Training in progress, step 160
Browse files- config.json +2 -2
- model.safetensors +2 -2
- tokenizer.json +35 -100
config.json
CHANGED
|
@@ -40,7 +40,7 @@
|
|
| 40 |
"tie_word_embeddings": true,
|
| 41 |
"type_vocab_size": 2,
|
| 42 |
"use_cache": true,
|
| 43 |
-
"vocab_size":
|
| 44 |
},
|
| 45 |
"decoder_start_token_id": 2,
|
| 46 |
"dtype": "float32",
|
|
@@ -82,7 +82,7 @@
|
|
| 82 |
"tie_word_embeddings": true,
|
| 83 |
"type_vocab_size": 2,
|
| 84 |
"use_cache": true,
|
| 85 |
-
"vocab_size":
|
| 86 |
},
|
| 87 |
"eos_token_id": 0,
|
| 88 |
"is_encoder_decoder": true,
|
|
|
|
| 40 |
"tie_word_embeddings": true,
|
| 41 |
"type_vocab_size": 2,
|
| 42 |
"use_cache": true,
|
| 43 |
+
"vocab_size": 52
|
| 44 |
},
|
| 45 |
"decoder_start_token_id": 2,
|
| 46 |
"dtype": "float32",
|
|
|
|
| 82 |
"tie_word_embeddings": true,
|
| 83 |
"type_vocab_size": 2,
|
| 84 |
"use_cache": true,
|
| 85 |
+
"vocab_size": 52
|
| 86 |
},
|
| 87 |
"eos_token_id": 0,
|
| 88 |
"is_encoder_decoder": true,
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7d3010dfeb4a3be09fbb08f4001231c3d61258e71807430cba39ad57faa7a900
|
| 3 |
+
size 31205552
|
tokenizer.json
CHANGED
|
@@ -117,28 +117,28 @@
|
|
| 117 |
"99": 17,
|
| 118 |
"10": 18,
|
| 119 |
"98": 19,
|
| 120 |
-
"
|
| 121 |
-
"
|
| 122 |
-
"
|
| 123 |
"96": 23,
|
| 124 |
"13": 24,
|
| 125 |
"95": 25,
|
| 126 |
"14": 26,
|
| 127 |
-
"
|
| 128 |
-
"
|
| 129 |
-
"
|
| 130 |
-
"
|
| 131 |
"17": 31,
|
| 132 |
"92": 32,
|
| 133 |
"91": 33,
|
| 134 |
"18": 34,
|
| 135 |
-
"
|
| 136 |
-
"
|
| 137 |
-
"
|
| 138 |
-
"
|
| 139 |
-
"
|
| 140 |
-
"
|
| 141 |
-
"
|
| 142 |
"87": 42,
|
| 143 |
"86": 43,
|
| 144 |
"23": 44,
|
|
@@ -148,20 +148,7 @@
|
|
| 148 |
"25": 48,
|
| 149 |
"26": 49,
|
| 150 |
"83": 50,
|
| 151 |
-
"
|
| 152 |
-
"81": 52,
|
| 153 |
-
"82": 53,
|
| 154 |
-
"28": 54,
|
| 155 |
-
"80": 55,
|
| 156 |
-
"30": 56,
|
| 157 |
-
"29": 57,
|
| 158 |
-
"31": 58,
|
| 159 |
-
"79": 59,
|
| 160 |
-
"78": 60,
|
| 161 |
-
"76": 61,
|
| 162 |
-
"34": 62,
|
| 163 |
-
"77": 63,
|
| 164 |
-
"33": 64
|
| 165 |
},
|
| 166 |
"merges": [
|
| 167 |
[
|
|
@@ -176,10 +163,6 @@
|
|
| 176 |
"9",
|
| 177 |
"8"
|
| 178 |
],
|
| 179 |
-
[
|
| 180 |
-
"1",
|
| 181 |
-
"1"
|
| 182 |
-
],
|
| 183 |
[
|
| 184 |
"1",
|
| 185 |
"2"
|
|
@@ -188,6 +171,10 @@
|
|
| 188 |
"9",
|
| 189 |
"7"
|
| 190 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
| 191 |
[
|
| 192 |
"9",
|
| 193 |
"6"
|
|
@@ -205,20 +192,20 @@
|
|
| 205 |
"4"
|
| 206 |
],
|
| 207 |
[
|
| 208 |
-
"
|
| 209 |
-
"
|
| 210 |
],
|
| 211 |
[
|
| 212 |
"9",
|
| 213 |
-
"
|
| 214 |
],
|
| 215 |
[
|
| 216 |
"1",
|
| 217 |
-
"
|
| 218 |
],
|
| 219 |
[
|
| 220 |
-
"
|
| 221 |
-
"
|
| 222 |
],
|
| 223 |
[
|
| 224 |
"1",
|
|
@@ -236,26 +223,18 @@
|
|
| 236 |
"1",
|
| 237 |
"8"
|
| 238 |
],
|
| 239 |
-
[
|
| 240 |
-
"9",
|
| 241 |
-
"0"
|
| 242 |
-
],
|
| 243 |
[
|
| 244 |
"1",
|
| 245 |
"9"
|
| 246 |
],
|
| 247 |
[
|
| 248 |
-
"
|
| 249 |
-
"
|
| 250 |
],
|
| 251 |
[
|
| 252 |
"2",
|
| 253 |
"0"
|
| 254 |
],
|
| 255 |
-
[
|
| 256 |
-
"8",
|
| 257 |
-
"8"
|
| 258 |
-
],
|
| 259 |
[
|
| 260 |
"2",
|
| 261 |
"1"
|
|
@@ -264,6 +243,14 @@
|
|
| 264 |
"2",
|
| 265 |
"2"
|
| 266 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 267 |
[
|
| 268 |
"8",
|
| 269 |
"7"
|
|
@@ -300,61 +287,9 @@
|
|
| 300 |
"8",
|
| 301 |
"3"
|
| 302 |
],
|
| 303 |
-
[
|
| 304 |
-
"2",
|
| 305 |
-
"7"
|
| 306 |
-
],
|
| 307 |
-
[
|
| 308 |
-
"8",
|
| 309 |
-
"1"
|
| 310 |
-
],
|
| 311 |
[
|
| 312 |
"8",
|
| 313 |
"2"
|
| 314 |
-
],
|
| 315 |
-
[
|
| 316 |
-
"2",
|
| 317 |
-
"8"
|
| 318 |
-
],
|
| 319 |
-
[
|
| 320 |
-
"8",
|
| 321 |
-
"0"
|
| 322 |
-
],
|
| 323 |
-
[
|
| 324 |
-
"3",
|
| 325 |
-
"0"
|
| 326 |
-
],
|
| 327 |
-
[
|
| 328 |
-
"2",
|
| 329 |
-
"9"
|
| 330 |
-
],
|
| 331 |
-
[
|
| 332 |
-
"3",
|
| 333 |
-
"1"
|
| 334 |
-
],
|
| 335 |
-
[
|
| 336 |
-
"7",
|
| 337 |
-
"9"
|
| 338 |
-
],
|
| 339 |
-
[
|
| 340 |
-
"7",
|
| 341 |
-
"8"
|
| 342 |
-
],
|
| 343 |
-
[
|
| 344 |
-
"7",
|
| 345 |
-
"6"
|
| 346 |
-
],
|
| 347 |
-
[
|
| 348 |
-
"3",
|
| 349 |
-
"4"
|
| 350 |
-
],
|
| 351 |
-
[
|
| 352 |
-
"7",
|
| 353 |
-
"7"
|
| 354 |
-
],
|
| 355 |
-
[
|
| 356 |
-
"3",
|
| 357 |
-
"3"
|
| 358 |
]
|
| 359 |
]
|
| 360 |
}
|
|
|
|
| 117 |
"99": 17,
|
| 118 |
"10": 18,
|
| 119 |
"98": 19,
|
| 120 |
+
"12": 20,
|
| 121 |
+
"97": 21,
|
| 122 |
+
"11": 22,
|
| 123 |
"96": 23,
|
| 124 |
"13": 24,
|
| 125 |
"95": 25,
|
| 126 |
"14": 26,
|
| 127 |
+
"94": 27,
|
| 128 |
+
"93": 28,
|
| 129 |
+
"15": 29,
|
| 130 |
+
"16": 30,
|
| 131 |
"17": 31,
|
| 132 |
"92": 32,
|
| 133 |
"91": 33,
|
| 134 |
"18": 34,
|
| 135 |
+
"19": 35,
|
| 136 |
+
"90": 36,
|
| 137 |
+
"20": 37,
|
| 138 |
+
"21": 38,
|
| 139 |
+
"22": 39,
|
| 140 |
+
"88": 40,
|
| 141 |
+
"89": 41,
|
| 142 |
"87": 42,
|
| 143 |
"86": 43,
|
| 144 |
"23": 44,
|
|
|
|
| 148 |
"25": 48,
|
| 149 |
"26": 49,
|
| 150 |
"83": 50,
|
| 151 |
+
"82": 51
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 152 |
},
|
| 153 |
"merges": [
|
| 154 |
[
|
|
|
|
| 163 |
"9",
|
| 164 |
"8"
|
| 165 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
| 166 |
[
|
| 167 |
"1",
|
| 168 |
"2"
|
|
|
|
| 171 |
"9",
|
| 172 |
"7"
|
| 173 |
],
|
| 174 |
+
[
|
| 175 |
+
"1",
|
| 176 |
+
"1"
|
| 177 |
+
],
|
| 178 |
[
|
| 179 |
"9",
|
| 180 |
"6"
|
|
|
|
| 192 |
"4"
|
| 193 |
],
|
| 194 |
[
|
| 195 |
+
"9",
|
| 196 |
+
"4"
|
| 197 |
],
|
| 198 |
[
|
| 199 |
"9",
|
| 200 |
+
"3"
|
| 201 |
],
|
| 202 |
[
|
| 203 |
"1",
|
| 204 |
+
"5"
|
| 205 |
],
|
| 206 |
[
|
| 207 |
+
"1",
|
| 208 |
+
"6"
|
| 209 |
],
|
| 210 |
[
|
| 211 |
"1",
|
|
|
|
| 223 |
"1",
|
| 224 |
"8"
|
| 225 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
| 226 |
[
|
| 227 |
"1",
|
| 228 |
"9"
|
| 229 |
],
|
| 230 |
[
|
| 231 |
+
"9",
|
| 232 |
+
"0"
|
| 233 |
],
|
| 234 |
[
|
| 235 |
"2",
|
| 236 |
"0"
|
| 237 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
| 238 |
[
|
| 239 |
"2",
|
| 240 |
"1"
|
|
|
|
| 243 |
"2",
|
| 244 |
"2"
|
| 245 |
],
|
| 246 |
+
[
|
| 247 |
+
"8",
|
| 248 |
+
"8"
|
| 249 |
+
],
|
| 250 |
+
[
|
| 251 |
+
"8",
|
| 252 |
+
"9"
|
| 253 |
+
],
|
| 254 |
[
|
| 255 |
"8",
|
| 256 |
"7"
|
|
|
|
| 287 |
"8",
|
| 288 |
"3"
|
| 289 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 290 |
[
|
| 291 |
"8",
|
| 292 |
"2"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 293 |
]
|
| 294 |
]
|
| 295 |
}
|