": 325,
+  "1517": 326,
+  "1522": 327,
+  "1529": 328,
+  "1540": 329,
+  "1543": 330,
+  "1557": 331,
+  "1558": 332,
+  "1560": 333,
+  "1564": 334,
+  "1589": 335,
+  "1592": 336,
+  "1601": 337,
+  "1604": 338,
+  "1610": 339,
+  "1612": 340,
+  "1624": 341,
+  "1645": 342,
+  "1646": 343,
+  "1652": 344,
+  "1658": 345,
+  "1662": 346,
+  "1665": 347,
+  "1678": 348,
+  "1685": 349,
+  "1709": 350,
+  "1744": 351,
+  "1762": 352,
+  "1764": 353,
+  "1768": 354,
+  "1773": 355,
+  "1794": 356,
+  "1826": 357,
+  "1832": 358,
+  "1841": 359,
+  "1860": 360,
+  "1864": 361,
+  "1867": 362,
+  "1871": 363,
+  "1880": 364,
+  "1897": 365,
+  "1922": 366,
+  "1955": 367,
+  "1962": 368,
+  "1964": 369,
+  "1978": 370,
+  "1989": 371,
+  "1995": 372,
+  "2000": 373,
+  "2016": 374,
+  "2041": 375,
+  "2047": 376,
+  "2067": 377,
+  "2068": 378,
+  "2070": 379,
+  "2092": 380,
+  "2099": 381,
+  "2106": 382,
+  "2107": 383,
+  "2117": 384,
+  "2119": 385,
+  "2145": 386,
+  "2159": 387,
+  "2171": 388,
+  "2172": 389,
+  "2177": 390,
+  "2200": 391,
+  "2201": 392,
+  "2215": 393,
+  "2223": 394,
+  "2227": 395,
+  "2252": 396,
+  "2254": 397,
+  "2259": 398,
+  "2269": 399,
+  "2295": 400,
+  "2308": 401,
+  "2322": 402,
+  "2347": 403,
+  "2385": 404,
+  "2389": 405,
+  "2392": 406,
+  "2423": 407,
+  "2442": 408,
+  "2445": 409,
+  "2473": 410,
+  "2480": 411,
+  "2483": 412,
+  "2486": 413,
+  "2497": 414,
+  "2512": 415,
+  "2558": 416,
+  "2579": 417,
+  "2605": 418,
+  "2629": 419,
+  "2642": 420,
+  "2649": 421,
+  "2679": 422,
+  "2700": 423,
+  "2709": 424,
+  "2727": 425,
+  "2739": 426,
+  "2749": 427,
+  "2751": 428,
+  "2762": 429,
+  "2807": 430,
+  "2827": 431,
+  "2836": 432,
+  "2840": 433,
+  "2852": 434,
+  "2853": 435,
+  "2854": 436,
+  "2910": 437,
+  "2917": 438,
+  "2960": 439,
+  "2961": 440,
+  "2967": 441,
+  "2979": 442,
+  "2995": 443,
+  "3026": 444,
+  "3055": 445,
+  "3067": 446,
+  "3074": 447,
+  "3078": 448,
+  "3091": 449,
+  "3105": 450,
+  "3117": 451,
+  "3127": 452,
+  "3197": 453,
+  "3209": 454,
+  "3212": 455,
+  "3213": 456,
+  "3223": 457,
+  "3237": 458,
+  "3276": 459,
+  "3282": 460,
+  "3289": 461,
+  "3315": 462,
+  "3326": 463,
+  "3352": 464,
+  "3375": 465,
+  "3394": 466,
+  "3395": 467,
+  "3415": 468,
+  "3429": 469,
+  "3444": 470,
+  "3496": 471,
+  "3497": 472,
+  "3510": 473,
+  "3512": 474,
+  "3521": 475,
+  "3527": 476,
+  "3528": 477,
+  "3536": 478,
+  "3557": 479,
+  "3567": 480,
+  "3609": 481,
+  "3614": 482,
+  "3635": 483,
+  "3667": 484,
+  "3678": 485,
+  "3687": 486,
+  "3689": 487,
+  "3698": 488,
+  "3703": 489,
+  "3714": 490,
+  "3716": 491,
+  "3769": 492,
+  "3833": 493,
+  "3841": 494,
+  "3846": 495,
+  "3887": 496,
+  "3904": 497,
+  "3913": 498,
+  "3920": 499,
+  "3935": 500,
+  "3969": 501,
+  "3972": 502,
+  "4039": 503,
+  "4042": 504,
+  "4046": 505,
+  "4057": 506,
+  "4087": 507,
+  "4247": 508,
+  "4298": 509,
+  "4306": 510,
+  "4317": 511,
+  "4319": 512,
+  "4338": 513,
+  "4371": 514,
+  "4378": 515,
+  "4381": 516,
+  "4406": 517,
+  "4411": 518,
+  "4415": 519,
+  "4428": 520,
+  "4433": 521,
+  "4457": 522,
+  "4458": 523,
+  "4528": 524,
+  "4543": 525,
+  "4545": 526,
+  "4547": 527,
+  "4581": 528,
+  "4590": 529,
+  "4633": 530,
+  "4634": 531,
+  "4637": 532,
+  "4647": 533,
+  "4669": 534,
+  "4675": 535,
+  "4692": 536,
+  "4704": 537,
+  "4744": 538,
+  "4749": 539,
+  "4766": 540,
+  "4781": 541,
+  "4849": 542,
+  "4902": 543,
+  "4915": 544,
+  "4917": 545,
+  "4919": 546,
+  "4941": 547,
+  "4988": 548,
+  "5048": 549,
+  "5066": 550,
+  "5169": 551,
+  "5171": 552,
+  "5178": 553,
+  "5208": 554,
+  "5276": 555,
+  "5308": 556,
+  "5320": 557,
+  "5325": 558,
+  "5335": 559,
+  "5347": 560,
+  "5362": 561,
+  "5367": 562,
+  "5389": 563,
+  "5477": 564,
+  "5485": 565,
+  "5512": 566,
+  "5518": 567,
+  "5531": 568,
+  "5541": 569,
+  "5554": 570,
+  "5615": 571,
+  "5641": 572,
+  "5642": 573,
+  "5670": 574,
+  "5682": 575,
+  "5724": 576,
+  "5730": 577,
+  "5770": 578,
+  "5797": 579,
+  "5801": 580,
+  "5808": 581,
+  "5919": 582,
+  "6000": 583,
+  "6027": 584,
+  "6033": 585,
+  "6091": 586,
+  "6133": 587,
+  "6181": 588,
+  "6200": 589,
+  "6216": 590,
+  "6217": 591,
+  "6225": 592,
+  "6229": 593,
+  "6252": 594,
+  "6263": 595,
+  "6292": 596,
+  "6293": 597,
+  "6316": 598,
+  "6321": 599,
+  "6347": 600,
+  "6348": 601,
+  "6388": 602,
+  "6489": 603,
+  "6491": 604,
+  "6519": 605,
+  "6526": 606,
+  "6632": 607,
+  "6697": 608,
+  "6723": 609,
+  "6730": 610,
+  "6733": 611,
+  "6747": 612,
+  "6754": 613,
+  "6792": 614,
+  "6824": 615,
+  "6885": 616,
+  "6902": 617,
+  "6916": 618,
+  "6960": 619,
+  "6962": 620,
+  "6987": 621,
+  "6993": 622,
+  "7005": 623,
+  "7006": 624,
+  "7018": 625,
+  "7063": 626,
+  "7138": 627,
+  "7143": 628,
+  "7164": 629,
+  "7206": 630,
+  "7210": 631,
+  "7304": 632,
+  "7363": 633,
+  "7385": 634,
+  "7453": 635,
+  "7495": 636,
+  "7515": 637,
+  "7528": 638,
+  "7710": 639,
+  "7835": 640,
+  "7836": 641,
+  "7870": 642,
+  "7893": 643,
+  "7962": 644,
+  "7992": 645,
+  "8083": 646,
+  "8110": 647,
+  "8112": 648,
+  "8122": 649,
+  "8126": 650,
+  "8136": 651,
+  "8186": 652,
+  "8213": 653,
+  "8249": 654,
+  "8393": 655,
+  "8407": 656,
+  "8424": 657,
+  "8458": 658,
+  "8481": 659,
+  "8536": 660,
+  "8566": 661,
+  "8568": 662,
+  "8629": 663,
+  "8733": 664,
+  "8747": 665,
+  "8770": 666,
+  "8778": 667,
+  "8783": 668,
+  "8801": 669,
+  "8811": 670,
+  "8835": 671,
+  "8862": 672,
+  "8869": 673,
+  "8890": 674,
+  "8938": 675,
+  "9004": 676,
+  "9008": 677,
+  "9011": 678,
+  "9048": 679,
+  "9110": 680,
+  "9120": 681,
+  "9157": 682,
+  "9225": 683,
+  "9252": 684,
+  "9355": 685,
+  "9427": 686,
+  "9465": 687,
+  "9466": 688,
+  "9511": 689,
+  "9594": 690,
+  "9630": 691,
+  "9637": 692,
+  "9652": 693,
+  "9686": 694,
+  "9700": 695,
+  "9730": 696,
+  "9769": 697,
+  "9831": 698,
+  "9988": 699,
+  "10024": 700,
+  "10025": 701,
+  "10045": 702,
+  "10071": 703,
+  "10073": 704,
+  "10084": 705,
+  "10097": 706,
+  "10098": 707,
+  "10126": 708,
+  "10194": 709,
+  "10216": 710,
+  "10253": 711,
+  "10274": 712,
+  "10317": 713,
+  "10333": 714,
+  "10337": 715,
+  "10444": 716,
+  "10478": 717,
+  "10484": 718,
+  "10524": 719,
+  "10540": 720,
+  "10574": 721,
+  "10581": 722,
+  "10609": 723,
+  "10628": 724,
+  "10649": 725,
+  "10672": 726,
+  "10696": 727,
+  "10727": 728,
+  "10728": 729,
+  "10732": 730,
+  "10737": 731,
+  "10746": 732,
+  "10773": 733,
+  "10776": 734,
+  "10832": 735,
+  "10900": 736,
+  "10910": 737,
+  "10946": 738,
+  "10994": 739,
+  "11015": 740,
+  "11091": 741,
+  "11222": 742,
+  "11249": 743,
+  "11269": 744,
+  "11280": 745,
+  "11305": 746,
+  "11354": 747,
+  "11368": 748,
+  "11369": 749,
+  "11388": 750,
+  "11403": 751,
+  "11594": 752,
+  "11620": 753,
+  "11707": 754,
+  "11715": 755,
+  "11752": 756,
+  "11810": 757,
+  "11814": 758,
+  "11857": 759,
+  "11860": 760,
+  "11885": 761,
+  "12052": 762,
+  "12077": 763,
+  "12108": 764,
+  "12134": 765,
+  "12155": 766,
+  "12183": 767,
+  "12260": 768,
+  "12341": 769,
+  "12440": 770,
+  "12489": 771,
+  "12635": 772,
+  "12646": 773,
+  "12674": 774,
+  "12717": 775,
+  "12812": 776,
+  "12821": 777,
+  "12826": 778,
+  "12997": 779,
+  "13028": 780,
+  "13081": 781,
+  "13095": 782,
+  "13102": 783,
+  "13115": 784,
+  "13140": 785,
+  "13154": 786,
+  "13178": 787,
+  "13264": 788,
+  "13279": 789,
+  "13387": 790,
+  "13441": 791,
+  "13493": 792,
+  "13510": 793,
+  "13559": 794,
+  "13576": 795,
+  "13616": 796,
+  "13645": 797,
+  "13652": 798,
+  "13827": 799,
+  "13910": 800,
+  "14080": 801,
+  "14107": 802,
+  "14113": 803,
+  "14122": 804,
+  "14151": 805,
+  "14210": 806,
+  "14287": 807,
+  "14448": 808,
+  "14605": 809,
+  "14606": 810,
+  "14609": 811,
+  "14635": 812,
+  "14720": 813,
+  "14746": 814,
+  "14796": 815,
+  "14949": 816,
+  "15003": 817,
+  "15104": 818,
+  "15184": 819,
+  "15190": 820,
+  "15192": 821,
+  "15222": 822,
+  "15279": 823,
+  "15286": 824,
+  "15362": 825,
+  "15377": 826,
+  "15433": 827,
+  "15462": 828,
+  "15482": 829,
+  "15601": 830,
+  "15607": 831,
+  "15627": 832,
+  "15715": 833,
+  "15769": 834,
+  "15772": 835,
+  "15887": 836,
+  "15938": 837,
+  "15959": 838,
+  "15962": 839,
+  "16080": 840,
+  "16262": 841,
+  "16338": 842,
+  "16339": 843,
+  "16382": 844,
+  "16697": 845,
+  "16711": 846,
+  "16833": 847,
+  "16845": 848,
+  "16869": 849,
+  "16903": 850,
+  "16915": 851,
+  "16948": 852,
+  "16960": 853,
+  "16977": 854,
+  "17028": 855,
+  "17060": 856,
+  "17105": 857,
+  "17187": 858,
+  "17251": 859,
+  "17258": 860,
+  "17270": 861,
+  "17358": 862,
+  "17392": 863,
+  "17434": 864,
+  "17449": 865,
+  "17452": 866,
+  "17514": 867,
+  "17544": 868,
+  "17580": 869,
+  "17606": 870,
+  "17684": 871,
+  "17757": 872,
+  "17797": 873,
+  "17872": 874,
+  "17930": 875,
+  "17996": 876,
+  "18205": 877,
+  "18274": 878,
+  "18316": 879,
+  "18439": 880,
+  "18525": 881,
+  "18536": 882,
+  "18541": 883,
+  "18595": 884,
+  "18619": 885,
+  "18638": 886,
+  "18680": 887,
+  "18812": 888,
+  "18931": 889,
+  "18970": 890,
+  "19031": 891,
+  "19067": 892,
+  "19126": 893,
+  "19216": 894,
+  "19227": 895,
+  "19290": 896,
+  "19329": 897,
+  "19394": 898,
+  "19540": 899,
+  "19733": 900,
+  "19736": 901,
+  "19776": 902,
+  "19866": 903,
+  "19907": 904,
+  "19972": 905,
+  "20053": 906,
+  "20100": 907,
+  "20103": 908,
+  "20108": 909,
+  "20110": 910,
+  "20233": 911,
+  "20243": 912,
+  "20245": 913,
+  "20358": 914,
+  "20370": 915,
+  "20393": 916,
+  "20399": 917,
+  "20531": 918,
+  "20568": 919,
+  "20605": 920,
+  "20607": 921,
+  "20708": 922,
+  "20735": 923,
+  "20822": 924,
+  "21047": 925,
+  "21075": 926,
+  "21143": 927,
+  "21152": 928,
+  "21200": 929,
+  "21434": 930,
+  "21519": 931,
+  "21538": 932,
+  "21830": 933,
+  "21887": 934,
+  "21920": 935,
+  "21946": 936,
+  "22059": 937,
+  "22377": 938,
+  "22460": 939,
+  "22514": 940,
+  "22544": 941,
+  "22631": 942,
+  "22711": 943,
+  "22997": 944,
+  "23095": 945,
+  "23114": 946,
+  "23135": 947,
+  "23254": 948,
+  "23423": 949,
+  "23553": 950,
+  "23576": 951,
+  "23577": 952,
+  "23637": 953,
+  "23671": 954,
+  "23697": 955,
+  "23774": 956,
+  "23834": 957,
+  "24108": 958,
+  "24201": 959,
+  "24301": 960,
+  "24409": 961,
+  "24470": 962,
+  "24520": 963,
+  "24614": 964,
+  "24625": 965,
+  "24705": 966,
+  "24823": 967,
+  "25075": 968,
+  "25088": 969,
+  "25089": 970,
+  "25098": 971,
+  "25155": 972,
+  "25160": 973,
+  "25196": 974,
+  "25228": 975,
+  "25282": 976,
+  "25286": 977,
+  "25295": 978,
+  "25356": 979,
+  "25455": 980,
+  "25526": 981,
+  "25540": 982,
+  "25568": 983,
+  "25626": 984,
+  "25632": 985,
+  "25738": 986,
+  "25746": 987,
+  "25985": 988,
+  "25994": 989,
+  "26047": 990,
+  "26205": 991,
+  "26407": 992,
+  "26414": 993,
+  "26419": 994,
+  "26427": 995,
+  "26458": 996,
+  "26483": 997,
+  "26494": 998,
+  "26514": 999,
+  "26577": 1000,
+  "26675": 1001,
+  "26931": 1002,
+  "27190": 1003,
+  "27282": 1004,
+  "27417": 1005,
+  "27428": 1006,
+  "27578": 1007,
+  "27835": 1008,
+  "27843": 1009,
+  "27845": 1010,
+  "27865": 1011,
+  "27990": 1012,
+  "28029": 1013,
+  "28043": 1014,
+  "28119": 1015,
+  "28246": 1016,
+  "28256": 1017,
+  "28344": 1018,
+  "28400": 1019,
+  "28625": 1020,
+  "28639": 1021,
+  "28658": 1022,
+  "28676": 1023,
+  "28705": 1024,
+  "28787": 1025,
+  "28793": 1026,
+  "28900": 1027,
+  "28963": 1028,
+  "29008": 1029,
+  "29028": 1030,
+  "29045": 1031,
+  "29141": 1032,
+  "29294": 1033,
+  "29350": 1034,
+  "29413": 1035,
+  "29465": 1036,
+  "29511": 1037,
+  "29856": 1038,
+  "29876": 1039,
+  "29928": 1040,
+  "29956": 1041,
+  "29963": 1042,
+  "30149": 1043,
+  "30223": 1044,
+  "30234": 1045,
+  "30331": 1046,
+  "30477": 1047,
+  "30528": 1048,
+  "30540": 1049,
+  "30607": 1050,
+  "30710": 1051,
+  "30805": 1052,
+  "31157": 1053,
+  "31225": 1054,
+  "31311": 1055,
+  "31349": 1056,
+  "31469": 1057,
+  "31531": 1058,
+  "31548": 1059,
+  "31674": 1060,
+  "31853": 1061,
+  "31952": 1062,
+  "31967": 1063,
+  "32190": 1064,
+  "32314": 1065,
+  "32710": 1066,
+  "32752": 1067,
+  "32759": 1068,
+  "32850": 1069,
+  "32878": 1070,
+  "33015": 1071,
+  "33192": 1072,
+  "33397": 1073,
+  "33515": 1074,
+  "33823": 1075,
+  "33853": 1076,
+  "33896": 1077,
+  "34081": 1078,
+  "34152": 1079,
+  "34182": 1080,
+  "34435": 1081,
+  "34437": 1082,
+  "34565": 1083,
+  "34815": 1084,
+  "34822": 1085,
+  "34830": 1086,
+  "34850": 1087,
+  "34860": 1088,
+  "34928": 1089,
+  "35010": 1090,
+  "35150": 1091,
+  "35220": 1092,
+  "35450": 1093,
+  "35497": 1094,
+  "35560": 1095,
+  "35662": 1096,
+  "35785": 1097,
+  "35838": 1098,
+  "35853": 1099,
+  "35904": 1100,
+  "36099": 1101,
+  "36436": 1102,
+  "36444": 1103,
+  "36643": 1104,
+  "36890": 1105,
+  "37017": 1106,
+  "37111": 1107,
+  "37232": 1108,
+  "37244": 1109,
+  "37511": 1110,
+  "37547": 1111,
+  "37930": 1112,
+  "38101": 1113,
+  "38169": 1114,
+  "38185": 1115,
+  "38279": 1116,
+  "38415": 1117,
+  "38610": 1118,
+  "38639": 1119,
+  "38812": 1120,
+  "38828": 1121,
+  "38893": 1122,
+  "39036": 1123,
+  "39073": 1124,
+  "39112": 1125,
+  "39124": 1126,
+  "39180": 1127,
+  "39225": 1128,
+  "39294": 1129,
+  "39298": 1130,
+  "39311": 1131,
+  "39451": 1132,
+  "39583": 1133,
+  "39667": 1134,
+  "39710": 1135,
+  "39830": 1136,
+  "39944": 1137,
+  "39961": 1138,
+  "39992": 1139,
+  "40055": 1140,
+  "40110": 1141,
+  "40492": 1142,
+  "40625": 1143,
+  "40860": 1144,
+  "40945": 1145,
+  "40967": 1146,
+  "41071": 1147,
+  "41093": 1148,
+  "41236": 1149,
+  "41278": 1150,
+  "41305": 1151,
+  "41376": 1152,
+  "41625": 1153,
+  "41664": 1154,
+  "41921": 1155,
+  "41934": 1156,
+  "41967": 1157,
+  "41987": 1158,
+  "42441": 1159,
+  "42502": 1160,
+  "42657": 1161,
+  "42723": 1162,
+  "42750": 1163,
+  "42935": 1164,
+  "43014": 1165,
+  "43222": 1166,
+  "43296": 1167,
+  "43301": 1168,
+  "44190": 1169,
+  "44338": 1170,
+  "44415": 1171,
+  "44469": 1172,
+  "44545": 1173,
+  "44554": 1174,
+  "44750": 1175,
+  "44797": 1176,
+  "44906": 1177,
+  "45009": 1178,
+  "45103": 1179,
+  "45111": 1180,
+  "45168": 1181,
+  "45558": 1182,
+  "45866": 1183,
+  "45876": 1184,
+  "45961": 1185,
+  "46245": 1186,
+  "46429": 1187,
+  "46494": 1188,
+  "46591": 1189,
+  "46931": 1190,
+  "47015": 1191,
+  "47026": 1192,
+  "47178": 1193,
+  "47206": 1194,
+  "47484": 1195,
+  "47535": 1196,
+  "47615": 1197,
+  "47767": 1198,
+  "47820": 1199,
+  "47822": 1200,
+  "48000": 1201,
+  "48067": 1202,
+  "48328": 1203,
+  "48416": 1204,
+  "48454": 1205,
+  "48850": 1206,
+  "48931": 1207,
+  "48999": 1208,
+  "49236": 1209,
+  "49551": 1210,
+  "49591": 1211,
+  "49602": 1212,
+  "50010": 1213,
+  "50103": 1214,
+  "50502": 1215,
+  "50914": 1216,
+  "51205": 1217,
+  "51441": 1218,
+  "51526": 1219,
+  "51737": 1220,
+  "52087": 1221,
+  "52311": 1222,
+  "52330": 1223,
+  "52360": 1224,
+  "52462": 1225,
+  "52596": 1226,
+  "52603": 1227,
+  "52762": 1228,
+  "52777": 1229,
+  "52962": 1230,
+  "53187": 1231,
+  "53358": 1232,
+  "53377": 1233,
+  "53403": 1234,
+  "53543": 1235,
+  "53565": 1236,
+  "53696": 1237,
+  "53722": 1238,
+  "53902": 1239,
+  "53990": 1240,
+  "54097": 1241,
+  "54176": 1242,
+  "54266": 1243,
+  "54272": 1244,
+  "54485": 1245,
+  "54681": 1246,
+  "54776": 1247,
+  "54788": 1248,
+  "54860": 1249,
+  "54921": 1250,
+  "55002": 1251,
+  "55122": 1252,
+  "55271": 1253,
+  "55279": 1254,
+  "55300": 1255,
+  "55424": 1256,
+  "55451": 1257,
+  "55672": 1258,
+  "55779": 1259,
+  "55862": 1260,
+  "55880": 1261,
+  "56094": 1262,
+  "56099": 1263,
+  "56184": 1264,
+  "56246": 1265,
+  "56503": 1266,
+  "56528": 1267,
+  "56552": 1268,
+  "56589": 1269,
+  "56692": 1270,
+  "56809": 1271,
+  "56868": 1272,
+  "56970": 1273,
+  "57258": 1274,
+  "57730": 1275,
+  "57766": 1276,
+  "57839": 1277,
+  "57973": 1278,
+  "58132": 1279,
+  "58147": 1280,
+  "58197": 1281,
+  "58315": 1282,
+  "58369": 1283,
+  "58576": 1284,
+  "58633": 1285,
+  "58708": 1286,
+  "58899": 1287,
+  "59004": 1288,
+  "59042": 1289,
+  "59548": 1290,
+  "60024": 1291,
+  "60050": 1292,
+  "60477": 1293,
+  "60577": 1294,
+  "61194": 1295,
+  "61344": 1296,
+  "61353": 1297,
+  "61387": 1298,
+  "61454": 1299,
+  "61745": 1300,
+  "62067": 1301,
+  "62090": 1302,
+  "62344": 1303,
+  "62444": 1304,
+  "62552": 1305,
+  "62822": 1306,
+  "63099": 1307,
+  "63278": 1308,
+  "63340": 1309,
+  "63423": 1310,
+  "63698": 1311,
+  "64032": 1312,
+  "64099": 1313,
+  "64115": 1314,
+  "64200": 1315,
+  "64281": 1316,
+  "64345": 1317,
+  "64487": 1318,
+  "64753": 1319,
+  "64934": 1320,
+  "65166": 1321,
+  "65856": 1322,
+  "66163": 1323,
+  "66652": 1324,
+  "66696": 1325,
+  "66780": 1326,
+  "66815": 1327,
+  "66842": 1328,
+  "66852": 1329,
+  "66999": 1330,
+  "67100": 1331,
+  "67443": 1332,
+  "67554": 1333,
+  "67717": 1334,
+  "68146": 1335,
+  "68151": 1336,
+  "68330": 1337,
+  "68396": 1338,
+  "68593": 1339,
+  "68669": 1340,
+  "68817": 1341,
+  "69034": 1342,
+  "69239": 1343,
+  "69406": 1344,
+  "69407": 1345,
+  "69432": 1346,
+  "69578": 1347,
+  "69694": 1348,
+  "69814": 1349,
+  "69947": 1350,
+  "70138": 1351,
+  "70257": 1352,
+  "70338": 1353,
+  "70383": 1354,
+  "70437": 1355,
+  "70854": 1356,
+  "70865": 1357,
+  "70866": 1358,
+  "71097": 1359,
+  "71253": 1360,
+  "71296": 1361,
+  "71398": 1362,
+  "71437": 1363,
+  "71544": 1364,
+  "71629": 1365,
+  "71826": 1366,
+  "72239": 1367,
+  "72343": 1368,
+  "72390": 1369,
+  "72583": 1370,
+  "72628": 1371,
+  "73077": 1372,
+  "73225": 1373,
+  "73512": 1374,
+  "73663": 1375,
+  "73866": 1376,
+  "73875": 1377,
+  "73911": 1378,
+  "73990": 1379,
+  "74261": 1380,
+  "74772": 1381,
+  "74961": 1382,
+  "75288": 1383,
+  "75401": 1384,
+  "75707": 1385,
+  "75802": 1386,
+  "75878": 1387,
+  "76063": 1388,
+  "76111": 1389,
+  "76181": 1390,
+  "76351": 1391,
+  "76535": 1392,
+  "76777": 1393,
+  "76871": 1394,
+  "77155": 1395,
+  "77467": 1396,
+  "77507": 1397,
+  "77510": 1398,
+  "77643": 1399,
+  "78418": 1400,
+  "78428": 1401,
+  "78540": 1402,
+  "78810": 1403,
+  "78968": 1404,
+  "79020": 1405,
+  "79212": 1406,
+  "79444": 1407,
+  "79807": 1408,
+  "79835": 1409,
+  "80024": 1410,
+  "80457": 1411,
+  "80701": 1412,
+  "80990": 1413,
+  "81585": 1414,
+  "81772": 1415,
+  "81807": 1416,
+  "82036": 1417,
+  "82123": 1418,
+  "82200": 1419,
+  "82629": 1420,
+  "82649": 1421,
+  "82863": 1422,
+  "83450": 1423,
+  "83571": 1424,
+  "83595": 1425,
+  "83921": 1426,
+  "83937": 1427,
+  "83977": 1428,
+  "84112": 1429,
+  "84189": 1430,
+  "84767": 1431,
+  "84807": 1432,
+  "84820": 1433,
+  "85210": 1434,
+  "85298": 1435,
+  "85369": 1436,
+  "85428": 1437,
+  "85530": 1438,
+  "85677": 1439,
+  "85724": 1440,
+  "85920": 1441,
+  "86084": 1442,
+  "86400": 1443,
+  "86518": 1444,
+  "86627": 1445,
+  "86656": 1446,
+  "86906": 1447,
+  "87034": 1448,
+  "87099": 1449,
+  "87120": 1450,
+  "87197": 1451,
+  "87200": 1452,
+  "87251": 1453,
+  "87512": 1454,
+  "87556": 1455,
+  "87685": 1456,
+  "87730": 1457,
+  "87770": 1458,
+  "87780": 1459,
+  "87802": 1460,
+  "87909": 1461,
+  "87996": 1462,
+  "88256": 1463,
+  "88295": 1464,
+  "88531": 1465,
+  "89049": 1466,
+  "89129": 1467,
+  "89615": 1468,
+  "90065": 1469,
+  "90099": 1470,
+  "90539": 1471,
+  "90745": 1472,
+  "90818": 1473,
+  "91010": 1474,
+  "91065": 1475,
+  "91427": 1476,
+  "91905": 1477,
+  "92347": 1478,
+  "92510": 1479,
+  "92644": 1480,
+  "92816": 1481,
+  "92835": 1482,
+  "93471": 1483,
+  "93540": 1484,
+  "94381": 1485,
+  "94487": 1486,
+  "94913": 1487,
+  "95102": 1488,
+  "95524": 1489,
+  "95531": 1490,
+  "95557": 1491,
+  "95614": 1492,
+  "95741": 1493,
+  "95952": 1494,
+  "95954": 1495,
+  "96099": 1496,
+  "96146": 1497,
+  "96266": 1498,
+  "96488": 1499,
+  "96521": 1500,
+  "96772": 1501,
+  "96899": 1502,
+  "97183": 1503,
+  "97219": 1504,
+  "97253": 1505,
+  "97306": 1506,
+  "97378": 1507,
+  "97587": 1508,
+  "97668": 1509,
+  "97686": 1510,
+  "97775": 1511,
+  "97800": 1512,
+  "97809": 1513,
+  "97914": 1514,
+  "97998": 1515,
+  "98040": 1516,
+  "98398": 1517,
+  "98443": 1518,
+  "98485": 1519,
+  "98765": 1520,
+  "98878": 1521,
+  "99156": 1522,
+  "99540": 1523,
+  "99603": 1524,
+  "99633": 1525,
+  "99954": 1526,
+  "99972": 1527,
+  "100127": 1528,
+  "<pad>": 0,
+  "<sos>": 1,
+  "<eos>": 2
+}

src/model/rellow-2.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a4a07540e4d12ec2919c9ae2004bc523efc5a27da9ec08c9945479a98a3e8e50
+size 14227945

src/services/model.py

@@ -0,0 +1,48 @@
+import os
+import json
+import torch
+# from pathlib import Path
+from services.transformer import TinyTransformer
+
+# Internal constants for file paths
+# _MODEL_DIR = os.path.join(os.path.dirname(os.path.dirname(__file__)), "model")
+# _VOCAB_DIR = os.path.join(os.path.dirname(os.path.dirname(__file__)), "data")
+_MODEL_PATH = os.path.join(os.path.dirname(os.path.dirname(__file__)), "model", "rellow-2.pt")
+_VOCAB_PATH = os.path.join(os.path.dirname(os.path.dirname(__file__)), "data", "vocab.json")
+
+# Internal device selection
+_DEVICE = torch.device("mps" if torch.backends.mps.is_available() else "cpu")
+
+# Save the model state and vocabulary to disk.
+# def save_model(model, vocab):
+#   # Create necessary directories
+#   os.makedirs(_MODEL_DIR, exist_ok=True)
+#   os.makedirs(_VOCAB_DIR, exist_ok=True)
+  
+#   # Save model state
+#   torch.save(model.state_dict(), _MODEL_PATH)
+  
+#   # Save vocabulary
+#   with open(_VOCAB_PATH, "w", encoding="utf-8") as f:
+#     json.dump(vocab, f, ensure_ascii=False, indent=2)
+  
+#   print(f"Model saved to {_MODEL_PATH}")
+#   print(f"Vocabulary saved to {_VOCAB_PATH}")
+
+# Load the model and its vocabulary from disk.
+def load_model():
+  # Load vocabulary
+  with open(_VOCAB_PATH, "r", encoding="utf-8") as f:
+    vocab = json.load(f)
+  inv_vocab = {int(v): k for k, v in vocab.items()}
+  
+  # Initialize and load model
+  model = TinyTransformer(vocab_size=len(vocab)).to(_DEVICE)
+  model.load_state_dict(torch.load(_MODEL_PATH, map_location=_DEVICE))
+  model.eval()
+  
+  return model, vocab, inv_vocab
+
+# Get the device being used for model operations.
+def get_device():
+  return _DEVICE 

src/services/transformer.py

@@ -0,0 +1,46 @@
+import torch
+import torch.nn as nn
+from constants.tokens import PAD_ID
+
+class TinyTransformer(nn.Module):
+  def __init__(self, vocab_size, d_model=256, nhead=4, num_layers=2, dim_feedforward=512, dropout=0.1):
+    super().__init__()
+    self.embedding = nn.Embedding(vocab_size, d_model, padding_idx=PAD_ID)
+    self.pos_encoder = PositionalEncoding(d_model, dropout)
+    encoder_layer = nn.TransformerEncoderLayer(d_model, nhead, dim_feedforward, dropout, batch_first=True)
+    self.encoder = nn.TransformerEncoder(encoder_layer, num_layers=num_layers)
+    decoder_layer = nn.TransformerDecoderLayer(d_model, nhead, dim_feedforward, dropout, batch_first=True)
+    self.decoder = nn.TransformerDecoder(decoder_layer, num_layers=num_layers)
+    self.out = nn.Linear(d_model, vocab_size)
+
+  # Keep tensors in batch-first format
+  def forward(self, src, tgt):
+    tgt_mask = nn.Transformer.generate_square_subsequent_mask(tgt.size(1)).to(src.device).bool()
+    src_emb = self.pos_encoder(self.embedding(src))
+    tgt_emb = self.pos_encoder(self.embedding(tgt))
+    # Create padding masks
+    src_padding_mask = (src == PAD_ID).bool()
+    tgt_padding_mask = (tgt == PAD_ID).bool()
+    memory = self.encoder(src_emb, src_key_padding_mask=src_padding_mask)
+    output = self.decoder(tgt_emb, memory, tgt_mask=tgt_mask, tgt_key_padding_mask=tgt_padding_mask)
+    return self.out(output)  # (batch, seq_len, vocab)
+
+  def generate_src_mask(self, size):
+    return torch.zeros((size, size), device='cpu').type(torch.bool)
+
+class PositionalEncoding(nn.Module):
+  def __init__(self, d_model, dropout=0.1, max_len=512):
+    super().__init__()
+    self.dropout = nn.Dropout(p=dropout)
+    position = torch.arange(0, max_len).unsqueeze(1)
+    div_term = torch.exp(
+      torch.arange(0, d_model, 2) * (-torch.log(torch.tensor(10000.0)) / d_model)
+    )
+    pe = torch.zeros(max_len, d_model)
+    pe[:, 0::2] = torch.sin(position * div_term)  # even indices
+    pe[:, 1::2] = torch.cos(position * div_term)  # odd indices
+    self.register_buffer('pe', pe.unsqueeze(0))
+
+  def forward(self, x):
+    x = x + self.pe[:, :x.size(1), :].to(x.device)
+    return self.dropout(x)