first model version

config.json

@@ -0,0 +1,139 @@
+{
+  "_name_or_path": "/home/user/naepli/noisepp/gsw_pos_best/DEbert_swisscrawl_ftDE-xpos-noise",
+  "architectures": [
+    "BertForTokenClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "classifier_dropout": null,
+  "finetuning_task": "ner",
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "$(",
+    "1": "$,",
+    "2": "$.",
+    "3": "ADJA",
+    "4": "ADJD",
+    "5": "ADV",
+    "6": "APPO",
+    "7": "APPR",
+    "8": "APPRART",
+    "9": "APZR",
+    "10": "ART",
+    "11": "CARD",
+    "12": "FM",
+    "13": "ITJ",
+    "14": "KOKOM",
+    "15": "KON",
+    "16": "KOUI",
+    "17": "KOUS",
+    "18": "NE",
+    "19": "NN",
+    "20": "PDAT",
+    "21": "PDS",
+    "22": "PIAT",
+    "23": "PIDAT",
+    "24": "PIS",
+    "25": "PPER",
+    "26": "PPOSAT",
+    "27": "PPOSS",
+    "28": "PRELAT",
+    "29": "PRELS",
+    "30": "PRF",
+    "31": "PROAV",
+    "32": "PTKA",
+    "33": "PTKANT",
+    "34": "PTKNEG",
+    "35": "PTKVZ",
+    "36": "PTKZU",
+    "37": "PWAT",
+    "38": "PWAV",
+    "39": "PWS",
+    "40": "TRUNC",
+    "41": "VAFIN",
+    "42": "VAIMP",
+    "43": "VAINF",
+    "44": "VAPP",
+    "45": "VMFIN",
+    "46": "VMINF",
+    "47": "VMPP",
+    "48": "VVFIN",
+    "49": "VVIMP",
+    "50": "VVINF",
+    "51": "VVIZU",
+    "52": "VVPP",
+    "53": "XY"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "$(": 0,
+    "$,": 1,
+    "$.": 2,
+    "ADJA": 3,
+    "ADJD": 4,
+    "ADV": 5,
+    "APPO": 6,
+    "APPR": 7,
+    "APPRART": 8,
+    "APZR": 9,
+    "ART": 10,
+    "CARD": 11,
+    "FM": 12,
+    "ITJ": 13,
+    "KOKOM": 14,
+    "KON": 15,
+    "KOUI": 16,
+    "KOUS": 17,
+    "NE": 18,
+    "NN": 19,
+    "PDAT": 20,
+    "PDS": 21,
+    "PIAT": 22,
+    "PIDAT": 23,
+    "PIS": 24,
+    "PPER": 25,
+    "PPOSAT": 26,
+    "PPOSS": 27,
+    "PRELAT": 28,
+    "PRELS": 29,
+    "PRF": 30,
+    "PROAV": 31,
+    "PTKA": 32,
+    "PTKANT": 33,
+    "PTKNEG": 34,
+    "PTKVZ": 35,
+    "PTKZU": 36,
+    "PWAT": 37,
+    "PWAV": 38,
+    "PWS": 39,
+    "TRUNC": 40,
+    "VAFIN": 41,
+    "VAIMP": 42,
+    "VAINF": 43,
+    "VAPP": 44,
+    "VMFIN": 45,
+    "VMINF": 46,
+    "VMPP": 47,
+    "VVFIN": 48,
+    "VVIMP": 49,
+    "VVINF": 50,
+    "VVIZU": 51,
+    "VVPP": 52,
+    "XY": 53
+  },
+  "layer_norm_eps": 1e-12,
+  "max_position_embeddings": 512,
+  "model_type": "bert",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "output_past": true,
+  "pad_token_id": 0,
+  "position_embedding_type": "absolute",
+  "torch_dtype": "float32",
+  "transformers_version": "4.25.0.dev0",
+  "type_vocab_size": 2,
+  "use_cache": true,
+  "vocab_size": 31102
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5b5130babacfc8b11f22774d275e98e9c4a2805bf7f3823f7b549f3c093b8f84
+size 437586605

special_tokens_map.json

@@ -0,0 +1,7 @@
+{
+  "cls_token": "[CLS]",
+  "mask_token": "[MASK]",
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "unk_token": "[UNK]"
+}

tokenizer_config.json

@@ -0,0 +1,16 @@
+{
+  "cls_token": "[CLS]",
+  "do_basic_tokenize": true,
+  "do_lower_case": false,
+  "mask_token": "[MASK]",
+  "max_len": 512,
+  "name_or_path": "/home/user/naepli/noisepp/gsw_pos_best/DEbert_swisscrawl_ftDE-xpos-noise",
+  "never_split": null,
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "special_tokens_map_file": null,
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "BertTokenizer",
+  "unk_token": "[UNK]"
+}

trainer_state.json

@@ -0,0 +1,367 @@
+{
+  "best_metric": 0.2828419506549835,
+  "best_model_checkpoint": "//home/user/naepli/noisepp/gsw_pos_best/DEbert_swisscrawl_ftDE-xpos-noise_ftGSW_v2/checkpoint-10000",
+  "epoch": 5.0,
+  "global_step": 14725,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.8302207130730054e-05,
+      "loss": 0.1584,
+      "step": 500
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.6604414261460106e-05,
+      "loss": 0.1276,
+      "step": 1000
+    },
+    {
+      "epoch": 0.34,
+      "eval_accuracy": 0.9205043369399982,
+      "eval_f1": 0.9128971962616822,
+      "eval_loss": 0.3129468858242035,
+      "eval_precision": 0.9137511693171187,
+      "eval_recall": 0.9120448179271708,
+      "eval_runtime": 2.3351,
+      "eval_samples_per_second": 313.471,
+      "eval_steps_per_second": 39.398,
+      "step": 1000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.490662139219016e-05,
+      "loss": 0.1257,
+      "step": 1500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.320882852292021e-05,
+      "loss": 0.1137,
+      "step": 2000
+    },
+    {
+      "epoch": 0.68,
+      "eval_accuracy": 0.9290887954931593,
+      "eval_f1": 0.9225095696013444,
+      "eval_loss": 0.28120967745780945,
+      "eval_precision": 0.9224234503360716,
+      "eval_recall": 0.9225957049486462,
+      "eval_runtime": 2.3612,
+      "eval_samples_per_second": 310.007,
+      "eval_steps_per_second": 38.963,
+      "step": 2000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.1511035653650255e-05,
+      "loss": 0.1123,
+      "step": 2500
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3.981324278438031e-05,
+      "loss": 0.102,
+      "step": 3000
+    },
+    {
+      "epoch": 1.02,
+      "eval_accuracy": 0.9374049897165341,
+      "eval_f1": 0.930929808994536,
+      "eval_loss": 0.2500903308391571,
+      "eval_precision": 0.9312342333925068,
+      "eval_recall": 0.9306255835667601,
+      "eval_runtime": 1.3009,
+      "eval_samples_per_second": 562.677,
+      "eval_steps_per_second": 70.719,
+      "step": 3000
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.811544991511036e-05,
+      "loss": 0.067,
+      "step": 3500
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 3.641765704584041e-05,
+      "loss": 0.0674,
+      "step": 4000
+    },
+    {
+      "epoch": 1.36,
+      "eval_accuracy": 0.9362425109541268,
+      "eval_f1": 0.9298311094522721,
+      "eval_loss": 0.25322550535202026,
+      "eval_precision": 0.9292241700857888,
+      "eval_recall": 0.9304388422035481,
+      "eval_runtime": 1.2863,
+      "eval_samples_per_second": 569.086,
+      "eval_steps_per_second": 71.524,
+      "step": 4000
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 3.471986417657046e-05,
+      "loss": 0.0659,
+      "step": 4500
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 3.3022071307300515e-05,
+      "loss": 0.0671,
+      "step": 5000
+    },
+    {
+      "epoch": 1.7,
+      "eval_accuracy": 0.9383886255924171,
+      "eval_f1": 0.9333582403213002,
+      "eval_loss": 0.257250040769577,
+      "eval_precision": 0.9336634588433149,
+      "eval_recall": 0.9330532212885154,
+      "eval_runtime": 2.4588,
+      "eval_samples_per_second": 297.711,
+      "eval_steps_per_second": 37.417,
+      "step": 5000
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 3.132427843803056e-05,
+      "loss": 0.0683,
+      "step": 5500
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 2.9626485568760615e-05,
+      "loss": 0.0547,
+      "step": 6000
+    },
+    {
+      "epoch": 2.04,
+      "eval_accuracy": 0.9407135831172315,
+      "eval_f1": 0.9349024002988698,
+      "eval_loss": 0.2575545907020569,
+      "eval_precision": 0.9351644245142003,
+      "eval_recall": 0.934640522875817,
+      "eval_runtime": 1.2898,
+      "eval_samples_per_second": 567.536,
+      "eval_steps_per_second": 71.33,
+      "step": 6000
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 2.7928692699490667e-05,
+      "loss": 0.0356,
+      "step": 6500
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 2.6230899830220713e-05,
+      "loss": 0.0368,
+      "step": 7000
+    },
+    {
+      "epoch": 2.38,
+      "eval_accuracy": 0.9416972189931145,
+      "eval_f1": 0.9363886685023568,
+      "eval_loss": 0.2715625464916229,
+      "eval_precision": 0.9360828590090511,
+      "eval_recall": 0.9366946778711485,
+      "eval_runtime": 1.2849,
+      "eval_samples_per_second": 569.686,
+      "eval_steps_per_second": 71.6,
+      "step": 7000
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 2.4533106960950768e-05,
+      "loss": 0.0367,
+      "step": 7500
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 2.2835314091680816e-05,
+      "loss": 0.0383,
+      "step": 8000
+    },
+    {
+      "epoch": 2.72,
+      "eval_accuracy": 0.9413395332200661,
+      "eval_f1": 0.9358543417366947,
+      "eval_loss": 0.2626907527446747,
+      "eval_precision": 0.9358543417366947,
+      "eval_recall": 0.9358543417366947,
+      "eval_runtime": 1.3044,
+      "eval_samples_per_second": 561.17,
+      "eval_steps_per_second": 70.53,
+      "step": 8000
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 2.1137521222410865e-05,
+      "loss": 0.0381,
+      "step": 8500
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 1.9439728353140917e-05,
+      "loss": 0.032,
+      "step": 9000
+    },
+    {
+      "epoch": 3.06,
+      "eval_accuracy": 0.9429491191987839,
+      "eval_f1": 0.9374211338038043,
+      "eval_loss": 0.2770785689353943,
+      "eval_precision": 0.938429868064003,
+      "eval_recall": 0.9364145658263305,
+      "eval_runtime": 1.3011,
+      "eval_samples_per_second": 562.587,
+      "eval_steps_per_second": 70.708,
+      "step": 9000
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 1.774193548387097e-05,
+      "loss": 0.0203,
+      "step": 9500
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 1.6044142614601017e-05,
+      "loss": 0.0195,
+      "step": 10000
+    },
+    {
+      "epoch": 3.4,
+      "eval_accuracy": 0.9438433336314048,
+      "eval_f1": 0.9383257855175311,
+      "eval_loss": 0.2828419506549835,
+      "eval_precision": 0.9383695956671958,
+      "eval_recall": 0.93828197945845,
+      "eval_runtime": 1.3079,
+      "eval_samples_per_second": 559.696,
+      "eval_steps_per_second": 70.344,
+      "step": 10000
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 1.434634974533107e-05,
+      "loss": 0.0176,
+      "step": 10500
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 1.2648556876061121e-05,
+      "loss": 0.0186,
+      "step": 11000
+    },
+    {
+      "epoch": 3.74,
+      "eval_accuracy": 0.9435750693016185,
+      "eval_f1": 0.9380398748657608,
+      "eval_loss": 0.30645471811294556,
+      "eval_precision": 0.9381712898104044,
+      "eval_recall": 0.9379084967320261,
+      "eval_runtime": 1.2883,
+      "eval_samples_per_second": 568.19,
+      "eval_steps_per_second": 71.412,
+      "step": 11000
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 1.0950764006791173e-05,
+      "loss": 0.0195,
+      "step": 11500
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 9.252971137521223e-06,
+      "loss": 0.0146,
+      "step": 12000
+    },
+    {
+      "epoch": 4.07,
+      "eval_accuracy": 0.9447375480640258,
+      "eval_f1": 0.9395509499136442,
+      "eval_loss": 0.3018830716609955,
+      "eval_precision": 0.939419396994306,
+      "eval_recall": 0.9396825396825397,
+      "eval_runtime": 1.2869,
+      "eval_samples_per_second": 568.801,
+      "eval_steps_per_second": 71.489,
+      "step": 12000
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 7.5551782682512745e-06,
+      "loss": 0.0104,
+      "step": 12500
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 5.857385398981325e-06,
+      "loss": 0.0088,
+      "step": 13000
+    },
+    {
+      "epoch": 4.41,
+      "eval_accuracy": 0.9442010194044532,
+      "eval_f1": 0.9391718407170534,
+      "eval_loss": 0.31471893191337585,
+      "eval_precision": 0.9391279992531043,
+      "eval_recall": 0.9392156862745098,
+      "eval_runtime": 1.283,
+      "eval_samples_per_second": 570.548,
+      "eval_steps_per_second": 71.708,
+      "step": 13000
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 4.159592529711375e-06,
+      "loss": 0.0091,
+      "step": 13500
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 2.461799660441426e-06,
+      "loss": 0.01,
+      "step": 14000
+    },
+    {
+      "epoch": 4.75,
+      "eval_accuracy": 0.9451846552803362,
+      "eval_f1": 0.939825404976425,
+      "eval_loss": 0.3139691650867462,
+      "eval_precision": 0.9397815330034544,
+      "eval_recall": 0.9398692810457516,
+      "eval_runtime": 1.3004,
+      "eval_samples_per_second": 562.918,
+      "eval_steps_per_second": 70.749,
+      "step": 14000
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 7.640067911714771e-07,
+      "loss": 0.0091,
+      "step": 14500
+    },
+    {
+      "epoch": 5.0,
+      "step": 14725,
+      "total_flos": 3264317950832928.0,
+      "train_loss": 0.05123547306287471,
+      "train_runtime": 906.0685,
+      "train_samples_per_second": 130.001,
+      "train_steps_per_second": 16.252
+    }
+  ],
+  "max_steps": 14725,
+  "num_train_epochs": 5,
+  "total_flos": 3264317950832928.0,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e0fffcdbbd109cc6eeccf4444f4fa5af5f4a3e614f99dba1d60e95397725adeb
+size 3643