Upload folder using huggingface_hub

config.json

@@ -0,0 +1,23 @@
+{
+    "n_layer": 6,
+    "n_head": 4,
+    "n_embd": 256,
+    "max_position_embeddings": 512,
+    "vocab_size": 194,
+    "model_type": "gpt2",
+    "architectures": [
+        "GPT2LMHeadModel"
+    ],
+    "attn_pdrop": 0.1,
+    "embd_pdrop": 0.1,
+    "resid_pdrop": 0.1,
+    "initializer_range": 0.02,
+    "layer_norm_epsilon": 1e-05,
+    "n_positions": 512,
+    "scale_attn_weights": true,
+    "use_cache": true,
+    "pad_token_id": 0,
+    "bos_token_id": 2,
+    "eos_token_id": 3,
+    "unk_token_id": 1
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:62dfad8cc0ff00af62a3ce18f5369a5428ee54337093da15d693b5a1d6e20f50
+size 26000829

tokenizer.json

@@ -0,0 +1,895 @@
+{
+  "version": "1.0",
+  "truncation": null,
+  "padding": null,
+  "added_tokens": [
+    {
+      "id": 0,
+      "content": "[PAD]",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 1,
+      "content": "[UNK]",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 2,
+      "content": "[BOS]",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 3,
+      "content": "[EOS]",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    }
+  ],
+  "normalizer": {
+    "type": "Sequence",
+    "normalizers": [
+      {
+        "type": "Lowercase"
+      },
+      {
+        "type": "NFKC"
+      }
+    ]
+  },
+  "pre_tokenizer": {
+    "type": "Whitespace"
+  },
+  "post_processor": null,
+  "decoder": null,
+  "model": {
+    "type": "BPE",
+    "dropout": null,
+    "unk_token": "[UNK]",
+    "continuing_subword_prefix": null,
+    "end_of_word_suffix": null,
+    "fuse_unk": false,
+    "byte_fallback": false,
+    "ignore_merges": false,
+    "vocab": {
+      "[PAD]": 0,
+      "[UNK]": 1,
+      "[BOS]": 2,
+      "[EOS]": 3,
+      "(": 4,
+      ")": 5,
+      ",": 6,
+      ".": 7,
+      "a": 8,
+      "b": 9,
+      "c": 10,
+      "d": 11,
+      "e": 12,
+      "f": 13,
+      "g": 14,
+      "h": 15,
+      "i": 16,
+      "j": 17,
+      "k": 18,
+      "l": 19,
+      "m": 20,
+      "n": 21,
+      "o": 22,
+      "p": 23,
+      "r": 24,
+      "s": 25,
+      "t": 26,
+      "u": 27,
+      "v": 28,
+      "y": 29,
+      "z": 30,
+      "ç": 31,
+      "ö": 32,
+      "ü": 33,
+      "ğ": 34,
+      "ı": 35,
+      "ş": 36,
+      "in": 37,
+      "la": 38,
+      "er": 39,
+      "el": 40,
+      "il": 41,
+      "ir": 42,
+      "ma": 43,
+      "me": 44,
+      "ır": 45,
+      "ka": 46,
+      "bir": 47,
+      "dil": 48,
+      "ar": 49,
+      "en": 50,
+      "or": 51,
+      "tü": 52,
+      "ve": 53,
+      "ya": 54,
+      "lar": 55,
+      "da": 56,
+      "im": 57,
+      "it": 58,
+      "mo": 59,
+      "mod": 60,
+      "em": 61,
+      "et": 62,
+      "eğ": 63,
+      "ha": 64,
+      "si": 65,
+      "yor": 66,
+      "ün": 67,
+      "lama": 68,
+      "eri": 69,
+      "model": 70,
+      "ay": 71,
+      "dır": 72,
+      "eka": 73,
+      "kç": 74,
+      "ok": 75,
+      "ol": 76,
+      "pay": 77,
+      "zeka": 78,
+      "ın": 79,
+      "lan": 80,
+      "tür": 81,
+      "yapay": 82,
+      "imi": 83,
+      "eğit": 84,
+      "al": 85,
+      "alan": 86,
+      "bu": 87,
+      "de": 88,
+      "ek": 89,
+      "gün": 90,
+      "iz": 91,
+      "iç": 92,
+      "kt": 93,
+      "li": 94,
+      "oj": 95,
+      "va": 96,
+      "ön": 97,
+      "ük": 98,
+      "ğr": 99,
+      "ğı": 100,
+      "şt": 101,
+      "erin": 102,
+      "eli": 103,
+      "kçe": 104,
+      "türkçe": 105,
+      "için": 106,
+      "an": 107,
+      "ba": 108,
+      "bil": 109,
+      "di": 110,
+      "do": 111,
+      "ec": 112,
+      "fa": 113,
+      "gu": 114,
+      "ku": 115,
+      "kü": 116,
+      "kin": 117,
+      "le": 118,
+      "lı": 119,
+      "pr": 120,
+      "sı": 121,
+      "uy": 122,
+      "veri": 123,
+      "çok": 124,
+      "öğr": 125,
+      "ür": 126,
+      "malar": 127,
+      "emli": 128,
+      "modelin": 129,
+      "önemli": 130,
+      "gulama": 131,
+      "uygulama": 132,
+      "öğren": 133,
+      "ah": 134,
+      "bü": 135,
+      "ca": 136,
+      "du": 137,
+      "dı": 138,
+      "dir": 139,
+      "dün": 140,
+      "eç": 141,
+      "esi": 142,
+      "edi": 143,
+      "gin": 144,
+      "ip": 145,
+      "isi": 146,
+      "keli": 147,
+      "lem": 148,
+      "leri": 149,
+      "nu": 150,
+      "nol": 151,
+      "onu": 152,
+      "par": 153,
+      "pha": 154,
+      "ra": 155,
+      "tu": 156,
+      "tok": 157,
+      "tek": 158,
+      "yi": 159,
+      "yük": 160,
+      "zla": 161,
+      "zel": 162,
+      "ğal": 163,
+      "inin": 164,
+      "elerin": 165,
+      "ile": 166,
+      "ilme": 167,
+      "makin": 168,
+      "met": 169,
+      "eni": 170,
+      "eniz": 171,
+      "tüpha": 172,
+      "ları": 173,
+      "etimi": 174,
+      "oldu": 175,
+      "eğitim": 176,
+      "ktır": 177,
+      "ğım": 178,
+      "ştir": 179,
+      "doğal": 180,
+      "kütüpha": 181,
+      "proj": 182,
+      "öğrenimi": 183,
+      "büyük": 184,
+      "dünya": 185,
+      "ediyor": 186,
+      "kelime": 187,
+      "noloj": 188,
+      "tokeniz": 189,
+      "teknoloj": 190,
+      "makine": 191,
+      "kütüphan": 192,
+      "tokenizer": 193
+    },
+    "merges": [
+      [
+        "i",
+        "n"
+      ],
+      [
+        "l",
+        "a"
+      ],
+      [
+        "e",
+        "r"
+      ],
+      [
+        "e",
+        "l"
+      ],
+      [
+        "i",
+        "l"
+      ],
+      [
+        "i",
+        "r"
+      ],
+      [
+        "m",
+        "a"
+      ],
+      [
+        "m",
+        "e"
+      ],
+      [
+        "ı",
+        "r"
+      ],
+      [
+        "k",
+        "a"
+      ],
+      [
+        "b",
+        "ir"
+      ],
+      [
+        "d",
+        "il"
+      ],
+      [
+        "a",
+        "r"
+      ],
+      [
+        "e",
+        "n"
+      ],
+      [
+        "o",
+        "r"
+      ],
+      [
+        "t",
+        "ü"
+      ],
+      [
+        "v",
+        "e"
+      ],
+      [
+        "y",
+        "a"
+      ],
+      [
+        "la",
+        "r"
+      ],
+      [
+        "d",
+        "a"
+      ],
+      [
+        "i",
+        "m"
+      ],
+      [
+        "i",
+        "t"
+      ],
+      [
+        "m",
+        "o"
+      ],
+      [
+        "mo",
+        "d"
+      ],
+      [
+        "e",
+        "m"
+      ],
+      [
+        "e",
+        "t"
+      ],
+      [
+        "e",
+        "ğ"
+      ],
+      [
+        "h",
+        "a"
+      ],
+      [
+        "s",
+        "i"
+      ],
+      [
+        "y",
+        "or"
+      ],
+      [
+        "ü",
+        "n"
+      ],
+      [
+        "la",
+        "ma"
+      ],
+      [
+        "er",
+        "i"
+      ],
+      [
+        "mod",
+        "el"
+      ],
+      [
+        "a",
+        "y"
+      ],
+      [
+        "d",
+        "ır"
+      ],
+      [
+        "e",
+        "ka"
+      ],
+      [
+        "k",
+        "ç"
+      ],
+      [
+        "o",
+        "k"
+      ],
+      [
+        "o",
+        "l"
+      ],
+      [
+        "p",
+        "ay"
+      ],
+      [
+        "z",
+        "eka"
+      ],
+      [
+        "ı",
+        "n"
+      ],
+      [
+        "la",
+        "n"
+      ],
+      [
+        "tü",
+        "r"
+      ],
+      [
+        "ya",
+        "pay"
+      ],
+      [
+        "im",
+        "i"
+      ],
+      [
+        "eğ",
+        "it"
+      ],
+      [
+        "a",
+        "l"
+      ],
+      [
+        "a",
+        "lan"
+      ],
+      [
+        "b",
+        "u"
+      ],
+      [
+        "d",
+        "e"
+      ],
+      [
+        "e",
+        "k"
+      ],
+      [
+        "g",
+        "ün"
+      ],
+      [
+        "i",
+        "z"
+      ],
+      [
+        "i",
+        "ç"
+      ],
+      [
+        "k",
+        "t"
+      ],
+      [
+        "l",
+        "i"
+      ],
+      [
+        "o",
+        "j"
+      ],
+      [
+        "v",
+        "a"
+      ],
+      [
+        "ö",
+        "n"
+      ],
+      [
+        "ü",
+        "k"
+      ],
+      [
+        "ğ",
+        "r"
+      ],
+      [
+        "ğ",
+        "ı"
+      ],
+      [
+        "ş",
+        "t"
+      ],
+      [
+        "er",
+        "in"
+      ],
+      [
+        "el",
+        "i"
+      ],
+      [
+        "kç",
+        "e"
+      ],
+      [
+        "tür",
+        "kçe"
+      ],
+      [
+        "iç",
+        "in"
+      ],
+      [
+        "a",
+        "n"
+      ],
+      [
+        "b",
+        "a"
+      ],
+      [
+        "b",
+        "il"
+      ],
+      [
+        "d",
+        "i"
+      ],
+      [
+        "d",
+        "o"
+      ],
+      [
+        "e",
+        "c"
+      ],
+      [
+        "f",
+        "a"
+      ],
+      [
+        "g",
+        "u"
+      ],
+      [
+        "k",
+        "u"
+      ],
+      [
+        "k",
+        "ü"
+      ],
+      [
+        "k",
+        "in"
+      ],
+      [
+        "l",
+        "e"
+      ],
+      [
+        "l",
+        "ı"
+      ],
+      [
+        "p",
+        "r"
+      ],
+      [
+        "s",
+        "ı"
+      ],
+      [
+        "u",
+        "y"
+      ],
+      [
+        "v",
+        "eri"
+      ],
+      [
+        "ç",
+        "ok"
+      ],
+      [
+        "ö",
+        "ğr"
+      ],
+      [
+        "ü",
+        "r"
+      ],
+      [
+        "ma",
+        "lar"
+      ],
+      [
+        "em",
+        "li"
+      ],
+      [
+        "model",
+        "in"
+      ],
+      [
+        "ön",
+        "emli"
+      ],
+      [
+        "gu",
+        "lama"
+      ],
+      [
+        "uy",
+        "gulama"
+      ],
+      [
+        "öğr",
+        "en"
+      ],
+      [
+        "a",
+        "h"
+      ],
+      [
+        "b",
+        "ü"
+      ],
+      [
+        "c",
+        "a"
+      ],
+      [
+        "d",
+        "u"
+      ],
+      [
+        "d",
+        "ı"
+      ],
+      [
+        "d",
+        "ir"
+      ],
+      [
+        "d",
+        "ün"
+      ],
+      [
+        "e",
+        "ç"
+      ],
+      [
+        "e",
+        "si"
+      ],
+      [
+        "e",
+        "di"
+      ],
+      [
+        "g",
+        "in"
+      ],
+      [
+        "i",
+        "p"
+      ],
+      [
+        "i",
+        "si"
+      ],
+      [
+        "k",
+        "eli"
+      ],
+      [
+        "l",
+        "em"
+      ],
+      [
+        "l",
+        "eri"
+      ],
+      [
+        "n",
+        "u"
+      ],
+      [
+        "n",
+        "ol"
+      ],
+      [
+        "o",
+        "nu"
+      ],
+      [
+        "p",
+        "ar"
+      ],
+      [
+        "p",
+        "ha"
+      ],
+      [
+        "r",
+        "a"
+      ],
+      [
+        "t",
+        "u"
+      ],
+      [
+        "t",
+        "ok"
+      ],
+      [
+        "t",
+        "ek"
+      ],
+      [
+        "y",
+        "i"
+      ],
+      [
+        "y",
+        "ük"
+      ],
+      [
+        "z",
+        "la"
+      ],
+      [
+        "z",
+        "el"
+      ],
+      [
+        "ğ",
+        "al"
+      ],
+      [
+        "in",
+        "in"
+      ],
+      [
+        "el",
+        "erin"
+      ],
+      [
+        "il",
+        "e"
+      ],
+      [
+        "il",
+        "me"
+      ],
+      [
+        "ma",
+        "kin"
+      ],
+      [
+        "me",
+        "t"
+      ],
+      [
+        "en",
+        "i"
+      ],
+      [
+        "en",
+        "iz"
+      ],
+      [
+        "tü",
+        "pha"
+      ],
+      [
+        "lar",
+        "ı"
+      ],
+      [
+        "et",
+        "imi"
+      ],
+      [
+        "ol",
+        "du"
+      ],
+      [
+        "eğit",
+        "im"
+      ],
+      [
+        "kt",
+        "ır"
+      ],
+      [
+        "ğı",
+        "m"
+      ],
+      [
+        "şt",
+        "ir"
+      ],
+      [
+        "do",
+        "ğal"
+      ],
+      [
+        "kü",
+        "tüpha"
+      ],
+      [
+        "pr",
+        "oj"
+      ],
+      [
+        "öğren",
+        "imi"
+      ],
+      [
+        "bü",
+        "yük"
+      ],
+      [
+        "dün",
+        "ya"
+      ],
+      [
+        "edi",
+        "yor"
+      ],
+      [
+        "keli",
+        "me"
+      ],
+      [
+        "nol",
+        "oj"
+      ],
+      [
+        "tok",
+        "eniz"
+      ],
+      [
+        "tek",
+        "noloj"
+      ],
+      [
+        "makin",
+        "e"
+      ],
+      [
+        "kütüpha",
+        "n"
+      ],
+      [
+        "tokeniz",
+        "er"
+      ]
+    ]
+  }
+}

tokenizer_config.json

@@ -0,0 +1,13 @@
+{
+    "model_max_length": 512,
+    "add_prefix_space": false,
+    "bos_token": "[BOS]",
+    "eos_token": "[EOS]",
+    "unk_token": "[UNK]",
+    "pad_token": "[PAD]",
+    "errors": "replace",
+    "tokenizer_class": "PreTrainedTokenizerFast",
+    "name_or_path": "/content",
+    "do_lower_case": true,
+    "special_tokens_map_file": null
+}

vocab.json

@@ -0,0 +1,196 @@
+{
+    "teknoloj": 190,
+    "makine": 191,
+    "ba": 108,
+    "yi": 159,
+    "it": 58,
+    "gu": 114,
+    "pha": 154,
+    "kt": 93,
+    "\u00fcr": 126,
+    "d\u0131r": 72,
+    "\u00e7ok": 124,
+    "dir": 139,
+    ".": 7,
+    "\u00e7": 31,
+    "ca": 136,
+    "or": 51,
+    ")": 5,
+    "t\u00fcrk\u00e7e": 105,
+    "par": 153,
+    "mod": 60,
+    "n": 21,
+    "[BOS]": 2,
+    "e\u011fitim": 176,
+    "li": 94,
+    "kin": 117,
+    "b": 9,
+    "d": 11,
+    "[UNK]": 1,
+    "d\u00fcn": 140,
+    "ha": 64,
+    "et": 62,
+    "kt\u0131r": 177,
+    "\u015ftir": 179,
+    "eniz": 171,
+    "yapay": 82,
+    "do\u011fal": 180,
+    "ve": 53,
+    "r": 24,
+    "g": 14,
+    "o": 22,
+    "e": 12,
+    "ra": 155,
+    "met": 169,
+    "er": 39,
+    "m": 20,
+    "z": 30,
+    "\u00f6\u011fr": 125,
+    "\u00f6\u011frenimi": 183,
+    "in": 37,
+    "oldu": 175,
+    "lar\u0131": 173,
+    "v": 28,
+    "el": 40,
+    "tu": 156,
+    "ma": 43,
+    "pr": 120,
+    "il": 41,
+    ",": 6,
+    "b\u00fcy\u00fck": 184,
+    "la": 38,
+    "im": 57,
+    "isi": 146,
+    "e\u011f": 63,
+    "ek": 89,
+    "em": 61,
+    "a": 8,
+    "ka": 46,
+    "\u00fck": 98,
+    "inin": 164,
+    "eni": 170,
+    "i": 16,
+    "e\u00e7": 141,
+    "\u0131": 35,
+    "du": 137,
+    "\u00fcn": 67,
+    "t\u00fc": 52,
+    "tokeniz": 189,
+    "noloj": 188,
+    "uy": 122,
+    "eri": 69,
+    "t\u00fcpha": 172,
+    "y\u00fck": 160,
+    "bil": 109,
+    "dil": 48,
+    "ediyor": 186,
+    "tek": 158,
+    "alan": 86,
+    "yor": 66,
+    "t\u00fcr": 81,
+    "\u011fal": 163,
+    "k\u00e7e": 104,
+    "(": 4,
+    "ar": 49,
+    "imi": 83,
+    "gulama": 131,
+    "\u011f\u0131m": 178,
+    "model": 70,
+    "bu": 87,
+    "zla": 161,
+    "f": 13,
+    "erin": 102,
+    "i\u00e7in": 106,
+    "\u011f": 34,
+    "uygulama": 132,
+    "emli": 128,
+    "makin": 168,
+    "veri": 123,
+    "elerin": 165,
+    "\u015ft": 101,
+    "k\u00e7": 74,
+    "[EOS]": 3,
+    "\u00f6": 32,
+    "di": 110,
+    "ah": 134,
+    "p": 23,
+    "esi": 142,
+    "me": 44,
+    "u": 27,
+    "zel": 162,
+    "nol": 151,
+    "modelin": 129,
+    "\u011f\u0131": 100,
+    "zeka": 78,
+    "tokenizer": 193,
+    "s\u0131": 121,
+    "ec": 112,
+    "edi": 143,
+    "le": 118,
+    "proj": 182,
+    "kelime": 187,
+    "y": 29,
+    "k\u00fct\u00fcpha": 181,
+    "c": 10,
+    "t": 26,
+    "tok": 157,
+    "ilme": 167,
+    "\u00f6n": 97,
+    "[PAD]": 0,
+    "ol": 76,
+    "lan": 80,
+    "\u015f": 36,
+    "ip": 145,
+    "gin": 144,
+    "an": 107,
+    "en": 50,
+    "\u011fr": 99,
+    "h": 15,
+    "si": 65,
+    "oj": 95,
+    "d\u00fcnya": 185,
+    "mo": 59,
+    "de": 88,
+    "ile": 166,
+    "keli": 147,
+    "fa": 113,
+    "etimi": 174,
+    "j": 17,
+    "g\u00fcn": 90,
+    "ok": 75,
+    "ay": 71,
+    "\u00fc": 33,
+    "ku": 115,
+    "va": 96,
+    "i\u00e7": 92,
+    "do": 111,
+    "onu": 152,
+    "\u0131n": 79,
+    "k": 18,
+    "bir": 47,
+    "ir": 42,
+    "l\u0131": 119,
+    "k\u00fc": 116,
+    "lama": 68,
+    "eli": 103,
+    "\u0131r": 45,
+    "eka": 73,
+    "pay": 77,
+    "l": 19,
+    "iz": 91,
+    "leri": 149,
+    "ya": 54,
+    "e\u011fit": 84,
+    "\u00f6nemli": 130,
+    "k\u00fct\u00fcphan": 192,
+    "\u00f6\u011fren": 133,
+    "d\u0131": 138,
+    "s": 25,
+    "da": 56,
+    "al": 85,
+    "b\u00fc": 135,
+    "lem": 148,
+    "malar": 127,
+    "lar": 55,
+    "nu": 150
+}