add models

checkpoint-3000/config.json

@@ -0,0 +1,281 @@
+{
+  "_name_or_path": "distilbert-base-uncased",
+  "activation": "gelu",
+  "architectures": [
+    "DistilBertForSequenceClassification"
+  ],
+  "attention_dropout": 0.1,
+  "dim": 768,
+  "dropout": 0.1,
+  "hidden_dim": 3072,
+  "id2label": {
+    "0": "cs.AI",
+    "1": "cs.CL",
+    "2": "stat.ML",
+    "3": "cs.CV",
+    "4": "cs.LG",
+    "5": "cs.NE",
+    "6": "cs.RO",
+    "7": "cs.IR",
+    "8": "cs.IT",
+    "9": "astro-ph.IM",
+    "10": "cs.DB",
+    "11": "cs.SE",
+    "12": "q-bio.NC",
+    "13": "eess.IV",
+    "14": "cs.SI",
+    "15": "cs.CR",
+    "16": "cs.SD",
+    "17": "cs.NI",
+    "18": "cond-mat.dis-nn",
+    "19": "math.OC",
+    "20": "math.NA",
+    "21": "eess.SP",
+    "22": "cs.CC",
+    "23": "cs.MM",
+    "24": "cs.DC",
+    "25": "physics.med-ph",
+    "26": "cs.HC",
+    "27": "cs.CY",
+    "28": "cs.FL",
+    "29": "q-bio.QM",
+    "30": "eess.AS",
+    "31": "quant-ph",
+    "32": "q-bio.BM",
+    "33": "q-fin.CP",
+    "34": "cs.SY",
+    "35": "stat.AP",
+    "36": "cs.NA",
+    "37": "cs.GT",
+    "38": "physics.optics",
+    "39": "gr-qc",
+    "40": "stat.CO",
+    "41": "physics.comp-ph",
+    "42": "math.ST",
+    "43": "nlin.AO",
+    "44": "cs.DS",
+    "45": "cs.DM",
+    "46": "cs.MA",
+    "47": "astro-ph",
+    "48": "cs.CE",
+    "49": "cs.AR",
+    "50": "cs.GR",
+    "51": "astro-ph.CO",
+    "52": "cs.ET",
+    "53": "q-fin.ST",
+    "54": "physics.data-an",
+    "55": "stat.ME",
+    "56": "math.PR",
+    "57": "cs.DL",
+    "58": "physics.soc-ph",
+    "59": "nlin.CG",
+    "60": "q-fin.PM",
+    "61": "cmp-lg",
+    "62": "cond-mat.stat-mech",
+    "63": "cs.OH",
+    "64": "math.DS",
+    "65": "econ.EM",
+    "66": "cs.PF",
+    "67": "math.FA",
+    "68": "math.AP",
+    "69": "astro-ph.GA",
+    "70": "math.GM",
+    "71": "q-bio.GN",
+    "72": "cs.LO",
+    "73": "cs.PL",
+    "74": "math.CT",
+    "75": "cs.MS",
+    "76": "q-bio.PE",
+    "77": "math.CO",
+    "78": "math.GT",
+    "79": "cond-mat.mtrl-sci",
+    "80": "physics.chem-ph",
+    "81": "cs.CG",
+    "82": "math.DG",
+    "83": "q-bio.MN",
+    "84": "q-fin.TR",
+    "85": "hep-ex",
+    "86": "astro-ph.EP",
+    "87": "math.HO",
+    "88": "hep-th",
+    "89": "math.AT",
+    "90": "physics.bio-ph",
+    "91": "nucl-th",
+    "92": "hep-ph",
+    "93": "math.MG",
+    "94": "hep-lat",
+    "95": "math.LO",
+    "96": "physics.class-ph",
+    "97": "q-bio.CB",
+    "98": "cond-mat.supr-con",
+    "99": "physics.ao-ph",
+    "100": "nlin.CD",
+    "101": "physics.gen-ph",
+    "102": "adap-org",
+    "103": "cond-mat",
+    "104": "cond-mat.other",
+    "105": "math.GR",
+    "106": "physics.ins-det",
+    "107": "nlin.PS",
+    "108": "cs.SC",
+    "109": "q-fin.RM",
+    "110": "stat.OT",
+    "111": "cs.GL",
+    "112": "physics.hist-ph",
+    "113": "math.NT",
+    "114": "q-fin.GN",
+    "115": "q-fin.EC",
+    "116": "cs.OS",
+    "117": "cond-mat.soft",
+    "118": "math.AG",
+    "119": "math.CA",
+    "120": "math.RT",
+    "121": "math.RA",
+    "122": "physics.geo-ph",
+    "123": "astro-ph.SR",
+    "124": "q-bio.TO",
+    "125": "math.GN"
+  },
+  "initializer_range": 0.02,
+  "label2id": {
+    "adap-org": 102,
+    "astro-ph": 47,
+    "astro-ph.CO": 51,
+    "astro-ph.EP": 86,
+    "astro-ph.GA": 69,
+    "astro-ph.IM": 9,
+    "astro-ph.SR": 123,
+    "cmp-lg": 61,
+    "cond-mat": 103,
+    "cond-mat.dis-nn": 18,
+    "cond-mat.mtrl-sci": 79,
+    "cond-mat.other": 104,
+    "cond-mat.soft": 117,
+    "cond-mat.stat-mech": 62,
+    "cond-mat.supr-con": 98,
+    "cs.AI": 0,
+    "cs.AR": 49,
+    "cs.CC": 22,
+    "cs.CE": 48,
+    "cs.CG": 81,
+    "cs.CL": 1,
+    "cs.CR": 15,
+    "cs.CV": 3,
+    "cs.CY": 27,
+    "cs.DB": 10,
+    "cs.DC": 24,
+    "cs.DL": 57,
+    "cs.DM": 45,
+    "cs.DS": 44,
+    "cs.ET": 52,
+    "cs.FL": 28,
+    "cs.GL": 111,
+    "cs.GR": 50,
+    "cs.GT": 37,
+    "cs.HC": 26,
+    "cs.IR": 7,
+    "cs.IT": 8,
+    "cs.LG": 4,
+    "cs.LO": 72,
+    "cs.MA": 46,
+    "cs.MM": 23,
+    "cs.MS": 75,
+    "cs.NA": 36,
+    "cs.NE": 5,
+    "cs.NI": 17,
+    "cs.OH": 63,
+    "cs.OS": 116,
+    "cs.PF": 66,
+    "cs.PL": 73,
+    "cs.RO": 6,
+    "cs.SC": 108,
+    "cs.SD": 16,
+    "cs.SE": 11,
+    "cs.SI": 14,
+    "cs.SY": 34,
+    "econ.EM": 65,
+    "eess.AS": 30,
+    "eess.IV": 13,
+    "eess.SP": 21,
+    "gr-qc": 39,
+    "hep-ex": 85,
+    "hep-lat": 94,
+    "hep-ph": 92,
+    "hep-th": 88,
+    "math.AG": 118,
+    "math.AP": 68,
+    "math.AT": 89,
+    "math.CA": 119,
+    "math.CO": 77,
+    "math.CT": 74,
+    "math.DG": 82,
+    "math.DS": 64,
+    "math.FA": 67,
+    "math.GM": 70,
+    "math.GN": 125,
+    "math.GR": 105,
+    "math.GT": 78,
+    "math.HO": 87,
+    "math.LO": 95,
+    "math.MG": 93,
+    "math.NA": 20,
+    "math.NT": 113,
+    "math.OC": 19,
+    "math.PR": 56,
+    "math.RA": 121,
+    "math.RT": 120,
+    "math.ST": 42,
+    "nlin.AO": 43,
+    "nlin.CD": 100,
+    "nlin.CG": 59,
+    "nlin.PS": 107,
+    "nucl-th": 91,
+    "physics.ao-ph": 99,
+    "physics.bio-ph": 90,
+    "physics.chem-ph": 80,
+    "physics.class-ph": 96,
+    "physics.comp-ph": 41,
+    "physics.data-an": 54,
+    "physics.gen-ph": 101,
+    "physics.geo-ph": 122,
+    "physics.hist-ph": 112,
+    "physics.ins-det": 106,
+    "physics.med-ph": 25,
+    "physics.optics": 38,
+    "physics.soc-ph": 58,
+    "q-bio.BM": 32,
+    "q-bio.CB": 97,
+    "q-bio.GN": 71,
+    "q-bio.MN": 83,
+    "q-bio.NC": 12,
+    "q-bio.PE": 76,
+    "q-bio.QM": 29,
+    "q-bio.TO": 124,
+    "q-fin.CP": 33,
+    "q-fin.EC": 115,
+    "q-fin.GN": 114,
+    "q-fin.PM": 60,
+    "q-fin.RM": 109,
+    "q-fin.ST": 53,
+    "q-fin.TR": 84,
+    "quant-ph": 31,
+    "stat.AP": 35,
+    "stat.CO": 40,
+    "stat.ME": 55,
+    "stat.ML": 2,
+    "stat.OT": 110
+  },
+  "max_position_embeddings": 512,
+  "model_type": "distilbert",
+  "n_heads": 12,
+  "n_layers": 6,
+  "pad_token_id": 0,
+  "problem_type": "single_label_classification",
+  "qa_dropout": 0.1,
+  "seq_classif_dropout": 0.2,
+  "sinusoidal_pos_embds": false,
+  "tie_weights_": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.23.1",
+  "vocab_size": 30522
+}

checkpoint-3000/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:934c31a7b760170b73f429c3928e4be27dab07dd40108a84529e2b61a71f9ba7
+size 268234289

checkpoint-3000/special_tokens_map.json

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

checkpoint-3000/tokenizer_config.json

@@ -0,0 +1,14 @@
+{
+  "cls_token": "[CLS]",
+  "do_lower_case": true,
+  "mask_token": "[MASK]",
+  "model_max_length": 512,
+  "name_or_path": "distilbert-base-uncased",
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "special_tokens_map_file": null,
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "DistilBertTokenizer",
+  "unk_token": "[UNK]"
+}

checkpoint-3000/trainer_state.json

@@ -0,0 +1,52 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.1779661016949152,
+  "global_step": 3000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.978813559322034e-05,
+      "loss": 2.0034,
+      "step": 500
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.9576271186440678e-05,
+      "loss": 1.3659,
+      "step": 1000
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 1.9364406779661017e-05,
+      "loss": 1.2089,
+      "step": 1500
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.9152542372881357e-05,
+      "loss": 1.1463,
+      "step": 2000
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.8940677966101697e-05,
+      "loss": 1.0185,
+      "step": 2500
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 1.8728813559322033e-05,
+      "loss": 0.9664,
+      "step": 3000
+    }
+  ],
+  "max_steps": 47200,
+  "num_train_epochs": 50,
+  "total_flos": 9490525857741888.0,
+  "trial_name": null,
+  "trial_params": null
+}