Training in progress, step 240
Browse files- config.json +2 -2
- model.safetensors +2 -2
- tokenizer.json +58 -52
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": 53
|
| 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": 53
|
| 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 sha256:61ec67b9a56b8ed19135e570e501f636af7489b821975a3b1b349e03f73bcf84
|
| 3 |
+
size 31207604
|
tokenizer.json
CHANGED
|
@@ -113,51 +113,53 @@
|
|
| 113 |
"7": 13,
|
| 114 |
"8": 14,
|
| 115 |
"9": 15,
|
| 116 |
-
"
|
| 117 |
"99": 17,
|
| 118 |
-
"
|
| 119 |
-
"
|
| 120 |
-
"
|
| 121 |
-
"
|
| 122 |
-
"
|
| 123 |
"96": 23,
|
| 124 |
-
"
|
| 125 |
-
"
|
| 126 |
-
"
|
| 127 |
-
"
|
| 128 |
-
"
|
| 129 |
"93": 29,
|
| 130 |
-
"
|
| 131 |
"17": 31,
|
| 132 |
-
"
|
| 133 |
-
"
|
| 134 |
-
"
|
| 135 |
-
"
|
| 136 |
-
"
|
| 137 |
-
"
|
| 138 |
-
"
|
| 139 |
-
"
|
| 140 |
-
"
|
| 141 |
-
"
|
| 142 |
-
"
|
| 143 |
"86": 43,
|
| 144 |
-
"
|
| 145 |
"85": 45,
|
| 146 |
-
"
|
| 147 |
-
"
|
| 148 |
-
"
|
| 149 |
"26": 49,
|
| 150 |
-
"
|
|
|
|
|
|
|
| 151 |
},
|
| 152 |
"merges": [
|
| 153 |
-
[
|
| 154 |
-
"1",
|
| 155 |
-
"0"
|
| 156 |
-
],
|
| 157 |
[
|
| 158 |
"9",
|
| 159 |
"9"
|
| 160 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
| 161 |
[
|
| 162 |
"1",
|
| 163 |
"1"
|
|
@@ -174,14 +176,14 @@
|
|
| 174 |
"1",
|
| 175 |
"2"
|
| 176 |
],
|
| 177 |
-
[
|
| 178 |
-
"1",
|
| 179 |
-
"3"
|
| 180 |
-
],
|
| 181 |
[
|
| 182 |
"9",
|
| 183 |
"6"
|
| 184 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
| 185 |
[
|
| 186 |
"1",
|
| 187 |
"4"
|
|
@@ -198,22 +200,22 @@
|
|
| 198 |
"9",
|
| 199 |
"4"
|
| 200 |
],
|
| 201 |
-
[
|
| 202 |
-
"1",
|
| 203 |
-
"6"
|
| 204 |
-
],
|
| 205 |
[
|
| 206 |
"9",
|
| 207 |
"3"
|
| 208 |
],
|
| 209 |
[
|
| 210 |
-
"
|
| 211 |
-
"
|
| 212 |
],
|
| 213 |
[
|
| 214 |
"1",
|
| 215 |
"7"
|
| 216 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
| 217 |
[
|
| 218 |
"1",
|
| 219 |
"8"
|
|
@@ -234,6 +236,10 @@
|
|
| 234 |
"8",
|
| 235 |
"9"
|
| 236 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
| 237 |
[
|
| 238 |
"2",
|
| 239 |
"0"
|
|
@@ -242,10 +248,6 @@
|
|
| 242 |
"2",
|
| 243 |
"1"
|
| 244 |
],
|
| 245 |
-
[
|
| 246 |
-
"8",
|
| 247 |
-
"8"
|
| 248 |
-
],
|
| 249 |
[
|
| 250 |
"2",
|
| 251 |
"2"
|
|
@@ -254,22 +256,22 @@
|
|
| 254 |
"8",
|
| 255 |
"7"
|
| 256 |
],
|
| 257 |
-
[
|
| 258 |
-
"2",
|
| 259 |
-
"3"
|
| 260 |
-
],
|
| 261 |
[
|
| 262 |
"8",
|
| 263 |
"6"
|
| 264 |
],
|
| 265 |
[
|
| 266 |
"2",
|
| 267 |
-
"
|
| 268 |
],
|
| 269 |
[
|
| 270 |
"8",
|
| 271 |
"5"
|
| 272 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
| 273 |
[
|
| 274 |
"2",
|
| 275 |
"5"
|
|
@@ -278,13 +280,17 @@
|
|
| 278 |
"8",
|
| 279 |
"4"
|
| 280 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
| 281 |
[
|
| 282 |
"8",
|
| 283 |
"3"
|
| 284 |
],
|
| 285 |
[
|
| 286 |
"2",
|
| 287 |
-
"
|
| 288 |
],
|
| 289 |
[
|
| 290 |
"8",
|
|
|
|
| 113 |
"7": 13,
|
| 114 |
"8": 14,
|
| 115 |
"9": 15,
|
| 116 |
+
"=": 16,
|
| 117 |
"99": 17,
|
| 118 |
+
"10": 18,
|
| 119 |
+
"11": 19,
|
| 120 |
+
"98": 20,
|
| 121 |
+
"97": 21,
|
| 122 |
+
"12": 22,
|
| 123 |
"96": 23,
|
| 124 |
+
"13": 24,
|
| 125 |
+
"14": 25,
|
| 126 |
+
"95": 26,
|
| 127 |
+
"15": 27,
|
| 128 |
+
"94": 28,
|
| 129 |
"93": 29,
|
| 130 |
+
"16": 30,
|
| 131 |
"17": 31,
|
| 132 |
+
"92": 32,
|
| 133 |
+
"18": 33,
|
| 134 |
+
"91": 34,
|
| 135 |
+
"90": 35,
|
| 136 |
+
"19": 36,
|
| 137 |
+
"89": 37,
|
| 138 |
+
"88": 38,
|
| 139 |
+
"20": 39,
|
| 140 |
+
"21": 40,
|
| 141 |
+
"22": 41,
|
| 142 |
+
"87": 42,
|
| 143 |
"86": 43,
|
| 144 |
+
"23": 44,
|
| 145 |
"85": 45,
|
| 146 |
+
"24": 46,
|
| 147 |
+
"25": 47,
|
| 148 |
+
"84": 48,
|
| 149 |
"26": 49,
|
| 150 |
+
"83": 50,
|
| 151 |
+
"27": 51,
|
| 152 |
+
"82": 52
|
| 153 |
},
|
| 154 |
"merges": [
|
|
|
|
|
|
|
|
|
|
|
|
|
| 155 |
[
|
| 156 |
"9",
|
| 157 |
"9"
|
| 158 |
],
|
| 159 |
+
[
|
| 160 |
+
"1",
|
| 161 |
+
"0"
|
| 162 |
+
],
|
| 163 |
[
|
| 164 |
"1",
|
| 165 |
"1"
|
|
|
|
| 176 |
"1",
|
| 177 |
"2"
|
| 178 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
| 179 |
[
|
| 180 |
"9",
|
| 181 |
"6"
|
| 182 |
],
|
| 183 |
+
[
|
| 184 |
+
"1",
|
| 185 |
+
"3"
|
| 186 |
+
],
|
| 187 |
[
|
| 188 |
"1",
|
| 189 |
"4"
|
|
|
|
| 200 |
"9",
|
| 201 |
"4"
|
| 202 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
| 203 |
[
|
| 204 |
"9",
|
| 205 |
"3"
|
| 206 |
],
|
| 207 |
[
|
| 208 |
+
"1",
|
| 209 |
+
"6"
|
| 210 |
],
|
| 211 |
[
|
| 212 |
"1",
|
| 213 |
"7"
|
| 214 |
],
|
| 215 |
+
[
|
| 216 |
+
"9",
|
| 217 |
+
"2"
|
| 218 |
+
],
|
| 219 |
[
|
| 220 |
"1",
|
| 221 |
"8"
|
|
|
|
| 236 |
"8",
|
| 237 |
"9"
|
| 238 |
],
|
| 239 |
+
[
|
| 240 |
+
"8",
|
| 241 |
+
"8"
|
| 242 |
+
],
|
| 243 |
[
|
| 244 |
"2",
|
| 245 |
"0"
|
|
|
|
| 248 |
"2",
|
| 249 |
"1"
|
| 250 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
| 251 |
[
|
| 252 |
"2",
|
| 253 |
"2"
|
|
|
|
| 256 |
"8",
|
| 257 |
"7"
|
| 258 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
| 259 |
[
|
| 260 |
"8",
|
| 261 |
"6"
|
| 262 |
],
|
| 263 |
[
|
| 264 |
"2",
|
| 265 |
+
"3"
|
| 266 |
],
|
| 267 |
[
|
| 268 |
"8",
|
| 269 |
"5"
|
| 270 |
],
|
| 271 |
+
[
|
| 272 |
+
"2",
|
| 273 |
+
"4"
|
| 274 |
+
],
|
| 275 |
[
|
| 276 |
"2",
|
| 277 |
"5"
|
|
|
|
| 280 |
"8",
|
| 281 |
"4"
|
| 282 |
],
|
| 283 |
+
[
|
| 284 |
+
"2",
|
| 285 |
+
"6"
|
| 286 |
+
],
|
| 287 |
[
|
| 288 |
"8",
|
| 289 |
"3"
|
| 290 |
],
|
| 291 |
[
|
| 292 |
"2",
|
| 293 |
+
"7"
|
| 294 |
],
|
| 295 |
[
|
| 296 |
"8",
|