Training in progress, step 240
Browse files- config.json +2 -2
- model.safetensors +2 -2
- tokenizer.json +52 -51
config.json
CHANGED
|
@@ -10,7 +10,7 @@
|
|
| 10 |
"bos_token_id": null,
|
| 11 |
"chunk_size_feed_forward": 0,
|
| 12 |
"classifier_dropout": null,
|
| 13 |
-
"dtype":
|
| 14 |
"eos_token_id": null,
|
| 15 |
"hidden_act": "gelu",
|
| 16 |
"hidden_dropout_prob": 0.1,
|
|
@@ -52,7 +52,7 @@
|
|
| 52 |
"bos_token_id": null,
|
| 53 |
"chunk_size_feed_forward": 0,
|
| 54 |
"classifier_dropout": null,
|
| 55 |
-
"dtype":
|
| 56 |
"eos_token_id": null,
|
| 57 |
"hidden_act": "gelu",
|
| 58 |
"hidden_dropout_prob": 0.1,
|
|
|
|
| 10 |
"bos_token_id": null,
|
| 11 |
"chunk_size_feed_forward": 0,
|
| 12 |
"classifier_dropout": null,
|
| 13 |
+
"dtype": "float32",
|
| 14 |
"eos_token_id": null,
|
| 15 |
"hidden_act": "gelu",
|
| 16 |
"hidden_dropout_prob": 0.1,
|
|
|
|
| 52 |
"bos_token_id": null,
|
| 53 |
"chunk_size_feed_forward": 0,
|
| 54 |
"classifier_dropout": null,
|
| 55 |
+
"dtype": "float32",
|
| 56 |
"eos_token_id": null,
|
| 57 |
"hidden_act": "gelu",
|
| 58 |
"hidden_dropout_prob": 0.1,
|
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:d599564c1eb2bc9a67391ddf7fe1a5a60e590441524ecbde95177ceb515e4b1f
|
| 3 |
+
size 31205528
|
tokenizer.json
CHANGED
|
@@ -113,51 +113,56 @@
|
|
| 113 |
"7": 13,
|
| 114 |
"8": 14,
|
| 115 |
"9": 15,
|
| 116 |
-
"
|
| 117 |
"10": 17,
|
| 118 |
-
"
|
| 119 |
"98": 19,
|
| 120 |
-
"
|
| 121 |
"97": 21,
|
| 122 |
-
"
|
| 123 |
"96": 23,
|
| 124 |
-
"
|
| 125 |
-
"
|
| 126 |
-
"
|
| 127 |
-
"
|
| 128 |
-
"
|
| 129 |
-
"
|
| 130 |
-
"
|
| 131 |
"17": 31,
|
| 132 |
-
"
|
| 133 |
-
"
|
| 134 |
-
"
|
| 135 |
"90": 35,
|
| 136 |
-
"
|
| 137 |
-
"
|
| 138 |
-
"
|
| 139 |
-
"
|
| 140 |
-
"
|
| 141 |
-
"
|
| 142 |
-
"
|
| 143 |
"23": 43,
|
| 144 |
-
"
|
| 145 |
"85": 45,
|
| 146 |
-
"
|
| 147 |
-
"
|
| 148 |
-
"
|
| 149 |
-
"
|
| 150 |
-
"
|
| 151 |
-
"
|
|
|
|
| 152 |
},
|
| 153 |
"merges": [
|
|
|
|
|
|
|
|
|
|
|
|
|
| 154 |
[
|
| 155 |
"9",
|
| 156 |
"9"
|
| 157 |
],
|
| 158 |
[
|
| 159 |
-
"
|
| 160 |
-
"
|
| 161 |
],
|
| 162 |
[
|
| 163 |
"1",
|
|
@@ -165,7 +170,7 @@
|
|
| 165 |
],
|
| 166 |
[
|
| 167 |
"9",
|
| 168 |
-
"
|
| 169 |
],
|
| 170 |
[
|
| 171 |
"1",
|
|
@@ -173,7 +178,7 @@
|
|
| 173 |
],
|
| 174 |
[
|
| 175 |
"9",
|
| 176 |
-
"
|
| 177 |
],
|
| 178 |
[
|
| 179 |
"1",
|
|
@@ -181,7 +186,7 @@
|
|
| 181 |
],
|
| 182 |
[
|
| 183 |
"9",
|
| 184 |
-
"
|
| 185 |
],
|
| 186 |
[
|
| 187 |
"1",
|
|
@@ -191,10 +196,6 @@
|
|
| 191 |
"9",
|
| 192 |
"4"
|
| 193 |
],
|
| 194 |
-
[
|
| 195 |
-
"9",
|
| 196 |
-
"5"
|
| 197 |
-
],
|
| 198 |
[
|
| 199 |
"1",
|
| 200 |
"5"
|
|
@@ -207,14 +208,14 @@
|
|
| 207 |
"1",
|
| 208 |
"6"
|
| 209 |
],
|
| 210 |
-
[
|
| 211 |
-
"9",
|
| 212 |
-
"2"
|
| 213 |
-
],
|
| 214 |
[
|
| 215 |
"1",
|
| 216 |
"7"
|
| 217 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
| 218 |
[
|
| 219 |
"1",
|
| 220 |
"8"
|
|
@@ -223,14 +224,14 @@
|
|
| 223 |
"9",
|
| 224 |
"1"
|
| 225 |
],
|
| 226 |
-
[
|
| 227 |
-
"1",
|
| 228 |
-
"9"
|
| 229 |
-
],
|
| 230 |
[
|
| 231 |
"9",
|
| 232 |
"0"
|
| 233 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
| 234 |
[
|
| 235 |
"2",
|
| 236 |
"0"
|
|
@@ -255,22 +256,22 @@
|
|
| 255 |
"2",
|
| 256 |
"2"
|
| 257 |
],
|
| 258 |
-
[
|
| 259 |
-
"8",
|
| 260 |
-
"6"
|
| 261 |
-
],
|
| 262 |
[
|
| 263 |
"2",
|
| 264 |
"3"
|
| 265 |
],
|
| 266 |
[
|
| 267 |
-
"
|
| 268 |
-
"
|
| 269 |
],
|
| 270 |
[
|
| 271 |
"8",
|
| 272 |
"5"
|
| 273 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
| 274 |
[
|
| 275 |
"2",
|
| 276 |
"5"
|
|
|
|
| 113 |
"7": 13,
|
| 114 |
"8": 14,
|
| 115 |
"9": 15,
|
| 116 |
+
"=": 16,
|
| 117 |
"10": 17,
|
| 118 |
+
"99": 18,
|
| 119 |
"98": 19,
|
| 120 |
+
"11": 20,
|
| 121 |
"97": 21,
|
| 122 |
+
"12": 22,
|
| 123 |
"96": 23,
|
| 124 |
+
"13": 24,
|
| 125 |
+
"95": 25,
|
| 126 |
+
"14": 26,
|
| 127 |
+
"94": 27,
|
| 128 |
+
"15": 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 |
+
"20": 37,
|
| 138 |
+
"89": 38,
|
| 139 |
+
"21": 39,
|
| 140 |
+
"88": 40,
|
| 141 |
+
"87": 41,
|
| 142 |
+
"22": 42,
|
| 143 |
"23": 43,
|
| 144 |
+
"86": 44,
|
| 145 |
"85": 45,
|
| 146 |
+
"24": 46,
|
| 147 |
+
"25": 47,
|
| 148 |
+
"84": 48,
|
| 149 |
+
"83": 49,
|
| 150 |
+
"26": 50,
|
| 151 |
+
"27": 51,
|
| 152 |
+
"82": 52
|
| 153 |
},
|
| 154 |
"merges": [
|
| 155 |
+
[
|
| 156 |
+
"1",
|
| 157 |
+
"0"
|
| 158 |
+
],
|
| 159 |
[
|
| 160 |
"9",
|
| 161 |
"9"
|
| 162 |
],
|
| 163 |
[
|
| 164 |
+
"9",
|
| 165 |
+
"8"
|
| 166 |
],
|
| 167 |
[
|
| 168 |
"1",
|
|
|
|
| 170 |
],
|
| 171 |
[
|
| 172 |
"9",
|
| 173 |
+
"7"
|
| 174 |
],
|
| 175 |
[
|
| 176 |
"1",
|
|
|
|
| 178 |
],
|
| 179 |
[
|
| 180 |
"9",
|
| 181 |
+
"6"
|
| 182 |
],
|
| 183 |
[
|
| 184 |
"1",
|
|
|
|
| 186 |
],
|
| 187 |
[
|
| 188 |
"9",
|
| 189 |
+
"5"
|
| 190 |
],
|
| 191 |
[
|
| 192 |
"1",
|
|
|
|
| 196 |
"9",
|
| 197 |
"4"
|
| 198 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
| 199 |
[
|
| 200 |
"1",
|
| 201 |
"5"
|
|
|
|
| 208 |
"1",
|
| 209 |
"6"
|
| 210 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
| 211 |
[
|
| 212 |
"1",
|
| 213 |
"7"
|
| 214 |
],
|
| 215 |
+
[
|
| 216 |
+
"9",
|
| 217 |
+
"2"
|
| 218 |
+
],
|
| 219 |
[
|
| 220 |
"1",
|
| 221 |
"8"
|
|
|
|
| 224 |
"9",
|
| 225 |
"1"
|
| 226 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
| 227 |
[
|
| 228 |
"9",
|
| 229 |
"0"
|
| 230 |
],
|
| 231 |
+
[
|
| 232 |
+
"1",
|
| 233 |
+
"9"
|
| 234 |
+
],
|
| 235 |
[
|
| 236 |
"2",
|
| 237 |
"0"
|
|
|
|
| 256 |
"2",
|
| 257 |
"2"
|
| 258 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
| 259 |
[
|
| 260 |
"2",
|
| 261 |
"3"
|
| 262 |
],
|
| 263 |
[
|
| 264 |
+
"8",
|
| 265 |
+
"6"
|
| 266 |
],
|
| 267 |
[
|
| 268 |
"8",
|
| 269 |
"5"
|
| 270 |
],
|
| 271 |
+
[
|
| 272 |
+
"2",
|
| 273 |
+
"4"
|
| 274 |
+
],
|
| 275 |
[
|
| 276 |
"2",
|
| 277 |
"5"
|