Upload 16 files

README.md

@@ -1,3 +1,52 @@
----
-license: mit
----
+---
+tags:
+- generated_from_trainer
+datasets:
+- PiC/phrase_sense_disambiguation
+model-index:
+- name: finetuned
+  results: []
+---
+
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
+
+# finetuned
+
+This model is a fine-tuned version of [../results/phrase_retrieval/PR-pass/qa/allenai/longformer-base-4096/finetuned](https://huggingface.co/../results/phrase_retrieval/PR-pass/qa/allenai/longformer-base-4096/finetuned) on the PiC/phrase_sense_disambiguation  dataset.
+
+## Model description
+
+More information needed
+
+## Intended uses & limitations
+
+More information needed
+
+## Training and evaluation data
+
+More information needed
+
+## Training procedure
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+- learning_rate: 3e-05
+- train_batch_size: 2
+- eval_batch_size: 2
+- seed: 42
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: linear
+- num_epochs: 2.0
+
+### Training results
+
+
+
+### Framework versions
+
+- Transformers 4.20.1
+- Pytorch 1.10.0+cu113
+- Datasets 2.3.2
+- Tokenizers 0.12.1

all_results.json

@@ -0,0 +1,11 @@
+{
+    "epoch": 2.0,
+    "eval_exact_match": 69.39999999999999,
+    "eval_f1": 69.43,
+    "eval_samples": 500,
+    "train_loss": 0.9066677579012784,
+    "train_runtime": 893.6642,
+    "train_samples": 1650,
+    "train_samples_per_second": 3.693,
+    "train_steps_per_second": 1.846
+}

config.json

@@ -0,0 +1,45 @@
+{
+  "_name_or_path": "../results/phrase_retrieval/PR-pass/qa/allenai/longformer-base-4096/finetuned",
+  "architectures": [
+    "LongformerForQuestionAnswering"
+  ],
+  "attention_mode": "longformer",
+  "attention_probs_dropout_prob": 0.1,
+  "attention_window": [
+    512,
+    512,
+    512,
+    512,
+    512,
+    512,
+    512,
+    512,
+    512,
+    512,
+    512,
+    512
+  ],
+  "bos_token_id": 0,
+  "classifier_dropout": null,
+  "eos_token_id": 2,
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "ignore_attention_mask": false,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 4098,
+  "model_type": "longformer",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "pad_token_id": 1,
+  "position_embedding_type": "absolute",
+  "sep_token_id": 2,
+  "torch_dtype": "float32",
+  "transformers_version": "4.20.1",
+  "type_vocab_size": 1,
+  "use_cache": true,
+  "vocab_size": 50265
+}

eval_predictions.json

@@ -0,0 +1,502 @@
+{
+    "6023-2": "better shot",
+    "8982-1": "fast release",
+    "8163-1": "immediate background",
+    "4019-2": "second column",
+    "1969-1": "maintenance work",
+    "847-2": "active person",
+    "12419-2": "last bar",
+    "5047-1": "street signs",
+    "9656-1": "greatest exposure",
+    "6861-2": "sudden death",
+    "9301-1": "quick run",
+    "8792-2": "free hit",
+    "12610-1": "Spring Rain",
+    "9645-1": "delivery model",
+    "14275-2": "warning signal",
+    "14699-2": "later conflict",
+    "1173-1": "occupied state",
+    "8990-2": "fledgling system",
+    "2100-1": "whole power",
+    "7224-2": "resting time",
+    "6198-1": "open book",
+    "4424-2": "altered production",
+    "12127-2": "film material",
+    "9371-2": "order forms",
+    "4373-2": "clear dividing line",
+    "12929-1": "bright light",
+    "14667-2": "Take Care",
+    "3361-1": "official guardian",
+    "9507-2": "top center",
+    "12411-1": "entire state",
+    "1889-1": "joint defense",
+    "10998-2": "regular period",
+    "6613-2": "universal manner",
+    "4376-2": "first hearing",
+    "11356-2": "single source",
+    "8130-2": "good value",
+    "8684-2": "sufficient material",
+    "8969-1": "principal types",
+    "8757-1": "one place",
+    "14613-2": "significant trail",
+    "3888-1": "greater clarity",
+    "3809-2": "physical performance",
+    "7088-1": "formal level",
+    "7192-1": "Local records",
+    "702-1": "second property",
+    "14056-1": "long relationship",
+    "13251-2": "clear position",
+    "8432-1": "similar record",
+    "4044-1": "live nights",
+    "8539-1": "religious organization",
+    "1206-1": "desired piece",
+    "5936-2": "standard character",
+    "8762-1": "bottom table",
+    "2812-2": "good links",
+    "10527-1": "independent operation",
+    "9940-2": "former shape",
+    "13640-1": "local assembly",
+    "1910-2": "supply water",
+    "14373-1": "US study",
+    "13707-1": "full order",
+    "8648-2": "visual direction",
+    "3743-1": "homecoming ceremony",
+    "7902-2": "political track",
+    "3207-1": "1st person",
+    "5964-2": "first seat",
+    "8704-1": "repayment period",
+    "3529-1": "common site",
+    "14975-1": "solid support",
+    "14545-1": "major order",
+    "11072-2": "secondary concern",
+    "4610-1": "boiling point",
+    "6740-1": "central activity",
+    "14278-1": "one \"unit",
+    "13627-2": "key tracks",
+    "8911-1": "small association",
+    "317-2": "death party",
+    "1746-2": "possible future development",
+    "14898-1": "big brand",
+    "10977-1": "second development",
+    "8244-2": "processing station",
+    "3992-1": "consistent exercise",
+    "14530-2": "fluid movement",
+    "14488-1": "whole cluster",
+    "6898-1": "regular school",
+    "14501-1": "natural limit",
+    "14165-2": "major west",
+    "11835-1": "larger body",
+    "9139-1": "notable event",
+    "4446-2": "healing force",
+    "11629-1": "entire channel",
+    "13755-1": "simple table",
+    "8523-1": "ideal unit",
+    "2530-2": "varying nature",
+    "14162-2": "first stroke",
+    "365-2": "long thread",
+    "7413-1": "new flagship",
+    "4560-2": "Winter Games",
+    "4381-1": "party tickets",
+    "4411-1": "another stem",
+    "8191-2": "general nature",
+    "3356-2": "hand positions",
+    "719-1": "force requirements",
+    "2023-1": "single side",
+    "376-2": "highly complex work",
+    "14432-2": "Current service",
+    "14943-2": "contact point",
+    "7823-2": "similar matter",
+    "13030-1": "14 and 16 year",
+    "11275-2": "deep division",
+    "2655-2": "one ex",
+    "8523-2": "ideal unit",
+    "9006-1": "last word",
+    "6674-2": "entire load",
+    "13889-2": "final top",
+    "5009-1": "job site",
+    "8130-1": "good value",
+    "8463-1": "two shoulder",
+    "864-1": "rough shape",
+    "624-2": "golden rule",
+    "11509-1": "metal material",
+    "3634-2": "education structure",
+    "5541-2": "next unit",
+    "12425-2": "local circuit",
+    "4313-2": "single spring",
+    "8848-1": "current subject",
+    "10521-1": "stronger claim",
+    "5972-2": "New Mexico game",
+    "4304-1": "first turn",
+    "5914-1": "true sense",
+    "8614-1": "one subdivision",
+    "2734-1": "Initial word",
+    "5314-2": "given decision",
+    "8785-1": "longest performance",
+    "9813-1": "reasonable level",
+    "14241-2": "immediate creation",
+    "3947-1": "useful material",
+    "3861-2": "one pack",
+    "11410-2": "one mass",
+    "12563-2": "Modern point",
+    "947-1": "early interception",
+    "14518-1": "modern port",
+    "4053-1": "informal economy",
+    "3846-2": "later piece",
+    "7133-1": "original volume",
+    "2905-2": "sufficient capacity",
+    "12791-2": "teaching practice",
+    "10565-2": "open structure",
+    "936-1": "latest party",
+    "5301-1": "black suit",
+    "12746-2": "initial activity",
+    "987-2": "stronger line",
+    "14073-1": "similar power",
+    "6750-1": "direct sight",
+    "11557-2": "ideal figure",
+    "10572-2": "practice level",
+    "8626-2": "common division",
+    "14262-1": "One Way Glass",
+    "14132-2": "one chart",
+    "11151-2": "available program",
+    "9817-2": "Greek system",
+    "9956-2": "specific connection",
+    "8967-2": "one mini",
+    "4908-2": "one possession",
+    "5626-2": "alternate point",
+    "4419-2": "transitional season",
+    "1953-1": "large spread",
+    "10901-1": "much voice",
+    "14349-2": "clear drop",
+    "8315-1": "precise formation",
+    "11420-1": "strong division",
+    "9348-1": "2018 update",
+    "3744-1": "main selection",
+    "2997-1": "landmark point",
+    "8213-1": "food dish",
+    "9037-2": "political deliberation",
+    "387-1": "direct pressure",
+    "3465-1": "random time",
+    "14275-1": "warning signal",
+    "7095-1": "popular voice",
+    "12200-1": "full leg",
+    "6694-2": "long stage",
+    "9791-1": "good exposure",
+    "14354-2": "absolute gift",
+    "417-1": "significant gap",
+    "9599-1": "large beam",
+    "12147-2": "live one",
+    "3077-1": "public desire",
+    "3845-2": "single unit",
+    "389-1": "last bit",
+    "6090-1": "well furnished",
+    "9486-2": "significant fall",
+    "8769-2": "Another resolution",
+    "11686-1": "Field's life",
+    "14541-1": "single sentence",
+    "2923-1": "better range",
+    "8065-1": "right approach",
+    "12662-2": "government advice",
+    "8661-2": "physical exchange",
+    "6160-2": "usual language",
+    "8580-2": "full structure",
+    "10177-1": "precious time",
+    "2851-2": "work measures",
+    "10935-1": "powerful spring",
+    "4181-2": "2D model",
+    "7139-1": "last two cases",
+    "4313-1": "single spring",
+    "433-1": "actual activity",
+    "2183-2": "major hunt",
+    "4140-2": "quick recognition",
+    "9077-2": "property types",
+    "8910-1": "solid center",
+    "7588-1": "strong movement",
+    "9215-1": "site visits",
+    "11029-1": "regular practice",
+    "14410-1": "first suggestion",
+    "14211-1": "stronger line",
+    "8895-2": "common source",
+    "3397-1": "certain condition",
+    "4230-1": "common program",
+    "719-2": "force requirements",
+    "14224-1": "controversial temperament",
+    "12144-1": "specialised nature",
+    "8990-1": "fledgling system",
+    "3465-2": "random time",
+    "4453-2": "certain evidence",
+    "4604-2": "text line",
+    "8108-1": "whole scene",
+    "9656-2": "greatest exposure",
+    "6241-2": "open framework",
+    "4700-1": "second case",
+    "13660-2": "release effort",
+    "7265-2": "overall position",
+    "9693-1": "Commercial material",
+    "9348-2": "2010 study",
+    "14196-1": "previous host",
+    "7107-1": "functional level",
+    "11350-2": "Activity Period",
+    "721-1": "quick rate",
+    "4373-1": "clear dividing line",
+    "6825-1": "broader market",
+    "1128-2": "initial degree",
+    "8441-1": "network's channels",
+    "14263-1": "unusual variation",
+    "14966-1": "exact expression",
+    "5138-2": "particular account",
+    "7445-1": "varying strength",
+    "3390-1": "narrow waist",
+    "4315-1": "quicker response",
+    "7925-2": "decisive point",
+    "25-1": "similar finish",
+    "10466-1": "one good",
+    "3812-2": "predictable course",
+    "2355-2": "primary season",
+    "8792-1": "free hit",
+    "3398-2": "certain condition",
+    "7503-1": "top figure",
+    "6694-1": "long stage",
+    "8065-2": "right approach",
+    "174-1": "normal number",
+    "6971-2": "Heritage groups",
+    "14643-2": "bright line",
+    "6556-2": "stop signal",
+    "10661-2": "field position",
+    "14914-2": "negative state",
+    "6086-1": "suitable ground",
+    "8011-1": "whole parts",
+    "9424-2": "corresponding range",
+    "8133-1": "fixed meeting place",
+    "13587-1": "heart land",
+    "3370-2": "quick transfer",
+    "489-1": "particular bond",
+    "14437-1": "whole club",
+    "3897-2": "5th stage",
+    "3366-1": "Community stage",
+    "12467-2": "another bar",
+    "10138-2": "single square",
+    "1668-2": "exceptional first season",
+    "13860-1": "labour time",
+    "6880-2": "full round",
+    "7837-1": "first component",
+    "11313-1": "primary class",
+    "14943-1": "contact point",
+    "12461-1": "personal identification",
+    "3220-1": "even empty shells",
+    "10059-1": "quick passing",
+    "191-1": "different case",
+    "936-2": "latest party",
+    "12579-1": "founder group",
+    "7926-2": "desired character",
+    "6014-2": "strong movement",
+    "9628-1": "higher place",
+    "8708-1": "key goods",
+    "14845-1": "default position",
+    "14234-1": "lower extension",
+    "8667-1": "surgical division",
+    "6257-2": "long cross",
+    "9920-2": "attached agency",
+    "399-2": "full contact",
+    "9700-1": "separate seat",
+    "2401-2": "light parts",
+    "14771-1": "bridge foundation",
+    "4155-2": "general point",
+    "13689-1": "wrong base",
+    "8766-1": "released material",
+    "14353-2": "cut line",
+    "4042-1": "archbishop's chair",
+    "13531-1": "common focus",
+    "3287-1": "German post",
+    "4482-2": "known culture",
+    "10718-2": "average looks",
+    "4600-1": "actual coverage",
+    "3356-1": "hand positions",
+    "4463-2": "second view",
+    "6572-1": "whole trouble",
+    "4865-1": "main class",
+    "572-2": "house members",
+    "11612-2": "general limitation",
+    "323-1": "modern work",
+    "10138-1": "single square",
+    "6030-2": "drum design",
+    "3259-1": "specific judge",
+    "11305-1": "normal home",
+    "14964-2": "Language Blocks",
+    "2455-1": "top five",
+    "13640-2": "Local assembly",
+    "12634-2": "particular property",
+    "3620-2": "zero sum game",
+    "14429-1": "new case",
+    "9133-2": "second radio",
+    "1018-2": "regular period",
+    "9645-2": "delivery model",
+    "14276-2": "fine position",
+    "6715-2": "normal action",
+    "8756-1": "second function",
+    "13506-2": "popular convention",
+    "14041-1": "general treatment",
+    "4436-2": "weaver family",
+    "13641-1": "next number",
+    "8149-2": "light fire",
+    "559-1": "little return",
+    "3451-1": "Dutch tool",
+    "14612-1": "external defense",
+    "13718-1": "current body",
+    "9494-2": "fourth line",
+    "3067-1": "Public Work",
+    "12776-1": "distribution side",
+    "2687-2": "actual figure",
+    "5185-2": "first master",
+    "6880-1": "full round",
+    "14473-2": "strict manner",
+    "4545-2": "formal office",
+    "9272-2": "increasing wave",
+    "3706-1": "consistent pattern",
+    "12195-2": "exact time period",
+    "9934-2": "small association",
+    "14549-2": "value conflicts",
+    "14458-1": "lightning strokes",
+    "10952-2": "form groups",
+    "4419-1": "transitional season",
+    "10141-2": "cold comfort",
+    "6233-2": "subsequent test",
+    "9337-1": "Standby power",
+    "14437-2": "whole club",
+    "3961-1": "remaining force",
+    "13553-1": "major breaches",
+    "12692-1": "property laws",
+    "3348-1": "unique series",
+    "14845-2": "default position",
+    "9324-1": "simple matter",
+    "8999-2": "every score",
+    "9020-2": "natural monopoly",
+    "3740-1": "official returns",
+    "14245-2": "lead female",
+    "5914-2": "true sense",
+    "3364-2": "set rate",
+    "8795-1": "lengthy set",
+    "11831-2": "following wave",
+    "4397-2": "one attribute",
+    "11063-1": "fine place",
+    "423-2": "existing class",
+    "14084-2": "land sharks",
+    "12723-1": "heavy rock",
+    "2724-1": "different play",
+    "9545-2": "rich world",
+    "9959-2": "particular action",
+    "9211-1": "overall capital",
+    "8637-1": "good record",
+    "12418-2": "fusion center",
+    "2692-1": "one piece",
+    "2544-1": "local action",
+    "4491-2": "sudden collapse",
+    "12817-1": "building houses",
+    "2139-2": "right version",
+    "10379-2": "third body",
+    "9705-1": "important spot",
+    "4170-1": "bicycle club",
+    "10706-1": "play structure",
+    "12325-2": "state transition",
+    "9744-2": "offensive manner",
+    "8930-1": "extra help",
+    "7051-1": "main vehicle",
+    "9210-2": "fast movements",
+    "6610-2": "small length",
+    "14476-1": "local players",
+    "9364-1": "defensive line",
+    "3863-1": "live one",
+    "882-2": "foundation experience",
+    "14026-1": "high concentrate",
+    "14361-2": "cheap way",
+    "3295-1": "regular period",
+    "3004-2": "open path",
+    "6399-2": "planned home",
+    "9165-1": "hollow area",
+    "4206-2": "total interest",
+    "3204-2": "Another address",
+    "14079-2": "power tools",
+    "10887-2": "fine thread",
+    "11569-1": "final frame",
+    "186-2": "Official surveys",
+    "12022-2": "immediate neighborhood",
+    "6613-1": "universal manner",
+    "5290-1": "official limits",
+    "11627-2": "major stature",
+    "9879-2": "former charge",
+    "1306-2": "negative material",
+    "6871-1": "stronger tone",
+    "8587-1": "modern restoration",
+    "13896-1": "one theatre",
+    "14615-1": "technical quality",
+    "14073-2": "similar power",
+    "11071-2": "secondary concern",
+    "2061-1": "old one",
+    "11467-2": "main test",
+    "11461-1": "desired tone",
+    "12147-1": "live one",
+    "11276-1": "quality production",
+    "4491-1": "sudden collapse",
+    "7313-1": "term \"body",
+    "14443-2": "local camps",
+    "10762-1": "original panel",
+    "8722-1": "good measure",
+    "7711-1": "high loads",
+    "2077-2": "popular track",
+    "4716-1": "real nature",
+    "14229-1": "possible range",
+    "8678-2": "longer course",
+    "8317-1": "licence area",
+    "10223-1": "long sleep",
+    "4287-2": "undocumented population",
+    "2314-1": "different unit",
+    "8389-2": "final disappearance",
+    "7589-2": "1 credit",
+    "962-1": "actual board",
+    "4146-2": "woman's side",
+    "2492-1": "get to know",
+    "7782-2": "either year",
+    "11009-2": "present breed",
+    "8819-2": "direct source",
+    "3459-1": "2010/11 campaign",
+    "3730-1": "complete protection",
+    "4604-1": "text line",
+    "8515-2": "significant foundation",
+    "7297-2": "notable facility",
+    "1313-1": "big cut",
+    "11072-1": "secondary concern",
+    "11232-1": "similar notice",
+    "4152-1": "full one",
+    "8219-1": "central process",
+    "14019-2": "similar direction",
+    "9375-1": "special meaning",
+    "11355-2": "sufficient interest",
+    "7727-2": "one classification",
+    "8516-2": "empty position",
+    "4181-1": "2D model",
+    "9237-1": "initial character",
+    "5041-2": "initial strain",
+    "4610-2": "boiling point",
+    "3715-1": "campaign success",
+    "4449-2": "major use",
+    "4395-2": "larger body",
+    "2622-2": "least one entry",
+    "7861-1": "correct accounts",
+    "6809-2": "good links",
+    "10239-1": "critical point",
+    "14839-2": "shared background",
+    "5632-2": "round table",
+    "8715-2": "entire test",
+    "3437-2": "community contributions",
+    "13956-1": "second release in 1973 with an orange sleeve. In both versions, the first side contained 10 stories narrated by Laura Olsher, complete with sound effects",
+    "2812-1": "good links",
+    "8198-2": "smaller sum",
+    "13843-2": "final sum",
+    "1098-2": "whole race",
+    "8765-1": "central reason",
+    "4216-1": "entire case",
+    "9762-2": "another edition",
+    "9554-2": "construction stage",
+    "1485-2": "official cables",
+    "6026-2": "ticker tape",
+    "679-1": "undefined point"
+}

eval_results.json

@@ -0,0 +1,6 @@
+{
+    "epoch": 2.0,
+    "eval_exact_match": 69.39999999999999,
+    "eval_f1": 69.43,
+    "eval_samples": 500
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:341e2423dc7b10c4c66b0cebdd8a7135cb3b4569a77189bfa8e8b59db0c2a06b
+size 592407201

special_tokens_map.json

@@ -0,0 +1,15 @@
+{
+  "bos_token": "<s>",
+  "cls_token": "<s>",
+  "eos_token": "</s>",
+  "mask_token": {
+    "content": "<mask>",
+    "lstrip": true,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<pad>",
+  "sep_token": "</s>",
+  "unk_token": "<unk>"
+}

tokenizer_config.json

@@ -0,0 +1,16 @@
+{
+  "add_prefix_space": false,
+  "bos_token": "<s>",
+  "cls_token": "<s>",
+  "eos_token": "</s>",
+  "errors": "replace",
+  "mask_token": "<mask>",
+  "model_max_length": 4096,
+  "name_or_path": "../results/phrase_retrieval/PR-pass/qa/allenai/longformer-base-4096/finetuned",
+  "pad_token": "<pad>",
+  "sep_token": "</s>",
+  "special_tokens_map_file": null,
+  "tokenizer_class": "LongformerTokenizer",
+  "trim_offsets": true,
+  "unk_token": "<unk>"
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 2.0,
+    "train_loss": 0.9066677579012784,
+    "train_runtime": 893.6642,
+    "train_samples": 1650,
+    "train_samples_per_second": 3.693,
+    "train_steps_per_second": 1.846
+}

trainer_state.json

@@ -0,0 +1,43 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.0,
+  "global_step": 1650,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.61,
+      "learning_rate": 2.090909090909091e-05,
+      "loss": 1.2525,
+      "step": 500
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 1.1818181818181819e-05,
+      "loss": 1.0478,
+      "step": 1000
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 2.7272727272727272e-06,
+      "loss": 0.5,
+      "step": 1500
+    },
+    {
+      "epoch": 2.0,
+      "step": 1650,
+      "total_flos": 8622543790080000.0,
+      "train_loss": 0.9066677579012784,
+      "train_runtime": 893.6642,
+      "train_samples_per_second": 3.693,
+      "train_steps_per_second": 1.846
+    }
+  ],
+  "max_steps": 1650,
+  "num_train_epochs": 2,
+  "total_flos": 8622543790080000.0,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aa622283c41db520921d00be29991b71b3a97d624437552ec2210f1e82901f31
+size 3375