Upload 6 files
Browse files- config (3).json +305 -0
- rng_state (2).pth +3 -0
- scheduler (2).pt +3 -0
- trainer_state (2).json +1249 -0
- training_args (2).bin +3 -0
config (3).json
ADDED
|
@@ -0,0 +1,305 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation": "gelu",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"DistilBertForSequenceClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_dropout": 0.1,
|
| 7 |
+
"dim": 768,
|
| 8 |
+
"dropout": 0.1,
|
| 9 |
+
"hidden_dim": 3072,
|
| 10 |
+
"id2label": {
|
| 11 |
+
"0": "LABEL_0",
|
| 12 |
+
"1": "LABEL_1",
|
| 13 |
+
"2": "LABEL_2",
|
| 14 |
+
"3": "LABEL_3",
|
| 15 |
+
"4": "LABEL_4",
|
| 16 |
+
"5": "LABEL_5",
|
| 17 |
+
"6": "LABEL_6",
|
| 18 |
+
"7": "LABEL_7",
|
| 19 |
+
"8": "LABEL_8",
|
| 20 |
+
"9": "LABEL_9",
|
| 21 |
+
"10": "LABEL_10",
|
| 22 |
+
"11": "LABEL_11",
|
| 23 |
+
"12": "LABEL_12",
|
| 24 |
+
"13": "LABEL_13",
|
| 25 |
+
"14": "LABEL_14",
|
| 26 |
+
"15": "LABEL_15",
|
| 27 |
+
"16": "LABEL_16",
|
| 28 |
+
"17": "LABEL_17",
|
| 29 |
+
"18": "LABEL_18",
|
| 30 |
+
"19": "LABEL_19",
|
| 31 |
+
"20": "LABEL_20",
|
| 32 |
+
"21": "LABEL_21",
|
| 33 |
+
"22": "LABEL_22",
|
| 34 |
+
"23": "LABEL_23",
|
| 35 |
+
"24": "LABEL_24",
|
| 36 |
+
"25": "LABEL_25",
|
| 37 |
+
"26": "LABEL_26",
|
| 38 |
+
"27": "LABEL_27",
|
| 39 |
+
"28": "LABEL_28",
|
| 40 |
+
"29": "LABEL_29",
|
| 41 |
+
"30": "LABEL_30",
|
| 42 |
+
"31": "LABEL_31",
|
| 43 |
+
"32": "LABEL_32",
|
| 44 |
+
"33": "LABEL_33",
|
| 45 |
+
"34": "LABEL_34",
|
| 46 |
+
"35": "LABEL_35",
|
| 47 |
+
"36": "LABEL_36",
|
| 48 |
+
"37": "LABEL_37",
|
| 49 |
+
"38": "LABEL_38",
|
| 50 |
+
"39": "LABEL_39",
|
| 51 |
+
"40": "LABEL_40",
|
| 52 |
+
"41": "LABEL_41",
|
| 53 |
+
"42": "LABEL_42",
|
| 54 |
+
"43": "LABEL_43",
|
| 55 |
+
"44": "LABEL_44",
|
| 56 |
+
"45": "LABEL_45",
|
| 57 |
+
"46": "LABEL_46",
|
| 58 |
+
"47": "LABEL_47",
|
| 59 |
+
"48": "LABEL_48",
|
| 60 |
+
"49": "LABEL_49",
|
| 61 |
+
"50": "LABEL_50",
|
| 62 |
+
"51": "LABEL_51",
|
| 63 |
+
"52": "LABEL_52",
|
| 64 |
+
"53": "LABEL_53",
|
| 65 |
+
"54": "LABEL_54",
|
| 66 |
+
"55": "LABEL_55",
|
| 67 |
+
"56": "LABEL_56",
|
| 68 |
+
"57": "LABEL_57",
|
| 69 |
+
"58": "LABEL_58",
|
| 70 |
+
"59": "LABEL_59",
|
| 71 |
+
"60": "LABEL_60",
|
| 72 |
+
"61": "LABEL_61",
|
| 73 |
+
"62": "LABEL_62",
|
| 74 |
+
"63": "LABEL_63",
|
| 75 |
+
"64": "LABEL_64",
|
| 76 |
+
"65": "LABEL_65",
|
| 77 |
+
"66": "LABEL_66",
|
| 78 |
+
"67": "LABEL_67",
|
| 79 |
+
"68": "LABEL_68",
|
| 80 |
+
"69": "LABEL_69",
|
| 81 |
+
"70": "LABEL_70",
|
| 82 |
+
"71": "LABEL_71",
|
| 83 |
+
"72": "LABEL_72",
|
| 84 |
+
"73": "LABEL_73",
|
| 85 |
+
"74": "LABEL_74",
|
| 86 |
+
"75": "LABEL_75",
|
| 87 |
+
"76": "LABEL_76",
|
| 88 |
+
"77": "LABEL_77",
|
| 89 |
+
"78": "LABEL_78",
|
| 90 |
+
"79": "LABEL_79",
|
| 91 |
+
"80": "LABEL_80",
|
| 92 |
+
"81": "LABEL_81",
|
| 93 |
+
"82": "LABEL_82",
|
| 94 |
+
"83": "LABEL_83",
|
| 95 |
+
"84": "LABEL_84",
|
| 96 |
+
"85": "LABEL_85",
|
| 97 |
+
"86": "LABEL_86",
|
| 98 |
+
"87": "LABEL_87",
|
| 99 |
+
"88": "LABEL_88",
|
| 100 |
+
"89": "LABEL_89",
|
| 101 |
+
"90": "LABEL_90",
|
| 102 |
+
"91": "LABEL_91",
|
| 103 |
+
"92": "LABEL_92",
|
| 104 |
+
"93": "LABEL_93",
|
| 105 |
+
"94": "LABEL_94",
|
| 106 |
+
"95": "LABEL_95",
|
| 107 |
+
"96": "LABEL_96",
|
| 108 |
+
"97": "LABEL_97",
|
| 109 |
+
"98": "LABEL_98",
|
| 110 |
+
"99": "LABEL_99",
|
| 111 |
+
"100": "LABEL_100",
|
| 112 |
+
"101": "LABEL_101",
|
| 113 |
+
"102": "LABEL_102",
|
| 114 |
+
"103": "LABEL_103",
|
| 115 |
+
"104": "LABEL_104",
|
| 116 |
+
"105": "LABEL_105",
|
| 117 |
+
"106": "LABEL_106",
|
| 118 |
+
"107": "LABEL_107",
|
| 119 |
+
"108": "LABEL_108",
|
| 120 |
+
"109": "LABEL_109",
|
| 121 |
+
"110": "LABEL_110",
|
| 122 |
+
"111": "LABEL_111",
|
| 123 |
+
"112": "LABEL_112",
|
| 124 |
+
"113": "LABEL_113",
|
| 125 |
+
"114": "LABEL_114",
|
| 126 |
+
"115": "LABEL_115",
|
| 127 |
+
"116": "LABEL_116",
|
| 128 |
+
"117": "LABEL_117",
|
| 129 |
+
"118": "LABEL_118",
|
| 130 |
+
"119": "LABEL_119",
|
| 131 |
+
"120": "LABEL_120",
|
| 132 |
+
"121": "LABEL_121",
|
| 133 |
+
"122": "LABEL_122",
|
| 134 |
+
"123": "LABEL_123",
|
| 135 |
+
"124": "LABEL_124",
|
| 136 |
+
"125": "LABEL_125",
|
| 137 |
+
"126": "LABEL_126",
|
| 138 |
+
"127": "LABEL_127",
|
| 139 |
+
"128": "LABEL_128",
|
| 140 |
+
"129": "LABEL_129",
|
| 141 |
+
"130": "LABEL_130",
|
| 142 |
+
"131": "LABEL_131",
|
| 143 |
+
"132": "LABEL_132",
|
| 144 |
+
"133": "LABEL_133",
|
| 145 |
+
"134": "LABEL_134",
|
| 146 |
+
"135": "LABEL_135",
|
| 147 |
+
"136": "LABEL_136",
|
| 148 |
+
"137": "LABEL_137"
|
| 149 |
+
},
|
| 150 |
+
"initializer_range": 0.02,
|
| 151 |
+
"label2id": {
|
| 152 |
+
"LABEL_0": 0,
|
| 153 |
+
"LABEL_1": 1,
|
| 154 |
+
"LABEL_10": 10,
|
| 155 |
+
"LABEL_100": 100,
|
| 156 |
+
"LABEL_101": 101,
|
| 157 |
+
"LABEL_102": 102,
|
| 158 |
+
"LABEL_103": 103,
|
| 159 |
+
"LABEL_104": 104,
|
| 160 |
+
"LABEL_105": 105,
|
| 161 |
+
"LABEL_106": 106,
|
| 162 |
+
"LABEL_107": 107,
|
| 163 |
+
"LABEL_108": 108,
|
| 164 |
+
"LABEL_109": 109,
|
| 165 |
+
"LABEL_11": 11,
|
| 166 |
+
"LABEL_110": 110,
|
| 167 |
+
"LABEL_111": 111,
|
| 168 |
+
"LABEL_112": 112,
|
| 169 |
+
"LABEL_113": 113,
|
| 170 |
+
"LABEL_114": 114,
|
| 171 |
+
"LABEL_115": 115,
|
| 172 |
+
"LABEL_116": 116,
|
| 173 |
+
"LABEL_117": 117,
|
| 174 |
+
"LABEL_118": 118,
|
| 175 |
+
"LABEL_119": 119,
|
| 176 |
+
"LABEL_12": 12,
|
| 177 |
+
"LABEL_120": 120,
|
| 178 |
+
"LABEL_121": 121,
|
| 179 |
+
"LABEL_122": 122,
|
| 180 |
+
"LABEL_123": 123,
|
| 181 |
+
"LABEL_124": 124,
|
| 182 |
+
"LABEL_125": 125,
|
| 183 |
+
"LABEL_126": 126,
|
| 184 |
+
"LABEL_127": 127,
|
| 185 |
+
"LABEL_128": 128,
|
| 186 |
+
"LABEL_129": 129,
|
| 187 |
+
"LABEL_13": 13,
|
| 188 |
+
"LABEL_130": 130,
|
| 189 |
+
"LABEL_131": 131,
|
| 190 |
+
"LABEL_132": 132,
|
| 191 |
+
"LABEL_133": 133,
|
| 192 |
+
"LABEL_134": 134,
|
| 193 |
+
"LABEL_135": 135,
|
| 194 |
+
"LABEL_136": 136,
|
| 195 |
+
"LABEL_137": 137,
|
| 196 |
+
"LABEL_14": 14,
|
| 197 |
+
"LABEL_15": 15,
|
| 198 |
+
"LABEL_16": 16,
|
| 199 |
+
"LABEL_17": 17,
|
| 200 |
+
"LABEL_18": 18,
|
| 201 |
+
"LABEL_19": 19,
|
| 202 |
+
"LABEL_2": 2,
|
| 203 |
+
"LABEL_20": 20,
|
| 204 |
+
"LABEL_21": 21,
|
| 205 |
+
"LABEL_22": 22,
|
| 206 |
+
"LABEL_23": 23,
|
| 207 |
+
"LABEL_24": 24,
|
| 208 |
+
"LABEL_25": 25,
|
| 209 |
+
"LABEL_26": 26,
|
| 210 |
+
"LABEL_27": 27,
|
| 211 |
+
"LABEL_28": 28,
|
| 212 |
+
"LABEL_29": 29,
|
| 213 |
+
"LABEL_3": 3,
|
| 214 |
+
"LABEL_30": 30,
|
| 215 |
+
"LABEL_31": 31,
|
| 216 |
+
"LABEL_32": 32,
|
| 217 |
+
"LABEL_33": 33,
|
| 218 |
+
"LABEL_34": 34,
|
| 219 |
+
"LABEL_35": 35,
|
| 220 |
+
"LABEL_36": 36,
|
| 221 |
+
"LABEL_37": 37,
|
| 222 |
+
"LABEL_38": 38,
|
| 223 |
+
"LABEL_39": 39,
|
| 224 |
+
"LABEL_4": 4,
|
| 225 |
+
"LABEL_40": 40,
|
| 226 |
+
"LABEL_41": 41,
|
| 227 |
+
"LABEL_42": 42,
|
| 228 |
+
"LABEL_43": 43,
|
| 229 |
+
"LABEL_44": 44,
|
| 230 |
+
"LABEL_45": 45,
|
| 231 |
+
"LABEL_46": 46,
|
| 232 |
+
"LABEL_47": 47,
|
| 233 |
+
"LABEL_48": 48,
|
| 234 |
+
"LABEL_49": 49,
|
| 235 |
+
"LABEL_5": 5,
|
| 236 |
+
"LABEL_50": 50,
|
| 237 |
+
"LABEL_51": 51,
|
| 238 |
+
"LABEL_52": 52,
|
| 239 |
+
"LABEL_53": 53,
|
| 240 |
+
"LABEL_54": 54,
|
| 241 |
+
"LABEL_55": 55,
|
| 242 |
+
"LABEL_56": 56,
|
| 243 |
+
"LABEL_57": 57,
|
| 244 |
+
"LABEL_58": 58,
|
| 245 |
+
"LABEL_59": 59,
|
| 246 |
+
"LABEL_6": 6,
|
| 247 |
+
"LABEL_60": 60,
|
| 248 |
+
"LABEL_61": 61,
|
| 249 |
+
"LABEL_62": 62,
|
| 250 |
+
"LABEL_63": 63,
|
| 251 |
+
"LABEL_64": 64,
|
| 252 |
+
"LABEL_65": 65,
|
| 253 |
+
"LABEL_66": 66,
|
| 254 |
+
"LABEL_67": 67,
|
| 255 |
+
"LABEL_68": 68,
|
| 256 |
+
"LABEL_69": 69,
|
| 257 |
+
"LABEL_7": 7,
|
| 258 |
+
"LABEL_70": 70,
|
| 259 |
+
"LABEL_71": 71,
|
| 260 |
+
"LABEL_72": 72,
|
| 261 |
+
"LABEL_73": 73,
|
| 262 |
+
"LABEL_74": 74,
|
| 263 |
+
"LABEL_75": 75,
|
| 264 |
+
"LABEL_76": 76,
|
| 265 |
+
"LABEL_77": 77,
|
| 266 |
+
"LABEL_78": 78,
|
| 267 |
+
"LABEL_79": 79,
|
| 268 |
+
"LABEL_8": 8,
|
| 269 |
+
"LABEL_80": 80,
|
| 270 |
+
"LABEL_81": 81,
|
| 271 |
+
"LABEL_82": 82,
|
| 272 |
+
"LABEL_83": 83,
|
| 273 |
+
"LABEL_84": 84,
|
| 274 |
+
"LABEL_85": 85,
|
| 275 |
+
"LABEL_86": 86,
|
| 276 |
+
"LABEL_87": 87,
|
| 277 |
+
"LABEL_88": 88,
|
| 278 |
+
"LABEL_89": 89,
|
| 279 |
+
"LABEL_9": 9,
|
| 280 |
+
"LABEL_90": 90,
|
| 281 |
+
"LABEL_91": 91,
|
| 282 |
+
"LABEL_92": 92,
|
| 283 |
+
"LABEL_93": 93,
|
| 284 |
+
"LABEL_94": 94,
|
| 285 |
+
"LABEL_95": 95,
|
| 286 |
+
"LABEL_96": 96,
|
| 287 |
+
"LABEL_97": 97,
|
| 288 |
+
"LABEL_98": 98,
|
| 289 |
+
"LABEL_99": 99
|
| 290 |
+
},
|
| 291 |
+
"max_position_embeddings": 512,
|
| 292 |
+
"model_type": "distilbert",
|
| 293 |
+
"n_heads": 12,
|
| 294 |
+
"n_layers": 6,
|
| 295 |
+
"output_past": true,
|
| 296 |
+
"pad_token_id": 0,
|
| 297 |
+
"problem_type": "multi_label_classification",
|
| 298 |
+
"qa_dropout": 0.1,
|
| 299 |
+
"seq_classif_dropout": 0.2,
|
| 300 |
+
"sinusoidal_pos_embds": false,
|
| 301 |
+
"tie_weights_": true,
|
| 302 |
+
"torch_dtype": "float32",
|
| 303 |
+
"transformers_version": "4.50.2",
|
| 304 |
+
"vocab_size": 28996
|
| 305 |
+
}
|
rng_state (2).pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:32761ad4532bb5e6091ca4c6c1b605035980b61e99bcfd3660390b24be2028c6
|
| 3 |
+
size 14244
|
scheduler (2).pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eb3a8012fb908180af318879d48681d0dc1ae007cb97e4885f1fa39e29882847
|
| 3 |
+
size 1064
|
trainer_state (2).json
ADDED
|
@@ -0,0 +1,1249 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1500,
|
| 3 |
+
"best_metric": 0.6884484673104748,
|
| 4 |
+
"best_model_checkpoint": "./results_2/checkpoint-1500",
|
| 5 |
+
"epoch": 0.7317073170731707,
|
| 6 |
+
"eval_steps": 100,
|
| 7 |
+
"global_step": 1500,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.004878048780487805,
|
| 14 |
+
"grad_norm": 0.09305039048194885,
|
| 15 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 16 |
+
"loss": 0.0173,
|
| 17 |
+
"step": 10
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.00975609756097561,
|
| 21 |
+
"grad_norm": 0.09173406660556793,
|
| 22 |
+
"learning_rate": 2.0000000000000003e-06,
|
| 23 |
+
"loss": 0.0167,
|
| 24 |
+
"step": 20
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.014634146341463415,
|
| 28 |
+
"grad_norm": 0.09779594093561172,
|
| 29 |
+
"learning_rate": 3e-06,
|
| 30 |
+
"loss": 0.0165,
|
| 31 |
+
"step": 30
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.01951219512195122,
|
| 35 |
+
"grad_norm": 0.1044146940112114,
|
| 36 |
+
"learning_rate": 4.000000000000001e-06,
|
| 37 |
+
"loss": 0.0144,
|
| 38 |
+
"step": 40
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.024390243902439025,
|
| 42 |
+
"grad_norm": 0.09117760509252548,
|
| 43 |
+
"learning_rate": 5e-06,
|
| 44 |
+
"loss": 0.0156,
|
| 45 |
+
"step": 50
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.02926829268292683,
|
| 49 |
+
"grad_norm": 0.08018726110458374,
|
| 50 |
+
"learning_rate": 6e-06,
|
| 51 |
+
"loss": 0.0149,
|
| 52 |
+
"step": 60
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.03414634146341464,
|
| 56 |
+
"grad_norm": 0.1090320348739624,
|
| 57 |
+
"learning_rate": 7.000000000000001e-06,
|
| 58 |
+
"loss": 0.0152,
|
| 59 |
+
"step": 70
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.03902439024390244,
|
| 63 |
+
"grad_norm": 0.09608811885118484,
|
| 64 |
+
"learning_rate": 8.000000000000001e-06,
|
| 65 |
+
"loss": 0.017,
|
| 66 |
+
"step": 80
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.04390243902439024,
|
| 70 |
+
"grad_norm": 0.08458913117647171,
|
| 71 |
+
"learning_rate": 9e-06,
|
| 72 |
+
"loss": 0.0167,
|
| 73 |
+
"step": 90
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.04878048780487805,
|
| 77 |
+
"grad_norm": 0.07629529386758804,
|
| 78 |
+
"learning_rate": 1e-05,
|
| 79 |
+
"loss": 0.0153,
|
| 80 |
+
"step": 100
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.04878048780487805,
|
| 84 |
+
"eval_f1": 0.6863881803425316,
|
| 85 |
+
"eval_loss": 0.018864748999476433,
|
| 86 |
+
"eval_precision": 0.6192098499790808,
|
| 87 |
+
"eval_recall": 0.7699167657550535,
|
| 88 |
+
"eval_runtime": 130.1056,
|
| 89 |
+
"eval_samples_per_second": 63.026,
|
| 90 |
+
"eval_steps_per_second": 0.992,
|
| 91 |
+
"step": 100
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 0.05365853658536585,
|
| 95 |
+
"grad_norm": 0.09111520648002625,
|
| 96 |
+
"learning_rate": 1.1000000000000001e-05,
|
| 97 |
+
"loss": 0.0164,
|
| 98 |
+
"step": 110
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.05853658536585366,
|
| 102 |
+
"grad_norm": 0.09741566330194473,
|
| 103 |
+
"learning_rate": 1.2e-05,
|
| 104 |
+
"loss": 0.0134,
|
| 105 |
+
"step": 120
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.06341463414634146,
|
| 109 |
+
"grad_norm": 0.07199473679065704,
|
| 110 |
+
"learning_rate": 1.3000000000000001e-05,
|
| 111 |
+
"loss": 0.0154,
|
| 112 |
+
"step": 130
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 0.06829268292682927,
|
| 116 |
+
"grad_norm": 0.08497001975774765,
|
| 117 |
+
"learning_rate": 1.4000000000000001e-05,
|
| 118 |
+
"loss": 0.0161,
|
| 119 |
+
"step": 140
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 0.07317073170731707,
|
| 123 |
+
"grad_norm": 0.08193599432706833,
|
| 124 |
+
"learning_rate": 1.5e-05,
|
| 125 |
+
"loss": 0.0147,
|
| 126 |
+
"step": 150
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 0.07804878048780488,
|
| 130 |
+
"grad_norm": 0.10380581766366959,
|
| 131 |
+
"learning_rate": 1.6000000000000003e-05,
|
| 132 |
+
"loss": 0.018,
|
| 133 |
+
"step": 160
|
| 134 |
+
},
|
| 135 |
+
{
|
| 136 |
+
"epoch": 0.08292682926829269,
|
| 137 |
+
"grad_norm": 0.11271784454584122,
|
| 138 |
+
"learning_rate": 1.7000000000000003e-05,
|
| 139 |
+
"loss": 0.0158,
|
| 140 |
+
"step": 170
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.08780487804878048,
|
| 144 |
+
"grad_norm": 0.0855347067117691,
|
| 145 |
+
"learning_rate": 1.8e-05,
|
| 146 |
+
"loss": 0.0166,
|
| 147 |
+
"step": 180
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.09268292682926829,
|
| 151 |
+
"grad_norm": 0.09468505531549454,
|
| 152 |
+
"learning_rate": 1.9e-05,
|
| 153 |
+
"loss": 0.0138,
|
| 154 |
+
"step": 190
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 0.0975609756097561,
|
| 158 |
+
"grad_norm": 0.06498368084430695,
|
| 159 |
+
"learning_rate": 2e-05,
|
| 160 |
+
"loss": 0.0138,
|
| 161 |
+
"step": 200
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 0.0975609756097561,
|
| 165 |
+
"eval_f1": 0.6788566213617548,
|
| 166 |
+
"eval_loss": 0.019105251878499985,
|
| 167 |
+
"eval_precision": 0.6045863570391873,
|
| 168 |
+
"eval_recall": 0.7739298454221165,
|
| 169 |
+
"eval_runtime": 126.9518,
|
| 170 |
+
"eval_samples_per_second": 64.591,
|
| 171 |
+
"eval_steps_per_second": 1.016,
|
| 172 |
+
"step": 200
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 0.1024390243902439,
|
| 176 |
+
"grad_norm": 0.08939243853092194,
|
| 177 |
+
"learning_rate": 2.1e-05,
|
| 178 |
+
"loss": 0.0177,
|
| 179 |
+
"step": 210
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 0.1073170731707317,
|
| 183 |
+
"grad_norm": 0.0708846002817154,
|
| 184 |
+
"learning_rate": 2.2000000000000003e-05,
|
| 185 |
+
"loss": 0.0155,
|
| 186 |
+
"step": 220
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 0.11219512195121951,
|
| 190 |
+
"grad_norm": 0.10868319869041443,
|
| 191 |
+
"learning_rate": 2.3000000000000003e-05,
|
| 192 |
+
"loss": 0.0172,
|
| 193 |
+
"step": 230
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 0.11707317073170732,
|
| 197 |
+
"grad_norm": 0.06580492854118347,
|
| 198 |
+
"learning_rate": 2.4e-05,
|
| 199 |
+
"loss": 0.0142,
|
| 200 |
+
"step": 240
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.12195121951219512,
|
| 204 |
+
"grad_norm": 0.08139664679765701,
|
| 205 |
+
"learning_rate": 2.5e-05,
|
| 206 |
+
"loss": 0.0181,
|
| 207 |
+
"step": 250
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.12682926829268293,
|
| 211 |
+
"grad_norm": 0.07353170961141586,
|
| 212 |
+
"learning_rate": 2.6000000000000002e-05,
|
| 213 |
+
"loss": 0.0149,
|
| 214 |
+
"step": 260
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"epoch": 0.13170731707317074,
|
| 218 |
+
"grad_norm": 0.09713993221521378,
|
| 219 |
+
"learning_rate": 2.7000000000000002e-05,
|
| 220 |
+
"loss": 0.0133,
|
| 221 |
+
"step": 270
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 0.13658536585365855,
|
| 225 |
+
"grad_norm": 0.0805194154381752,
|
| 226 |
+
"learning_rate": 2.8000000000000003e-05,
|
| 227 |
+
"loss": 0.0187,
|
| 228 |
+
"step": 280
|
| 229 |
+
},
|
| 230 |
+
{
|
| 231 |
+
"epoch": 0.14146341463414633,
|
| 232 |
+
"grad_norm": 0.11184240132570267,
|
| 233 |
+
"learning_rate": 2.9e-05,
|
| 234 |
+
"loss": 0.0169,
|
| 235 |
+
"step": 290
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 0.14634146341463414,
|
| 239 |
+
"grad_norm": 0.11055561900138855,
|
| 240 |
+
"learning_rate": 3e-05,
|
| 241 |
+
"loss": 0.0179,
|
| 242 |
+
"step": 300
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.14634146341463414,
|
| 246 |
+
"eval_f1": 0.6787136697277903,
|
| 247 |
+
"eval_loss": 0.019281357526779175,
|
| 248 |
+
"eval_precision": 0.6091580502215658,
|
| 249 |
+
"eval_recall": 0.7662009512485137,
|
| 250 |
+
"eval_runtime": 126.6013,
|
| 251 |
+
"eval_samples_per_second": 64.77,
|
| 252 |
+
"eval_steps_per_second": 1.019,
|
| 253 |
+
"step": 300
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 0.15121951219512195,
|
| 257 |
+
"grad_norm": 0.07896488904953003,
|
| 258 |
+
"learning_rate": 3.1e-05,
|
| 259 |
+
"loss": 0.0152,
|
| 260 |
+
"step": 310
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.15609756097560976,
|
| 264 |
+
"grad_norm": 0.10503584146499634,
|
| 265 |
+
"learning_rate": 3.2000000000000005e-05,
|
| 266 |
+
"loss": 0.0181,
|
| 267 |
+
"step": 320
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.16097560975609757,
|
| 271 |
+
"grad_norm": 0.10049130022525787,
|
| 272 |
+
"learning_rate": 3.3e-05,
|
| 273 |
+
"loss": 0.0169,
|
| 274 |
+
"step": 330
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 0.16585365853658537,
|
| 278 |
+
"grad_norm": 0.10558082908391953,
|
| 279 |
+
"learning_rate": 3.4000000000000007e-05,
|
| 280 |
+
"loss": 0.0158,
|
| 281 |
+
"step": 340
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 0.17073170731707318,
|
| 285 |
+
"grad_norm": 0.11450464278459549,
|
| 286 |
+
"learning_rate": 3.5e-05,
|
| 287 |
+
"loss": 0.0168,
|
| 288 |
+
"step": 350
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 0.17560975609756097,
|
| 292 |
+
"grad_norm": 0.15014208853244781,
|
| 293 |
+
"learning_rate": 3.6e-05,
|
| 294 |
+
"loss": 0.0145,
|
| 295 |
+
"step": 360
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 0.18048780487804877,
|
| 299 |
+
"grad_norm": 0.06404370069503784,
|
| 300 |
+
"learning_rate": 3.7e-05,
|
| 301 |
+
"loss": 0.0143,
|
| 302 |
+
"step": 370
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.18536585365853658,
|
| 306 |
+
"grad_norm": 0.09269659221172333,
|
| 307 |
+
"learning_rate": 3.8e-05,
|
| 308 |
+
"loss": 0.018,
|
| 309 |
+
"step": 380
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.1902439024390244,
|
| 313 |
+
"grad_norm": 0.09888678789138794,
|
| 314 |
+
"learning_rate": 3.9000000000000006e-05,
|
| 315 |
+
"loss": 0.0142,
|
| 316 |
+
"step": 390
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 0.1951219512195122,
|
| 320 |
+
"grad_norm": 0.16841690242290497,
|
| 321 |
+
"learning_rate": 4e-05,
|
| 322 |
+
"loss": 0.0207,
|
| 323 |
+
"step": 400
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 0.1951219512195122,
|
| 327 |
+
"eval_f1": 0.670402918255679,
|
| 328 |
+
"eval_loss": 0.02026216685771942,
|
| 329 |
+
"eval_precision": 0.6053057069285586,
|
| 330 |
+
"eval_recall": 0.7511890606420928,
|
| 331 |
+
"eval_runtime": 125.682,
|
| 332 |
+
"eval_samples_per_second": 65.244,
|
| 333 |
+
"eval_steps_per_second": 1.026,
|
| 334 |
+
"step": 400
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 0.2,
|
| 338 |
+
"grad_norm": 0.10583117604255676,
|
| 339 |
+
"learning_rate": 4.1e-05,
|
| 340 |
+
"loss": 0.0161,
|
| 341 |
+
"step": 410
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 0.2048780487804878,
|
| 345 |
+
"grad_norm": 0.08883402496576309,
|
| 346 |
+
"learning_rate": 4.2e-05,
|
| 347 |
+
"loss": 0.0167,
|
| 348 |
+
"step": 420
|
| 349 |
+
},
|
| 350 |
+
{
|
| 351 |
+
"epoch": 0.2097560975609756,
|
| 352 |
+
"grad_norm": 0.08243865519762039,
|
| 353 |
+
"learning_rate": 4.3e-05,
|
| 354 |
+
"loss": 0.0161,
|
| 355 |
+
"step": 430
|
| 356 |
+
},
|
| 357 |
+
{
|
| 358 |
+
"epoch": 0.2146341463414634,
|
| 359 |
+
"grad_norm": 0.06643176078796387,
|
| 360 |
+
"learning_rate": 4.4000000000000006e-05,
|
| 361 |
+
"loss": 0.0163,
|
| 362 |
+
"step": 440
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.21951219512195122,
|
| 366 |
+
"grad_norm": 0.05310194939374924,
|
| 367 |
+
"learning_rate": 4.5e-05,
|
| 368 |
+
"loss": 0.0149,
|
| 369 |
+
"step": 450
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.22439024390243903,
|
| 373 |
+
"grad_norm": 0.08369413018226624,
|
| 374 |
+
"learning_rate": 4.600000000000001e-05,
|
| 375 |
+
"loss": 0.0149,
|
| 376 |
+
"step": 460
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 0.22926829268292684,
|
| 380 |
+
"grad_norm": 0.06853285431861877,
|
| 381 |
+
"learning_rate": 4.7e-05,
|
| 382 |
+
"loss": 0.0142,
|
| 383 |
+
"step": 470
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 0.23414634146341465,
|
| 387 |
+
"grad_norm": 0.09447991102933884,
|
| 388 |
+
"learning_rate": 4.8e-05,
|
| 389 |
+
"loss": 0.0188,
|
| 390 |
+
"step": 480
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 0.23902439024390243,
|
| 394 |
+
"grad_norm": 0.09374509006738663,
|
| 395 |
+
"learning_rate": 4.9e-05,
|
| 396 |
+
"loss": 0.0174,
|
| 397 |
+
"step": 490
|
| 398 |
+
},
|
| 399 |
+
{
|
| 400 |
+
"epoch": 0.24390243902439024,
|
| 401 |
+
"grad_norm": 0.08009267598390579,
|
| 402 |
+
"learning_rate": 5e-05,
|
| 403 |
+
"loss": 0.0178,
|
| 404 |
+
"step": 500
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.24390243902439024,
|
| 408 |
+
"eval_f1": 0.6727972910165327,
|
| 409 |
+
"eval_loss": 0.019859502092003822,
|
| 410 |
+
"eval_precision": 0.6080043201728069,
|
| 411 |
+
"eval_recall": 0.7530469678953626,
|
| 412 |
+
"eval_runtime": 125.7109,
|
| 413 |
+
"eval_samples_per_second": 65.229,
|
| 414 |
+
"eval_steps_per_second": 1.026,
|
| 415 |
+
"step": 500
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.24878048780487805,
|
| 419 |
+
"grad_norm": 0.10110995173454285,
|
| 420 |
+
"learning_rate": 4.9911504424778765e-05,
|
| 421 |
+
"loss": 0.0176,
|
| 422 |
+
"step": 510
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.25365853658536586,
|
| 426 |
+
"grad_norm": 0.13534288108348846,
|
| 427 |
+
"learning_rate": 4.982300884955752e-05,
|
| 428 |
+
"loss": 0.019,
|
| 429 |
+
"step": 520
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.25853658536585367,
|
| 433 |
+
"grad_norm": 0.08551128208637238,
|
| 434 |
+
"learning_rate": 4.9734513274336284e-05,
|
| 435 |
+
"loss": 0.0156,
|
| 436 |
+
"step": 530
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.2634146341463415,
|
| 440 |
+
"grad_norm": 0.10786890238523483,
|
| 441 |
+
"learning_rate": 4.964601769911505e-05,
|
| 442 |
+
"loss": 0.0141,
|
| 443 |
+
"step": 540
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.2682926829268293,
|
| 447 |
+
"grad_norm": 0.10076329857110977,
|
| 448 |
+
"learning_rate": 4.955752212389381e-05,
|
| 449 |
+
"loss": 0.0175,
|
| 450 |
+
"step": 550
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.2731707317073171,
|
| 454 |
+
"grad_norm": 0.10950438678264618,
|
| 455 |
+
"learning_rate": 4.946902654867257e-05,
|
| 456 |
+
"loss": 0.0204,
|
| 457 |
+
"step": 560
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.2780487804878049,
|
| 461 |
+
"grad_norm": 0.14525355398654938,
|
| 462 |
+
"learning_rate": 4.938053097345133e-05,
|
| 463 |
+
"loss": 0.0188,
|
| 464 |
+
"step": 570
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.28292682926829266,
|
| 468 |
+
"grad_norm": 0.10379478335380554,
|
| 469 |
+
"learning_rate": 4.929203539823009e-05,
|
| 470 |
+
"loss": 0.0172,
|
| 471 |
+
"step": 580
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.28780487804878047,
|
| 475 |
+
"grad_norm": 0.11898482590913773,
|
| 476 |
+
"learning_rate": 4.9203539823008854e-05,
|
| 477 |
+
"loss": 0.0182,
|
| 478 |
+
"step": 590
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.2926829268292683,
|
| 482 |
+
"grad_norm": 0.09698235988616943,
|
| 483 |
+
"learning_rate": 4.911504424778761e-05,
|
| 484 |
+
"loss": 0.0161,
|
| 485 |
+
"step": 600
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.2926829268292683,
|
| 489 |
+
"eval_f1": 0.679977521404251,
|
| 490 |
+
"eval_loss": 0.01935453712940216,
|
| 491 |
+
"eval_precision": 0.612384638285204,
|
| 492 |
+
"eval_recall": 0.7643430439952438,
|
| 493 |
+
"eval_runtime": 126.1147,
|
| 494 |
+
"eval_samples_per_second": 65.02,
|
| 495 |
+
"eval_steps_per_second": 1.023,
|
| 496 |
+
"step": 600
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 0.2975609756097561,
|
| 500 |
+
"grad_norm": 0.11621833592653275,
|
| 501 |
+
"learning_rate": 4.902654867256637e-05,
|
| 502 |
+
"loss": 0.0172,
|
| 503 |
+
"step": 610
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 0.3024390243902439,
|
| 507 |
+
"grad_norm": 0.09230925887823105,
|
| 508 |
+
"learning_rate": 4.893805309734513e-05,
|
| 509 |
+
"loss": 0.0184,
|
| 510 |
+
"step": 620
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 0.3073170731707317,
|
| 514 |
+
"grad_norm": 0.07670127600431442,
|
| 515 |
+
"learning_rate": 4.88495575221239e-05,
|
| 516 |
+
"loss": 0.0157,
|
| 517 |
+
"step": 630
|
| 518 |
+
},
|
| 519 |
+
{
|
| 520 |
+
"epoch": 0.3121951219512195,
|
| 521 |
+
"grad_norm": 0.09875659644603729,
|
| 522 |
+
"learning_rate": 4.876106194690266e-05,
|
| 523 |
+
"loss": 0.0153,
|
| 524 |
+
"step": 640
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.3170731707317073,
|
| 528 |
+
"grad_norm": 0.07239075750112534,
|
| 529 |
+
"learning_rate": 4.867256637168142e-05,
|
| 530 |
+
"loss": 0.0137,
|
| 531 |
+
"step": 650
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 0.32195121951219513,
|
| 535 |
+
"grad_norm": 0.0903860479593277,
|
| 536 |
+
"learning_rate": 4.858407079646018e-05,
|
| 537 |
+
"loss": 0.0157,
|
| 538 |
+
"step": 660
|
| 539 |
+
},
|
| 540 |
+
{
|
| 541 |
+
"epoch": 0.32682926829268294,
|
| 542 |
+
"grad_norm": 0.10013315826654434,
|
| 543 |
+
"learning_rate": 4.849557522123894e-05,
|
| 544 |
+
"loss": 0.0163,
|
| 545 |
+
"step": 670
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 0.33170731707317075,
|
| 549 |
+
"grad_norm": 0.10445868223905563,
|
| 550 |
+
"learning_rate": 4.84070796460177e-05,
|
| 551 |
+
"loss": 0.0162,
|
| 552 |
+
"step": 680
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 0.33658536585365856,
|
| 556 |
+
"grad_norm": 0.09091856330633163,
|
| 557 |
+
"learning_rate": 4.831858407079646e-05,
|
| 558 |
+
"loss": 0.0173,
|
| 559 |
+
"step": 690
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 0.34146341463414637,
|
| 563 |
+
"grad_norm": 0.07674919068813324,
|
| 564 |
+
"learning_rate": 4.823008849557522e-05,
|
| 565 |
+
"loss": 0.0186,
|
| 566 |
+
"step": 700
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 0.34146341463414637,
|
| 570 |
+
"eval_f1": 0.6647128916943094,
|
| 571 |
+
"eval_loss": 0.019945966079831123,
|
| 572 |
+
"eval_precision": 0.5865302642796248,
|
| 573 |
+
"eval_recall": 0.7669441141498217,
|
| 574 |
+
"eval_runtime": 125.0727,
|
| 575 |
+
"eval_samples_per_second": 65.562,
|
| 576 |
+
"eval_steps_per_second": 1.031,
|
| 577 |
+
"step": 700
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 0.3463414634146341,
|
| 581 |
+
"grad_norm": 0.10548862814903259,
|
| 582 |
+
"learning_rate": 4.814159292035398e-05,
|
| 583 |
+
"loss": 0.0206,
|
| 584 |
+
"step": 710
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.35121951219512193,
|
| 588 |
+
"grad_norm": 0.0873124822974205,
|
| 589 |
+
"learning_rate": 4.805309734513275e-05,
|
| 590 |
+
"loss": 0.0178,
|
| 591 |
+
"step": 720
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 0.35609756097560974,
|
| 595 |
+
"grad_norm": 0.13294565677642822,
|
| 596 |
+
"learning_rate": 4.7964601769911506e-05,
|
| 597 |
+
"loss": 0.0171,
|
| 598 |
+
"step": 730
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 0.36097560975609755,
|
| 602 |
+
"grad_norm": 0.09687939286231995,
|
| 603 |
+
"learning_rate": 4.787610619469027e-05,
|
| 604 |
+
"loss": 0.021,
|
| 605 |
+
"step": 740
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 0.36585365853658536,
|
| 609 |
+
"grad_norm": 0.0876525342464447,
|
| 610 |
+
"learning_rate": 4.778761061946903e-05,
|
| 611 |
+
"loss": 0.0167,
|
| 612 |
+
"step": 750
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 0.37073170731707317,
|
| 616 |
+
"grad_norm": 0.09069986641407013,
|
| 617 |
+
"learning_rate": 4.769911504424779e-05,
|
| 618 |
+
"loss": 0.0159,
|
| 619 |
+
"step": 760
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 0.375609756097561,
|
| 623 |
+
"grad_norm": 0.09903474152088165,
|
| 624 |
+
"learning_rate": 4.761061946902655e-05,
|
| 625 |
+
"loss": 0.0179,
|
| 626 |
+
"step": 770
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.3804878048780488,
|
| 630 |
+
"grad_norm": 0.09864596277475357,
|
| 631 |
+
"learning_rate": 4.752212389380531e-05,
|
| 632 |
+
"loss": 0.0186,
|
| 633 |
+
"step": 780
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 0.3853658536585366,
|
| 637 |
+
"grad_norm": 0.14535708725452423,
|
| 638 |
+
"learning_rate": 4.743362831858407e-05,
|
| 639 |
+
"loss": 0.0163,
|
| 640 |
+
"step": 790
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 0.3902439024390244,
|
| 644 |
+
"grad_norm": 0.09460794180631638,
|
| 645 |
+
"learning_rate": 4.734513274336283e-05,
|
| 646 |
+
"loss": 0.0188,
|
| 647 |
+
"step": 800
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 0.3902439024390244,
|
| 651 |
+
"eval_f1": 0.6686207562971075,
|
| 652 |
+
"eval_loss": 0.019426193088293076,
|
| 653 |
+
"eval_precision": 0.5900511654349062,
|
| 654 |
+
"eval_recall": 0.7713287752675386,
|
| 655 |
+
"eval_runtime": 125.9109,
|
| 656 |
+
"eval_samples_per_second": 65.125,
|
| 657 |
+
"eval_steps_per_second": 1.025,
|
| 658 |
+
"step": 800
|
| 659 |
+
},
|
| 660 |
+
{
|
| 661 |
+
"epoch": 0.3951219512195122,
|
| 662 |
+
"grad_norm": 0.10055914521217346,
|
| 663 |
+
"learning_rate": 4.7256637168141595e-05,
|
| 664 |
+
"loss": 0.0163,
|
| 665 |
+
"step": 810
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 0.4,
|
| 669 |
+
"grad_norm": 0.09273848682641983,
|
| 670 |
+
"learning_rate": 4.716814159292036e-05,
|
| 671 |
+
"loss": 0.0169,
|
| 672 |
+
"step": 820
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 0.40487804878048783,
|
| 676 |
+
"grad_norm": 0.11031738668680191,
|
| 677 |
+
"learning_rate": 4.707964601769912e-05,
|
| 678 |
+
"loss": 0.0175,
|
| 679 |
+
"step": 830
|
| 680 |
+
},
|
| 681 |
+
{
|
| 682 |
+
"epoch": 0.4097560975609756,
|
| 683 |
+
"grad_norm": 0.09007064253091812,
|
| 684 |
+
"learning_rate": 4.699115044247788e-05,
|
| 685 |
+
"loss": 0.0161,
|
| 686 |
+
"step": 840
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 0.4146341463414634,
|
| 690 |
+
"grad_norm": 0.10067761689424515,
|
| 691 |
+
"learning_rate": 4.690265486725664e-05,
|
| 692 |
+
"loss": 0.0177,
|
| 693 |
+
"step": 850
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 0.4195121951219512,
|
| 697 |
+
"grad_norm": 0.15128421783447266,
|
| 698 |
+
"learning_rate": 4.6814159292035396e-05,
|
| 699 |
+
"loss": 0.0183,
|
| 700 |
+
"step": 860
|
| 701 |
+
},
|
| 702 |
+
{
|
| 703 |
+
"epoch": 0.424390243902439,
|
| 704 |
+
"grad_norm": 0.10395248234272003,
|
| 705 |
+
"learning_rate": 4.672566371681416e-05,
|
| 706 |
+
"loss": 0.0151,
|
| 707 |
+
"step": 870
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"epoch": 0.4292682926829268,
|
| 711 |
+
"grad_norm": 0.08134312182664871,
|
| 712 |
+
"learning_rate": 4.663716814159292e-05,
|
| 713 |
+
"loss": 0.0184,
|
| 714 |
+
"step": 880
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 0.43414634146341463,
|
| 718 |
+
"grad_norm": 0.09317316114902496,
|
| 719 |
+
"learning_rate": 4.6548672566371684e-05,
|
| 720 |
+
"loss": 0.0177,
|
| 721 |
+
"step": 890
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 0.43902439024390244,
|
| 725 |
+
"grad_norm": 0.09974437952041626,
|
| 726 |
+
"learning_rate": 4.646017699115045e-05,
|
| 727 |
+
"loss": 0.0164,
|
| 728 |
+
"step": 900
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 0.43902439024390244,
|
| 732 |
+
"eval_f1": 0.6717267552182163,
|
| 733 |
+
"eval_loss": 0.01958908513188362,
|
| 734 |
+
"eval_precision": 0.6,
|
| 735 |
+
"eval_recall": 0.7629310344827587,
|
| 736 |
+
"eval_runtime": 125.8067,
|
| 737 |
+
"eval_samples_per_second": 65.179,
|
| 738 |
+
"eval_steps_per_second": 1.025,
|
| 739 |
+
"step": 900
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"epoch": 0.44390243902439025,
|
| 743 |
+
"grad_norm": 0.10190931707620621,
|
| 744 |
+
"learning_rate": 4.637168141592921e-05,
|
| 745 |
+
"loss": 0.0219,
|
| 746 |
+
"step": 910
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 0.44878048780487806,
|
| 750 |
+
"grad_norm": 0.08647562563419342,
|
| 751 |
+
"learning_rate": 4.6283185840707966e-05,
|
| 752 |
+
"loss": 0.0152,
|
| 753 |
+
"step": 920
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 0.45365853658536587,
|
| 757 |
+
"grad_norm": 0.07675183564424515,
|
| 758 |
+
"learning_rate": 4.619469026548673e-05,
|
| 759 |
+
"loss": 0.0164,
|
| 760 |
+
"step": 930
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 0.4585365853658537,
|
| 764 |
+
"grad_norm": 0.11960858851671219,
|
| 765 |
+
"learning_rate": 4.6106194690265485e-05,
|
| 766 |
+
"loss": 0.0172,
|
| 767 |
+
"step": 940
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 0.4634146341463415,
|
| 771 |
+
"grad_norm": 0.09075548499822617,
|
| 772 |
+
"learning_rate": 4.601769911504425e-05,
|
| 773 |
+
"loss": 0.0159,
|
| 774 |
+
"step": 950
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 0.4682926829268293,
|
| 778 |
+
"grad_norm": 0.05601793900132179,
|
| 779 |
+
"learning_rate": 4.592920353982301e-05,
|
| 780 |
+
"loss": 0.0171,
|
| 781 |
+
"step": 960
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 0.47317073170731705,
|
| 785 |
+
"grad_norm": 0.0845816358923912,
|
| 786 |
+
"learning_rate": 4.584070796460177e-05,
|
| 787 |
+
"loss": 0.0185,
|
| 788 |
+
"step": 970
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 0.47804878048780486,
|
| 792 |
+
"grad_norm": 0.08288878947496414,
|
| 793 |
+
"learning_rate": 4.5752212389380536e-05,
|
| 794 |
+
"loss": 0.017,
|
| 795 |
+
"step": 980
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 0.48292682926829267,
|
| 799 |
+
"grad_norm": 0.1054663434624672,
|
| 800 |
+
"learning_rate": 4.56637168141593e-05,
|
| 801 |
+
"loss": 0.0148,
|
| 802 |
+
"step": 990
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 0.4878048780487805,
|
| 806 |
+
"grad_norm": 0.08182746171951294,
|
| 807 |
+
"learning_rate": 4.5575221238938055e-05,
|
| 808 |
+
"loss": 0.0164,
|
| 809 |
+
"step": 1000
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 0.4878048780487805,
|
| 813 |
+
"eval_f1": 0.6684826669260953,
|
| 814 |
+
"eval_loss": 0.019462432712316513,
|
| 815 |
+
"eval_precision": 0.5930038547839594,
|
| 816 |
+
"eval_recall": 0.7659780023781213,
|
| 817 |
+
"eval_runtime": 125.7047,
|
| 818 |
+
"eval_samples_per_second": 65.232,
|
| 819 |
+
"eval_steps_per_second": 1.026,
|
| 820 |
+
"step": 1000
|
| 821 |
+
},
|
| 822 |
+
{
|
| 823 |
+
"epoch": 0.4926829268292683,
|
| 824 |
+
"grad_norm": 0.09717566519975662,
|
| 825 |
+
"learning_rate": 4.548672566371682e-05,
|
| 826 |
+
"loss": 0.0182,
|
| 827 |
+
"step": 1010
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"epoch": 0.4975609756097561,
|
| 831 |
+
"grad_norm": 0.09170341491699219,
|
| 832 |
+
"learning_rate": 4.5398230088495574e-05,
|
| 833 |
+
"loss": 0.0175,
|
| 834 |
+
"step": 1020
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 0.5024390243902439,
|
| 838 |
+
"grad_norm": 0.08827481418848038,
|
| 839 |
+
"learning_rate": 4.5309734513274336e-05,
|
| 840 |
+
"loss": 0.0177,
|
| 841 |
+
"step": 1030
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"epoch": 0.5073170731707317,
|
| 845 |
+
"grad_norm": 0.08792293816804886,
|
| 846 |
+
"learning_rate": 4.52212389380531e-05,
|
| 847 |
+
"loss": 0.0184,
|
| 848 |
+
"step": 1040
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 0.5121951219512195,
|
| 852 |
+
"grad_norm": 0.13856299221515656,
|
| 853 |
+
"learning_rate": 4.5132743362831855e-05,
|
| 854 |
+
"loss": 0.0182,
|
| 855 |
+
"step": 1050
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 0.5170731707317073,
|
| 859 |
+
"grad_norm": 0.0926881730556488,
|
| 860 |
+
"learning_rate": 4.5044247787610625e-05,
|
| 861 |
+
"loss": 0.0177,
|
| 862 |
+
"step": 1060
|
| 863 |
+
},
|
| 864 |
+
{
|
| 865 |
+
"epoch": 0.5219512195121951,
|
| 866 |
+
"grad_norm": 0.0846000462770462,
|
| 867 |
+
"learning_rate": 4.495575221238939e-05,
|
| 868 |
+
"loss": 0.0157,
|
| 869 |
+
"step": 1070
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 0.526829268292683,
|
| 873 |
+
"grad_norm": 0.10093377530574799,
|
| 874 |
+
"learning_rate": 4.4867256637168144e-05,
|
| 875 |
+
"loss": 0.0188,
|
| 876 |
+
"step": 1080
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 0.5317073170731708,
|
| 880 |
+
"grad_norm": 0.09538795799016953,
|
| 881 |
+
"learning_rate": 4.4778761061946906e-05,
|
| 882 |
+
"loss": 0.0184,
|
| 883 |
+
"step": 1090
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 0.5365853658536586,
|
| 887 |
+
"grad_norm": 0.14831095933914185,
|
| 888 |
+
"learning_rate": 4.469026548672566e-05,
|
| 889 |
+
"loss": 0.0159,
|
| 890 |
+
"step": 1100
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 0.5365853658536586,
|
| 894 |
+
"eval_f1": 0.6750937261537441,
|
| 895 |
+
"eval_loss": 0.020070159807801247,
|
| 896 |
+
"eval_precision": 0.6097692538207972,
|
| 897 |
+
"eval_recall": 0.7560939357907254,
|
| 898 |
+
"eval_runtime": 126.5385,
|
| 899 |
+
"eval_samples_per_second": 64.802,
|
| 900 |
+
"eval_steps_per_second": 1.019,
|
| 901 |
+
"step": 1100
|
| 902 |
+
},
|
| 903 |
+
{
|
| 904 |
+
"epoch": 0.5414634146341464,
|
| 905 |
+
"grad_norm": 0.09911943227052689,
|
| 906 |
+
"learning_rate": 4.4601769911504425e-05,
|
| 907 |
+
"loss": 0.0174,
|
| 908 |
+
"step": 1110
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 0.5463414634146342,
|
| 912 |
+
"grad_norm": 0.07620345056056976,
|
| 913 |
+
"learning_rate": 4.451327433628319e-05,
|
| 914 |
+
"loss": 0.0157,
|
| 915 |
+
"step": 1120
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 0.551219512195122,
|
| 919 |
+
"grad_norm": 0.09838444739580154,
|
| 920 |
+
"learning_rate": 4.4424778761061944e-05,
|
| 921 |
+
"loss": 0.0146,
|
| 922 |
+
"step": 1130
|
| 923 |
+
},
|
| 924 |
+
{
|
| 925 |
+
"epoch": 0.5560975609756098,
|
| 926 |
+
"grad_norm": 0.08282183855772018,
|
| 927 |
+
"learning_rate": 4.433628318584071e-05,
|
| 928 |
+
"loss": 0.0178,
|
| 929 |
+
"step": 1140
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 0.5609756097560976,
|
| 933 |
+
"grad_norm": 0.09383049607276917,
|
| 934 |
+
"learning_rate": 4.4247787610619477e-05,
|
| 935 |
+
"loss": 0.0164,
|
| 936 |
+
"step": 1150
|
| 937 |
+
},
|
| 938 |
+
{
|
| 939 |
+
"epoch": 0.5658536585365853,
|
| 940 |
+
"grad_norm": 0.1041053757071495,
|
| 941 |
+
"learning_rate": 4.415929203539823e-05,
|
| 942 |
+
"loss": 0.018,
|
| 943 |
+
"step": 1160
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 0.5707317073170731,
|
| 947 |
+
"grad_norm": 0.0997629165649414,
|
| 948 |
+
"learning_rate": 4.4070796460176995e-05,
|
| 949 |
+
"loss": 0.016,
|
| 950 |
+
"step": 1170
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 0.5756097560975609,
|
| 954 |
+
"grad_norm": 0.1428525149822235,
|
| 955 |
+
"learning_rate": 4.398230088495575e-05,
|
| 956 |
+
"loss": 0.018,
|
| 957 |
+
"step": 1180
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 0.5804878048780487,
|
| 961 |
+
"grad_norm": 0.1270400732755661,
|
| 962 |
+
"learning_rate": 4.3893805309734514e-05,
|
| 963 |
+
"loss": 0.0172,
|
| 964 |
+
"step": 1190
|
| 965 |
+
},
|
| 966 |
+
{
|
| 967 |
+
"epoch": 0.5853658536585366,
|
| 968 |
+
"grad_norm": 0.08549308031797409,
|
| 969 |
+
"learning_rate": 4.380530973451328e-05,
|
| 970 |
+
"loss": 0.0147,
|
| 971 |
+
"step": 1200
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"epoch": 0.5853658536585366,
|
| 975 |
+
"eval_f1": 0.6807967485136542,
|
| 976 |
+
"eval_loss": 0.019661063328385353,
|
| 977 |
+
"eval_precision": 0.621146184492798,
|
| 978 |
+
"eval_recall": 0.7531212841854935,
|
| 979 |
+
"eval_runtime": 125.9122,
|
| 980 |
+
"eval_samples_per_second": 65.125,
|
| 981 |
+
"eval_steps_per_second": 1.025,
|
| 982 |
+
"step": 1200
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.5902439024390244,
|
| 986 |
+
"grad_norm": 0.05166243761777878,
|
| 987 |
+
"learning_rate": 4.371681415929203e-05,
|
| 988 |
+
"loss": 0.0167,
|
| 989 |
+
"step": 1210
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.5951219512195122,
|
| 993 |
+
"grad_norm": 0.0989900454878807,
|
| 994 |
+
"learning_rate": 4.3628318584070796e-05,
|
| 995 |
+
"loss": 0.0185,
|
| 996 |
+
"step": 1220
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.6,
|
| 1000 |
+
"grad_norm": 0.09084061533212662,
|
| 1001 |
+
"learning_rate": 4.353982300884956e-05,
|
| 1002 |
+
"loss": 0.0175,
|
| 1003 |
+
"step": 1230
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.6048780487804878,
|
| 1007 |
+
"grad_norm": 0.08588280528783798,
|
| 1008 |
+
"learning_rate": 4.345132743362832e-05,
|
| 1009 |
+
"loss": 0.0154,
|
| 1010 |
+
"step": 1240
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.6097560975609756,
|
| 1014 |
+
"grad_norm": 0.07876714318990707,
|
| 1015 |
+
"learning_rate": 4.3362831858407084e-05,
|
| 1016 |
+
"loss": 0.0155,
|
| 1017 |
+
"step": 1250
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.6146341463414634,
|
| 1021 |
+
"grad_norm": 0.10766426473855972,
|
| 1022 |
+
"learning_rate": 4.327433628318584e-05,
|
| 1023 |
+
"loss": 0.0162,
|
| 1024 |
+
"step": 1260
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.6195121951219512,
|
| 1028 |
+
"grad_norm": 0.08020301163196564,
|
| 1029 |
+
"learning_rate": 4.31858407079646e-05,
|
| 1030 |
+
"loss": 0.0174,
|
| 1031 |
+
"step": 1270
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.624390243902439,
|
| 1035 |
+
"grad_norm": 0.11564213037490845,
|
| 1036 |
+
"learning_rate": 4.3097345132743366e-05,
|
| 1037 |
+
"loss": 0.0167,
|
| 1038 |
+
"step": 1280
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.6292682926829268,
|
| 1042 |
+
"grad_norm": 0.11487080901861191,
|
| 1043 |
+
"learning_rate": 4.300884955752212e-05,
|
| 1044 |
+
"loss": 0.0155,
|
| 1045 |
+
"step": 1290
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.6341463414634146,
|
| 1049 |
+
"grad_norm": 0.09950511902570724,
|
| 1050 |
+
"learning_rate": 4.2920353982300885e-05,
|
| 1051 |
+
"loss": 0.0159,
|
| 1052 |
+
"step": 1300
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.6341463414634146,
|
| 1056 |
+
"eval_f1": 0.687796188619635,
|
| 1057 |
+
"eval_loss": 0.01919134519994259,
|
| 1058 |
+
"eval_precision": 0.6278456157575014,
|
| 1059 |
+
"eval_recall": 0.7604042806183116,
|
| 1060 |
+
"eval_runtime": 125.3379,
|
| 1061 |
+
"eval_samples_per_second": 65.423,
|
| 1062 |
+
"eval_steps_per_second": 1.029,
|
| 1063 |
+
"step": 1300
|
| 1064 |
+
},
|
| 1065 |
+
{
|
| 1066 |
+
"epoch": 0.6390243902439025,
|
| 1067 |
+
"grad_norm": 0.08389411866664886,
|
| 1068 |
+
"learning_rate": 4.283185840707965e-05,
|
| 1069 |
+
"loss": 0.0156,
|
| 1070 |
+
"step": 1310
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 0.6439024390243903,
|
| 1074 |
+
"grad_norm": 0.08979038149118423,
|
| 1075 |
+
"learning_rate": 4.274336283185841e-05,
|
| 1076 |
+
"loss": 0.0173,
|
| 1077 |
+
"step": 1320
|
| 1078 |
+
},
|
| 1079 |
+
{
|
| 1080 |
+
"epoch": 0.6487804878048781,
|
| 1081 |
+
"grad_norm": 0.069185771048069,
|
| 1082 |
+
"learning_rate": 4.265486725663717e-05,
|
| 1083 |
+
"loss": 0.0178,
|
| 1084 |
+
"step": 1330
|
| 1085 |
+
},
|
| 1086 |
+
{
|
| 1087 |
+
"epoch": 0.6536585365853659,
|
| 1088 |
+
"grad_norm": 0.07186874747276306,
|
| 1089 |
+
"learning_rate": 4.256637168141593e-05,
|
| 1090 |
+
"loss": 0.0153,
|
| 1091 |
+
"step": 1340
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 0.6585365853658537,
|
| 1095 |
+
"grad_norm": 0.11203644424676895,
|
| 1096 |
+
"learning_rate": 4.247787610619469e-05,
|
| 1097 |
+
"loss": 0.0167,
|
| 1098 |
+
"step": 1350
|
| 1099 |
+
},
|
| 1100 |
+
{
|
| 1101 |
+
"epoch": 0.6634146341463415,
|
| 1102 |
+
"grad_norm": 0.11567346006631851,
|
| 1103 |
+
"learning_rate": 4.2389380530973455e-05,
|
| 1104 |
+
"loss": 0.017,
|
| 1105 |
+
"step": 1360
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 0.6682926829268293,
|
| 1109 |
+
"grad_norm": 0.11276240646839142,
|
| 1110 |
+
"learning_rate": 4.230088495575221e-05,
|
| 1111 |
+
"loss": 0.0158,
|
| 1112 |
+
"step": 1370
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 0.6731707317073171,
|
| 1116 |
+
"grad_norm": 0.07067442685365677,
|
| 1117 |
+
"learning_rate": 4.2212389380530974e-05,
|
| 1118 |
+
"loss": 0.0156,
|
| 1119 |
+
"step": 1380
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 0.6780487804878049,
|
| 1123 |
+
"grad_norm": 0.08775879442691803,
|
| 1124 |
+
"learning_rate": 4.2123893805309737e-05,
|
| 1125 |
+
"loss": 0.0155,
|
| 1126 |
+
"step": 1390
|
| 1127 |
+
},
|
| 1128 |
+
{
|
| 1129 |
+
"epoch": 0.6829268292682927,
|
| 1130 |
+
"grad_norm": 0.07028288394212723,
|
| 1131 |
+
"learning_rate": 4.20353982300885e-05,
|
| 1132 |
+
"loss": 0.0142,
|
| 1133 |
+
"step": 1400
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 0.6829268292682927,
|
| 1137 |
+
"eval_f1": 0.6818793753365644,
|
| 1138 |
+
"eval_loss": 0.019489064812660217,
|
| 1139 |
+
"eval_precision": 0.6231545275590551,
|
| 1140 |
+
"eval_recall": 0.7528240190249703,
|
| 1141 |
+
"eval_runtime": 125.915,
|
| 1142 |
+
"eval_samples_per_second": 65.123,
|
| 1143 |
+
"eval_steps_per_second": 1.025,
|
| 1144 |
+
"step": 1400
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 0.6878048780487804,
|
| 1148 |
+
"grad_norm": 0.08612460643053055,
|
| 1149 |
+
"learning_rate": 4.194690265486726e-05,
|
| 1150 |
+
"loss": 0.0163,
|
| 1151 |
+
"step": 1410
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 0.6926829268292682,
|
| 1155 |
+
"grad_norm": 0.1031985655426979,
|
| 1156 |
+
"learning_rate": 4.185840707964602e-05,
|
| 1157 |
+
"loss": 0.0146,
|
| 1158 |
+
"step": 1420
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 0.697560975609756,
|
| 1162 |
+
"grad_norm": 0.10174574702978134,
|
| 1163 |
+
"learning_rate": 4.176991150442478e-05,
|
| 1164 |
+
"loss": 0.017,
|
| 1165 |
+
"step": 1430
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 0.7024390243902439,
|
| 1169 |
+
"grad_norm": 0.06620911508798599,
|
| 1170 |
+
"learning_rate": 4.1681415929203544e-05,
|
| 1171 |
+
"loss": 0.0171,
|
| 1172 |
+
"step": 1440
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 0.7073170731707317,
|
| 1176 |
+
"grad_norm": 0.11920668184757233,
|
| 1177 |
+
"learning_rate": 4.15929203539823e-05,
|
| 1178 |
+
"loss": 0.0191,
|
| 1179 |
+
"step": 1450
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 0.7121951219512195,
|
| 1183 |
+
"grad_norm": 0.10696449875831604,
|
| 1184 |
+
"learning_rate": 4.150442477876106e-05,
|
| 1185 |
+
"loss": 0.0164,
|
| 1186 |
+
"step": 1460
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 0.7170731707317073,
|
| 1190 |
+
"grad_norm": 0.09047581255435944,
|
| 1191 |
+
"learning_rate": 4.1415929203539825e-05,
|
| 1192 |
+
"loss": 0.0185,
|
| 1193 |
+
"step": 1470
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 0.7219512195121951,
|
| 1197 |
+
"grad_norm": 0.08915913105010986,
|
| 1198 |
+
"learning_rate": 4.132743362831858e-05,
|
| 1199 |
+
"loss": 0.0172,
|
| 1200 |
+
"step": 1480
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 0.7268292682926829,
|
| 1204 |
+
"grad_norm": 0.10107399523258209,
|
| 1205 |
+
"learning_rate": 4.123893805309735e-05,
|
| 1206 |
+
"loss": 0.0188,
|
| 1207 |
+
"step": 1490
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 0.7317073170731707,
|
| 1211 |
+
"grad_norm": 0.1132921576499939,
|
| 1212 |
+
"learning_rate": 4.115044247787611e-05,
|
| 1213 |
+
"loss": 0.0166,
|
| 1214 |
+
"step": 1500
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 0.7317073170731707,
|
| 1218 |
+
"eval_f1": 0.6884484673104748,
|
| 1219 |
+
"eval_loss": 0.0191908348351717,
|
| 1220 |
+
"eval_precision": 0.6290352333517801,
|
| 1221 |
+
"eval_recall": 0.7602556480380499,
|
| 1222 |
+
"eval_runtime": 126.3472,
|
| 1223 |
+
"eval_samples_per_second": 64.901,
|
| 1224 |
+
"eval_steps_per_second": 1.021,
|
| 1225 |
+
"step": 1500
|
| 1226 |
+
}
|
| 1227 |
+
],
|
| 1228 |
+
"logging_steps": 10,
|
| 1229 |
+
"max_steps": 6150,
|
| 1230 |
+
"num_input_tokens_seen": 0,
|
| 1231 |
+
"num_train_epochs": 3,
|
| 1232 |
+
"save_steps": 100,
|
| 1233 |
+
"stateful_callbacks": {
|
| 1234 |
+
"TrainerControl": {
|
| 1235 |
+
"args": {
|
| 1236 |
+
"should_epoch_stop": false,
|
| 1237 |
+
"should_evaluate": false,
|
| 1238 |
+
"should_log": false,
|
| 1239 |
+
"should_save": true,
|
| 1240 |
+
"should_training_stop": false
|
| 1241 |
+
},
|
| 1242 |
+
"attributes": {}
|
| 1243 |
+
}
|
| 1244 |
+
},
|
| 1245 |
+
"total_flos": 3186928336896000.0,
|
| 1246 |
+
"train_batch_size": 16,
|
| 1247 |
+
"trial_name": null,
|
| 1248 |
+
"trial_params": null
|
| 1249 |
+
}
|
training_args (2).bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ab1a6d23376ba1a8cf0a27a77657f45f0d973b85e017358496c06d962db4f23d
|
| 3 |
+
size 5304
|