Training in progress, step 80
Browse files
config.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
| 1 |
{
|
| 2 |
"_name_or_path": "bert-large-uncased",
|
| 3 |
"architectures": [
|
| 4 |
-
"
|
| 5 |
],
|
| 6 |
"attention_probs_dropout_prob": 0.1,
|
| 7 |
"classifier_dropout": null,
|
|
@@ -9,212 +9,8 @@
|
|
| 9 |
"hidden_act": "gelu",
|
| 10 |
"hidden_dropout_prob": 0.1,
|
| 11 |
"hidden_size": 1024,
|
| 12 |
-
"id2label": {
|
| 13 |
-
"0": "Section 2",
|
| 14 |
-
"1": "Section 3",
|
| 15 |
-
"2": "Section 4",
|
| 16 |
-
"3": "Section 5",
|
| 17 |
-
"4": "Section 13",
|
| 18 |
-
"5": "Section 34",
|
| 19 |
-
"6": "Section 107",
|
| 20 |
-
"7": "Section 109",
|
| 21 |
-
"8": "Section 114",
|
| 22 |
-
"9": "Section 120",
|
| 23 |
-
"10": "Section 120B",
|
| 24 |
-
"11": "Section 143",
|
| 25 |
-
"12": "Section 147",
|
| 26 |
-
"13": "Section 148",
|
| 27 |
-
"14": "Section 149",
|
| 28 |
-
"15": "Section 155",
|
| 29 |
-
"16": "Section 156",
|
| 30 |
-
"17": "Section 161",
|
| 31 |
-
"18": "Section 164",
|
| 32 |
-
"19": "Section 173",
|
| 33 |
-
"20": "Section 174A",
|
| 34 |
-
"21": "Section 186",
|
| 35 |
-
"22": "Section 188",
|
| 36 |
-
"23": "Section 190",
|
| 37 |
-
"24": "Section 193",
|
| 38 |
-
"25": "Section 200",
|
| 39 |
-
"26": "Section 201",
|
| 40 |
-
"27": "Section 228",
|
| 41 |
-
"28": "Section 229A",
|
| 42 |
-
"29": "Section 279",
|
| 43 |
-
"30": "Section 294",
|
| 44 |
-
"31": "Section 294(b)",
|
| 45 |
-
"32": "Section 299",
|
| 46 |
-
"33": "Section 300",
|
| 47 |
-
"34": "Section 302",
|
| 48 |
-
"35": "Section 304",
|
| 49 |
-
"36": "Section 304A",
|
| 50 |
-
"37": "Section 304B",
|
| 51 |
-
"38": "Section 306",
|
| 52 |
-
"39": "Section 307",
|
| 53 |
-
"40": "Section 308",
|
| 54 |
-
"41": "Section 313",
|
| 55 |
-
"42": "Section 320",
|
| 56 |
-
"43": "Section 323",
|
| 57 |
-
"44": "Section 324",
|
| 58 |
-
"45": "Section 325",
|
| 59 |
-
"46": "Section 326",
|
| 60 |
-
"47": "Section 332",
|
| 61 |
-
"48": "Section 336",
|
| 62 |
-
"49": "Section 337",
|
| 63 |
-
"50": "Section 338",
|
| 64 |
-
"51": "Section 341",
|
| 65 |
-
"52": "Section 342",
|
| 66 |
-
"53": "Section 353",
|
| 67 |
-
"54": "Section 354",
|
| 68 |
-
"55": "Section 363",
|
| 69 |
-
"56": "Section 364",
|
| 70 |
-
"57": "Section 365",
|
| 71 |
-
"58": "Section 366",
|
| 72 |
-
"59": "Section 366A",
|
| 73 |
-
"60": "Section 375",
|
| 74 |
-
"61": "Section 376",
|
| 75 |
-
"62": "Section 376(2)",
|
| 76 |
-
"63": "Section 379",
|
| 77 |
-
"64": "Section 380",
|
| 78 |
-
"65": "Section 384",
|
| 79 |
-
"66": "Section 389",
|
| 80 |
-
"67": "Section 392",
|
| 81 |
-
"68": "Section 394",
|
| 82 |
-
"69": "Section 395",
|
| 83 |
-
"70": "Section 397",
|
| 84 |
-
"71": "Section 406",
|
| 85 |
-
"72": "Section 409",
|
| 86 |
-
"73": "Section 411",
|
| 87 |
-
"74": "Section 415",
|
| 88 |
-
"75": "Section 417",
|
| 89 |
-
"76": "Section 419",
|
| 90 |
-
"77": "Section 420",
|
| 91 |
-
"78": "Section 427",
|
| 92 |
-
"79": "Section 436",
|
| 93 |
-
"80": "Section 437",
|
| 94 |
-
"81": "Section 438",
|
| 95 |
-
"82": "Section 447",
|
| 96 |
-
"83": "Section 448",
|
| 97 |
-
"84": "Section 450",
|
| 98 |
-
"85": "Section 452",
|
| 99 |
-
"86": "Section 457",
|
| 100 |
-
"87": "Section 465",
|
| 101 |
-
"88": "Section 467",
|
| 102 |
-
"89": "Section 468",
|
| 103 |
-
"90": "Section 471",
|
| 104 |
-
"91": "Section 482",
|
| 105 |
-
"92": "Section 494",
|
| 106 |
-
"93": "Section 498",
|
| 107 |
-
"94": "Section 498A",
|
| 108 |
-
"95": "Section 500",
|
| 109 |
-
"96": "Section 504",
|
| 110 |
-
"97": "Section 506",
|
| 111 |
-
"98": "Section 509",
|
| 112 |
-
"99": "Section 511"
|
| 113 |
-
},
|
| 114 |
"initializer_range": 0.02,
|
| 115 |
"intermediate_size": 4096,
|
| 116 |
-
"label2id": {
|
| 117 |
-
"Section 107": 6,
|
| 118 |
-
"Section 109": 7,
|
| 119 |
-
"Section 114": 8,
|
| 120 |
-
"Section 120": 9,
|
| 121 |
-
"Section 120B": 10,
|
| 122 |
-
"Section 13": 4,
|
| 123 |
-
"Section 143": 11,
|
| 124 |
-
"Section 147": 12,
|
| 125 |
-
"Section 148": 13,
|
| 126 |
-
"Section 149": 14,
|
| 127 |
-
"Section 155": 15,
|
| 128 |
-
"Section 156": 16,
|
| 129 |
-
"Section 161": 17,
|
| 130 |
-
"Section 164": 18,
|
| 131 |
-
"Section 173": 19,
|
| 132 |
-
"Section 174A": 20,
|
| 133 |
-
"Section 186": 21,
|
| 134 |
-
"Section 188": 22,
|
| 135 |
-
"Section 190": 23,
|
| 136 |
-
"Section 193": 24,
|
| 137 |
-
"Section 2": 0,
|
| 138 |
-
"Section 200": 25,
|
| 139 |
-
"Section 201": 26,
|
| 140 |
-
"Section 228": 27,
|
| 141 |
-
"Section 229A": 28,
|
| 142 |
-
"Section 279": 29,
|
| 143 |
-
"Section 294": 30,
|
| 144 |
-
"Section 294(b)": 31,
|
| 145 |
-
"Section 299": 32,
|
| 146 |
-
"Section 3": 1,
|
| 147 |
-
"Section 300": 33,
|
| 148 |
-
"Section 302": 34,
|
| 149 |
-
"Section 304": 35,
|
| 150 |
-
"Section 304A": 36,
|
| 151 |
-
"Section 304B": 37,
|
| 152 |
-
"Section 306": 38,
|
| 153 |
-
"Section 307": 39,
|
| 154 |
-
"Section 308": 40,
|
| 155 |
-
"Section 313": 41,
|
| 156 |
-
"Section 320": 42,
|
| 157 |
-
"Section 323": 43,
|
| 158 |
-
"Section 324": 44,
|
| 159 |
-
"Section 325": 45,
|
| 160 |
-
"Section 326": 46,
|
| 161 |
-
"Section 332": 47,
|
| 162 |
-
"Section 336": 48,
|
| 163 |
-
"Section 337": 49,
|
| 164 |
-
"Section 338": 50,
|
| 165 |
-
"Section 34": 5,
|
| 166 |
-
"Section 341": 51,
|
| 167 |
-
"Section 342": 52,
|
| 168 |
-
"Section 353": 53,
|
| 169 |
-
"Section 354": 54,
|
| 170 |
-
"Section 363": 55,
|
| 171 |
-
"Section 364": 56,
|
| 172 |
-
"Section 365": 57,
|
| 173 |
-
"Section 366": 58,
|
| 174 |
-
"Section 366A": 59,
|
| 175 |
-
"Section 375": 60,
|
| 176 |
-
"Section 376": 61,
|
| 177 |
-
"Section 376(2)": 62,
|
| 178 |
-
"Section 379": 63,
|
| 179 |
-
"Section 380": 64,
|
| 180 |
-
"Section 384": 65,
|
| 181 |
-
"Section 389": 66,
|
| 182 |
-
"Section 392": 67,
|
| 183 |
-
"Section 394": 68,
|
| 184 |
-
"Section 395": 69,
|
| 185 |
-
"Section 397": 70,
|
| 186 |
-
"Section 4": 2,
|
| 187 |
-
"Section 406": 71,
|
| 188 |
-
"Section 409": 72,
|
| 189 |
-
"Section 411": 73,
|
| 190 |
-
"Section 415": 74,
|
| 191 |
-
"Section 417": 75,
|
| 192 |
-
"Section 419": 76,
|
| 193 |
-
"Section 420": 77,
|
| 194 |
-
"Section 427": 78,
|
| 195 |
-
"Section 436": 79,
|
| 196 |
-
"Section 437": 80,
|
| 197 |
-
"Section 438": 81,
|
| 198 |
-
"Section 447": 82,
|
| 199 |
-
"Section 448": 83,
|
| 200 |
-
"Section 450": 84,
|
| 201 |
-
"Section 452": 85,
|
| 202 |
-
"Section 457": 86,
|
| 203 |
-
"Section 465": 87,
|
| 204 |
-
"Section 467": 88,
|
| 205 |
-
"Section 468": 89,
|
| 206 |
-
"Section 471": 90,
|
| 207 |
-
"Section 482": 91,
|
| 208 |
-
"Section 494": 92,
|
| 209 |
-
"Section 498": 93,
|
| 210 |
-
"Section 498A": 94,
|
| 211 |
-
"Section 5": 3,
|
| 212 |
-
"Section 500": 95,
|
| 213 |
-
"Section 504": 96,
|
| 214 |
-
"Section 506": 97,
|
| 215 |
-
"Section 509": 98,
|
| 216 |
-
"Section 511": 99
|
| 217 |
-
},
|
| 218 |
"layer_norm_eps": 1e-12,
|
| 219 |
"max_position_embeddings": 512,
|
| 220 |
"model_type": "bert",
|
|
@@ -222,7 +18,6 @@
|
|
| 222 |
"num_hidden_layers": 24,
|
| 223 |
"pad_token_id": 0,
|
| 224 |
"position_embedding_type": "absolute",
|
| 225 |
-
"problem_type": "multi_label_classification",
|
| 226 |
"torch_dtype": "float32",
|
| 227 |
"transformers_version": "4.46.2",
|
| 228 |
"type_vocab_size": 2,
|
|
|
|
| 1 |
{
|
| 2 |
"_name_or_path": "bert-large-uncased",
|
| 3 |
"architectures": [
|
| 4 |
+
"DualTaskBert"
|
| 5 |
],
|
| 6 |
"attention_probs_dropout_prob": 0.1,
|
| 7 |
"classifier_dropout": null,
|
|
|
|
| 9 |
"hidden_act": "gelu",
|
| 10 |
"hidden_dropout_prob": 0.1,
|
| 11 |
"hidden_size": 1024,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 12 |
"initializer_range": 0.02,
|
| 13 |
"intermediate_size": 4096,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 14 |
"layer_norm_eps": 1e-12,
|
| 15 |
"max_position_embeddings": 512,
|
| 16 |
"model_type": "bert",
|
|
|
|
| 18 |
"num_hidden_layers": 24,
|
| 19 |
"pad_token_id": 0,
|
| 20 |
"position_embedding_type": "absolute",
|
|
|
|
| 21 |
"torch_dtype": "float32",
|
| 22 |
"transformers_version": "4.46.2",
|
| 23 |
"type_vocab_size": 2,
|
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 sha256:05fb0bae3808a8fc3953d896ceb0958ecd3f6b7b8cf0a0363f43464b35342f76
|
| 3 |
+
size 1341147752
|
runs/Apr03_14-46-46_raptor/events.out.tfevents.1743671811.raptor.286607.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:700657376e46cc17c0543cb33004fa18b01f7552db0783e6bbaea1240cdc5c3b
|
| 3 |
+
size 5135
|
runs/Apr03_14-47-54_raptor/events.out.tfevents.1743671876.raptor.286988.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:16a941ee11a832770b53ed4ddb68843ccecb0078df71809387aef17c9578dcce
|
| 3 |
+
size 5483
|
tokenizer.json
CHANGED
|
@@ -1,7 +1,21 @@
|
|
| 1 |
{
|
| 2 |
"version": "1.0",
|
| 3 |
-
"truncation":
|
| 4 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5 |
"added_tokens": [
|
| 6 |
{
|
| 7 |
"id": 0,
|
|
|
|
| 1 |
{
|
| 2 |
"version": "1.0",
|
| 3 |
+
"truncation": {
|
| 4 |
+
"direction": "Right",
|
| 5 |
+
"max_length": 512,
|
| 6 |
+
"strategy": "LongestFirst",
|
| 7 |
+
"stride": 0
|
| 8 |
+
},
|
| 9 |
+
"padding": {
|
| 10 |
+
"strategy": {
|
| 11 |
+
"Fixed": 512
|
| 12 |
+
},
|
| 13 |
+
"direction": "Right",
|
| 14 |
+
"pad_to_multiple_of": null,
|
| 15 |
+
"pad_id": 0,
|
| 16 |
+
"pad_type_id": 0,
|
| 17 |
+
"pad_token": "[PAD]"
|
| 18 |
+
},
|
| 19 |
"added_tokens": [
|
| 20 |
{
|
| 21 |
"id": 0,
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 5304
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:92cf605d054cd75c75af0145f32a38c117ca27cf3dec777517d8b869406178bb
|
| 3 |
size 5304
|