Training in progress, step 500

config.json

@@ -0,0 +1,330 @@
+{
+  "_name_or_path": "allenai/scibert_scivocab_cased",
+  "architectures": [
+    "BertForSequenceClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "classifier_dropout": null,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "cs.AI",
+    "1": "cs.CL",
+    "2": "cs.CV",
+    "3": "cs.NE",
+    "4": "stat.ML",
+    "5": "cs.LG",
+    "6": "physics.soc-ph",
+    "7": "stat.AP",
+    "8": "cs.RO",
+    "9": "cs.SE",
+    "10": "cs.MA",
+    "11": "math.OC",
+    "12": "cs.IR",
+    "13": "cond-mat.dis-nn",
+    "14": "stat.ME",
+    "15": "physics.chem-ph",
+    "16": "cs.DC",
+    "17": "stat.CO",
+    "18": "q-bio.NC",
+    "19": "cs.GT",
+    "20": "cs.MM",
+    "21": "cs.CG",
+    "22": "cs.CR",
+    "23": "cs.HC",
+    "24": "cs.GL",
+    "25": "eess.AS",
+    "26": "cs.SD",
+    "27": "math.DS",
+    "28": "cs.GR",
+    "29": "math.NA",
+    "30": "cs.CY",
+    "31": "physics.data-an",
+    "32": "math.ST",
+    "33": "stat.TH",
+    "34": "cs.IT",
+    "35": "math.IT",
+    "36": "quant-ph",
+    "37": "astro-ph.IM",
+    "38": "astro-ph.GA",
+    "39": "cs.SI",
+    "40": "cs.DB",
+    "41": "cs.LO",
+    "42": "nlin.AO",
+    "43": "cs.PF",
+    "44": "cs.ET",
+    "45": "eess.IV",
+    "46": "cs.AR",
+    "47": "cs.SY",
+    "48": "cs.CC",
+    "49": "q-bio.BM",
+    "50": "q-bio.QM",
+    "51": "cs.NI",
+    "52": "cs.DS",
+    "53": "cond-mat.stat-mech",
+    "54": "cs.NA",
+    "55": "cs.DM",
+    "56": "eess.SP",
+    "57": "cs.MS",
+    "58": "physics.med-ph",
+    "59": "physics.optics",
+    "60": "q-fin.CP",
+    "61": "cs.FL",
+    "62": "cs.SC",
+    "63": "q-fin.EC",
+    "64": "q-fin.TR",
+    "65": "cond-mat.mes-hall",
+    "66": "math.PR",
+    "67": "q-fin.RM",
+    "68": "nlin.CD",
+    "69": "cs.CE",
+    "70": "math.AT",
+    "71": "stat.OT",
+    "72": "physics.ao-ph",
+    "73": "math.SP",
+    "74": "cs.PL",
+    "75": "math.AP",
+    "76": "math.FA",
+    "77": "gr-qc",
+    "78": "physics.geo-ph",
+    "79": "q-bio.TO",
+    "80": "physics.comp-ph",
+    "81": "cs.DL",
+    "82": "math.CO",
+    "83": "physics.flu-dyn",
+    "84": "math.MG",
+    "85": "astro-ph.EP",
+    "86": "q-bio.CB",
+    "87": "hep-th",
+    "88": "astro-ph",
+    "89": "math.RA",
+    "90": "q-bio",
+    "91": "astro-ph.CO",
+    "92": "cond-mat.mtrl-sci",
+    "93": "q-fin.ST",
+    "94": "q-bio.GN",
+    "95": "hep-ex",
+    "96": "nlin.CG",
+    "97": "nlin.PS",
+    "98": "math.HO",
+    "99": "q-fin.GN",
+    "100": "math.LO",
+    "101": "math.CT",
+    "102": "cs.CL, cs.AI, math.CT",
+    "103": "q-bio.PE",
+    "104": "astro-ph.SR",
+    "105": "q-fin.PM",
+    "106": "physics.bio-ph",
+    "107": "math.AG",
+    "108": "cs.OH",
+    "109": "math.DG",
+    "110": "astro-ph.HE",
+    "111": "econ.EM",
+    "112": "math.QA",
+    "113": "q-bio.SC",
+    "114": "math.GM",
+    "115": "q-bio.MN",
+    "116": "math.GT",
+    "117": "math.AC",
+    "118": "math.CA",
+    "119": "cond-mat.str-el",
+    "120": "math.GN",
+    "121": "hep-ph",
+    "122": "aaai.org",
+    "123": "cond-mat.supr-con",
+    "124": "q-bio.OT",
+    "125": "nucl-th",
+    "126": "physics.ins-det",
+    "127": "hep-lat",
+    "128": "physics.app-ph",
+    "129": "math.RT",
+    "130": "cond-mat",
+    "131": "math-ph",
+    "132": "math.MP",
+    "133": "physics.class-ph",
+    "134": "q-fin.PR",
+    "135": "physics.space-ph",
+    "136": "physics.gen-ph",
+    "137": "cond-mat.other",
+    "138": "math.GR",
+    "139": "nucl-ex",
+    "140": "cond-mat.quant-gas",
+    "141": "math.OA",
+    "142": "physics.hist-ph",
+    "143": "math.NT",
+    "144": "cs.OS",
+    "145": "cond-mat.soft",
+    "146": "q-bio.BM, q-bio.MN, q-bio.NC, nlin.AO, nlin.CD",
+    "147": "nlin.AO, nlin.CD, q-bio.NC, physics.bio-ph, cond-mat.dis-nn",
+    "148": "physics.pop-ph",
+    "149": "math.CV"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "aaai.org": 122,
+    "astro-ph": 88,
+    "astro-ph.CO": 91,
+    "astro-ph.EP": 85,
+    "astro-ph.GA": 38,
+    "astro-ph.HE": 110,
+    "astro-ph.IM": 37,
+    "astro-ph.SR": 104,
+    "cond-mat": 130,
+    "cond-mat.dis-nn": 13,
+    "cond-mat.mes-hall": 65,
+    "cond-mat.mtrl-sci": 92,
+    "cond-mat.other": 137,
+    "cond-mat.quant-gas": 140,
+    "cond-mat.soft": 145,
+    "cond-mat.stat-mech": 53,
+    "cond-mat.str-el": 119,
+    "cond-mat.supr-con": 123,
+    "cs.AI": 0,
+    "cs.AR": 46,
+    "cs.CC": 48,
+    "cs.CE": 69,
+    "cs.CG": 21,
+    "cs.CL": 1,
+    "cs.CL, cs.AI, math.CT": 102,
+    "cs.CR": 22,
+    "cs.CV": 2,
+    "cs.CY": 30,
+    "cs.DB": 40,
+    "cs.DC": 16,
+    "cs.DL": 81,
+    "cs.DM": 55,
+    "cs.DS": 52,
+    "cs.ET": 44,
+    "cs.FL": 61,
+    "cs.GL": 24,
+    "cs.GR": 28,
+    "cs.GT": 19,
+    "cs.HC": 23,
+    "cs.IR": 12,
+    "cs.IT": 34,
+    "cs.LG": 5,
+    "cs.LO": 41,
+    "cs.MA": 10,
+    "cs.MM": 20,
+    "cs.MS": 57,
+    "cs.NA": 54,
+    "cs.NE": 3,
+    "cs.NI": 51,
+    "cs.OH": 108,
+    "cs.OS": 144,
+    "cs.PF": 43,
+    "cs.PL": 74,
+    "cs.RO": 8,
+    "cs.SC": 62,
+    "cs.SD": 26,
+    "cs.SE": 9,
+    "cs.SI": 39,
+    "cs.SY": 47,
+    "econ.EM": 111,
+    "eess.AS": 25,
+    "eess.IV": 45,
+    "eess.SP": 56,
+    "gr-qc": 77,
+    "hep-ex": 95,
+    "hep-lat": 127,
+    "hep-ph": 121,
+    "hep-th": 87,
+    "math-ph": 131,
+    "math.AC": 117,
+    "math.AG": 107,
+    "math.AP": 75,
+    "math.AT": 70,
+    "math.CA": 118,
+    "math.CO": 82,
+    "math.CT": 101,
+    "math.CV": 149,
+    "math.DG": 109,
+    "math.DS": 27,
+    "math.FA": 76,
+    "math.GM": 114,
+    "math.GN": 120,
+    "math.GR": 138,
+    "math.GT": 116,
+    "math.HO": 98,
+    "math.IT": 35,
+    "math.LO": 100,
+    "math.MG": 84,
+    "math.MP": 132,
+    "math.NA": 29,
+    "math.NT": 143,
+    "math.OA": 141,
+    "math.OC": 11,
+    "math.PR": 66,
+    "math.QA": 112,
+    "math.RA": 89,
+    "math.RT": 129,
+    "math.SP": 73,
+    "math.ST": 32,
+    "nlin.AO": 42,
+    "nlin.AO, nlin.CD, q-bio.NC, physics.bio-ph, cond-mat.dis-nn": 147,
+    "nlin.CD": 68,
+    "nlin.CG": 96,
+    "nlin.PS": 97,
+    "nucl-ex": 139,
+    "nucl-th": 125,
+    "physics.ao-ph": 72,
+    "physics.app-ph": 128,
+    "physics.bio-ph": 106,
+    "physics.chem-ph": 15,
+    "physics.class-ph": 133,
+    "physics.comp-ph": 80,
+    "physics.data-an": 31,
+    "physics.flu-dyn": 83,
+    "physics.gen-ph": 136,
+    "physics.geo-ph": 78,
+    "physics.hist-ph": 142,
+    "physics.ins-det": 126,
+    "physics.med-ph": 58,
+    "physics.optics": 59,
+    "physics.pop-ph": 148,
+    "physics.soc-ph": 6,
+    "physics.space-ph": 135,
+    "q-bio": 90,
+    "q-bio.BM": 49,
+    "q-bio.BM, q-bio.MN, q-bio.NC, nlin.AO, nlin.CD": 146,
+    "q-bio.CB": 86,
+    "q-bio.GN": 94,
+    "q-bio.MN": 115,
+    "q-bio.NC": 18,
+    "q-bio.OT": 124,
+    "q-bio.PE": 103,
+    "q-bio.QM": 50,
+    "q-bio.SC": 113,
+    "q-bio.TO": 79,
+    "q-fin.CP": 60,
+    "q-fin.EC": 63,
+    "q-fin.GN": 99,
+    "q-fin.PM": 105,
+    "q-fin.PR": 134,
+    "q-fin.RM": 67,
+    "q-fin.ST": 93,
+    "q-fin.TR": 64,
+    "quant-ph": 36,
+    "stat.AP": 7,
+    "stat.CO": 17,
+    "stat.ME": 14,
+    "stat.ML": 4,
+    "stat.OT": 71,
+    "stat.TH": 33
+  },
+  "layer_norm_eps": 1e-12,
+  "max_position_embeddings": 512,
+  "model_type": "bert",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "pad_token_id": 0,
+  "position_embedding_type": "absolute",
+  "problem_type": "single_label_classification",
+  "torch_dtype": "float32",
+  "transformers_version": "4.43.4",
+  "type_vocab_size": 2,
+  "use_cache": true,
+  "vocab_size": 31116
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6207b4e429a9aab305b6baf10ce5afc9034990bfa814056445c4e1da8a8d379b
+size 440238664

runs/Apr06_15-14-22_t4-4eee8d2a-2d33-34ff-95a5-9e673c42f315/events.out.tfevents.1775488464.t4-4eee8d2a-2d33-34ff-95a5-9e673c42f315.3211.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ee978a8adc6ae642ac7d18e2332c799b6c00a0ceb443e770c7f1e8a493d6685c
+size 11523

runs/Apr06_15-16-52_t4-4eee8d2a-2d33-34ff-95a5-9e673c42f315/events.out.tfevents.1775488612.t4-4eee8d2a-2d33-34ff-95a5-9e673c42f315.3211.3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20c5a04c6ed9b1d796d80620d9814a9756edb5924048631e20d2a9f207bc4bd6
+size 8280

runs/Apr06_15-18-36_t4-4eee8d2a-2d33-34ff-95a5-9e673c42f315/events.out.tfevents.1775488717.t4-4eee8d2a-2d33-34ff-95a5-9e673c42f315.8803.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4ca9b14c0a47acb3f2a2b719eaa4888235e7f5fa59eaa460d8fcc1b189f1973d
+size 11523

runs/Apr06_15-20-11_t4-4eee8d2a-2d33-34ff-95a5-9e673c42f315/events.out.tfevents.1775488812.t4-4eee8d2a-2d33-34ff-95a5-9e673c42f315.8803.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dae2199944611d000eb7b382f02f305d7bd94141fbca6693f1e93ea417106e42
+size 8280

runs/Apr06_15-22-19_t4-4eee8d2a-2d33-34ff-95a5-9e673c42f315/events.out.tfevents.1775488940.t4-4eee8d2a-2d33-34ff-95a5-9e673c42f315.9302.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:619fda3bee127963273e00253655fcecee50cbb3f9395a76759caa7274693d0b
+size 12057

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 +1,57 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100": {
+      "content": "[UNK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "101": {
+      "content": "[CLS]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "102": {
+      "content": "[SEP]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "103": {
+      "content": "[MASK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "clean_up_tokenization_spaces": true,
+  "cls_token": "[CLS]",
+  "do_basic_tokenize": true,
+  "do_lower_case": true,
+  "mask_token": "[MASK]",
+  "model_max_length": 1000000000000000019884624838656,
+  "never_split": null,
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "BertTokenizer",
+  "unk_token": "[UNK]"
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77c99fc3a67e2e4098ccb8ed5d2936aeed71c804c3ced4681c5ddc1e0d0af171
+size 4731