Add dataset map configuration

app/config/dataset_map.json

@@ -0,0 +1,416 @@
+{
+  "datasets": {
+    "causal-lm": [
+      {
+        "id": "wikitext",
+        "name": "WikiText",
+        "description": "Wikipedia text dataset for language modeling",
+        "configs": ["wikitext-2-raw-v1", "wikitext-103-raw-v1"],
+        "splits": ["train", "validation", "test"],
+        "text_column": "text",
+        "default_config": "wikitext-2-raw-v1",
+        "size_categories": ["100K-1M", "1M-10M"],
+        "recommended": true
+      },
+      {
+        "id": "openwebtext",
+        "name": "OpenWebText",
+        "description": "Open source recreation of WebText dataset",
+        "configs": [],
+        "splits": ["train"],
+        "text_column": "text",
+        "size_categories": [">10M"],
+        "recommended": true
+      },
+      {
+        "id": "the_pile",
+        "name": "The Pile",
+        "description": "Large-scale text corpus for language modeling",
+        "configs": ["all", "enron_emails", "europarl", "hacker_news", "pubmed", "ubuntu_irc"],
+        "splits": ["train", "validation", "test"],
+        "text_column": "text",
+        "size_categories": [">10M"],
+        "recommended": false
+      },
+      {
+        "id": "c4",
+        "name": "C4 (Colossal Clean Crawled Corpus)",
+        "description": "Huge cleaned web text dataset",
+        "configs": ["en", "realnewslike", "en.noblocklist", "en.noclean"],
+        "splits": ["train", "validation"],
+        "text_column": "text",
+        "size_categories": [">10M"],
+        "recommended": false
+      },
+      {
+        "id": "tiny_shakespeare",
+        "name": "Tiny Shakespeare",
+        "description": "Small Shakespeare text for quick testing",
+        "configs": [],
+        "splits": ["train", "validation", "test"],
+        "text_column": "text",
+        "size_categories": ["<10K"],
+        "recommended": true
+      }
+    ],
+    "seq2seq": [
+      {
+        "id": "cnn_dailymail",
+        "name": "CNN/DailyMail",
+        "description": "News article summarization dataset",
+        "configs": ["1.0.0", "2.0.0", "3.0.0"],
+        "splits": ["train", "validation", "test"],
+        "text_column": "article",
+        "label_column": "highlights",
+        "default_config": "3.0.0",
+        "size_categories": ["100K-1M"],
+        "recommended": true
+      },
+      {
+        "id": "xsum",
+        "name": "XSum",
+        "description": "BBC article summarization",
+        "configs": [],
+        "splits": ["train", "validation", "test"],
+        "text_column": "document",
+        "label_column": "summary",
+        "size_categories": ["10K-100K"],
+        "recommended": true
+      },
+      {
+        "id": "samsum",
+        "name": "SAMSum",
+        "description": "Dialogue summarization dataset",
+        "configs": [],
+        "splits": ["train", "validation", "test"],
+        "text_column": "dialogue",
+        "label_column": "summary",
+        "size_categories": ["10K-100K"],
+        "recommended": true
+      },
+      {
+        "id": "wmt16",
+        "name": "WMT16 Translation",
+        "description": "Machine translation dataset",
+        "configs": ["de-en", "en-de", "ro-en", "en-ro", "cs-en", "en-cs"],
+        "splits": ["train", "validation", "test"],
+        "size_categories": ["1M-10M"],
+        "recommended": false
+      },
+      {
+        "id": "billsum",
+        "name": "BillSum",
+        "description": "US Congressional bill summarization",
+        "configs": [],
+        "splits": ["train", "test"],
+        "text_column": "text",
+        "label_column": "summary",
+        "size_categories": ["10K-100K"],
+        "recommended": true
+      }
+    ],
+    "token-classification": [
+      {
+        "id": "conll2003",
+        "name": "CoNLL-2003",
+        "description": "Named entity recognition dataset",
+        "configs": [],
+        "splits": ["train", "validation", "test"],
+        "text_column": "tokens",
+        "label_column": "ner_tags",
+        "labels": ["O", "B-PER", "I-PER", "B-ORG", "I-ORG", "B-LOC", "I-LOC", "B-MISC", "I-MISC"],
+        "size_categories": ["10K-100K"],
+        "recommended": true
+      },
+      {
+        "id": "wnut_17",
+        "name": "WNUT 17",
+        "description": "Emerging entity recognition from social media",
+        "configs": [],
+        "splits": ["train", "validation", "test"],
+        "text_column": "tokens",
+        "label_column": "ner_tags",
+        "labels": ["O", "B-corporation", "B-creative-work", "B-group", "B-location", "B-person", "B-product", "I-corporation", "I-creative-work", "I-group", "I-location", "I-person", "I-product"],
+        "size_categories": ["<10K"],
+        "recommended": true
+      },
+      {
+        "id": "ontonotes5",
+        "name": "OntoNotes 5.0",
+        "description": "Multi-genre NER and coreference",
+        "configs": ["english_v4", "english_v12", "chinese_v4", "arabic_v4"],
+        "splits": ["train", "validation", "test"],
+        "text_column": "document",
+        "label_column": "named_entities",
+        "size_categories": ["100K-1M"],
+        "recommended": false
+      }
+    ],
+    "text-classification": [
+      {
+        "id": "imdb",
+        "name": "IMDB",
+        "description": "Movie review sentiment classification",
+        "configs": [],
+        "splits": ["train", "test", "unsupervised"],
+        "text_column": "text",
+        "label_column": "label",
+        "labels": ["negative", "positive"],
+        "size_categories": ["10K-100K"],
+        "recommended": true
+      },
+      {
+        "id": "yelp_polarity",
+        "name": "Yelp Polarity",
+        "description": "Yelp review sentiment classification",
+        "configs": [],
+        "splits": ["train", "test"],
+        "text_column": "text",
+        "label_column": "label",
+        "labels": ["negative", "positive"],
+        "size_categories": ["100K-1M"],
+        "recommended": true
+      },
+      {
+        "id": "ag_news",
+        "name": "AG News",
+        "description": "News article categorization",
+        "configs": [],
+        "splits": ["train", "test"],
+        "text_column": "text",
+        "label_column": "label",
+        "labels": ["World", "Sports", "Business", "Sci/Tech"],
+        "size_categories": ["100K-1M"],
+        "recommended": true
+      },
+      {
+        "id": "glue",
+        "name": "GLUE",
+        "description": "General Language Understanding Evaluation",
+        "configs": ["cola", "mnli", "mnli_matched", "mnli_mismatched", "mrpc", "qnli", "qqp", "rte", "sst2", "stsb", "wnli"],
+        "splits": ["train", "validation", "test"],
+        "size_categories": ["varies"],
+        "recommended": true
+      },
+      {
+        "id": "emotion",
+        "name": "Emotion",
+        "description": "Twitter emotion classification",
+        "configs": [],
+        "splits": ["train", "validation", "test"],
+        "text_column": "text",
+        "label_column": "label",
+        "labels": ["sadness", "joy", "love", "anger", "fear", "surprise"],
+        "size_categories": ["10K-100K"],
+        "recommended": true
+      }
+    ],
+    "question-answering": [
+      {
+        "id": "squad",
+        "name": "SQuAD",
+        "description": "Stanford Question Answering Dataset",
+        "configs": ["plain_text"],
+        "splits": ["train", "validation"],
+        "text_column": "context",
+        "question_column": "question",
+        "answer_column": "answers",
+        "size_categories": ["10K-100K"],
+        "recommended": true
+      },
+      {
+        "id": "squad_v2",
+        "name": "SQuAD 2.0",
+        "description": "SQuAD with unanswerable questions",
+        "configs": ["squad_v2"],
+        "splits": ["train", "validation"],
+        "size_categories": ["100K-1M"],
+        "recommended": true
+      },
+      {
+        "id": "natural_questions",
+        "name": "Natural Questions",
+        "description": "Real user questions with Wikipedia answers",
+        "configs": ["default"],
+        "splits": ["train", "validation"],
+        "size_categories": [">10M"],
+        "recommended": false
+      },
+      {
+        "id": "coqa",
+        "name": "CoQA",
+        "description": "Conversational Question Answering",
+        "configs": [],
+        "splits": ["train", "validation"],
+        "size_categories": ["100K-1M"],
+        "recommended": true
+      }
+    ],
+    "translation": [
+      {
+        "id": "wmt14",
+        "name": "WMT14 Translation",
+        "description": "Large-scale machine translation",
+        "configs": ["de-en", "en-de", "fr-en", "en-fr"],
+        "splits": ["train", "validation", "test"],
+        "size_categories": [">10M"],
+        "recommended": false
+      },
+      {
+        "id": "opus100",
+        "name": "OPUS-100",
+        "description": "Multi-lingual parallel corpus",
+        "configs": ["en-de", "en-fr", "en-es", "en-ru", "en-zh"],
+        "splits": ["train", "validation", "test"],
+        "size_categories": ["1M-10M"],
+        "recommended": true
+      }
+    ],
+    "image-classification": [
+      {
+        "id": "cifar10",
+        "name": "CIFAR-10",
+        "description": "10-class image classification",
+        "configs": [],
+        "splits": ["train", "test"],
+        "image_column": "img",
+        "label_column": "label",
+        "labels": ["airplane", "automobile", "bird", "cat", "deer", "dog", "frog", "horse", "ship", "truck"],
+        "size_categories": ["10K-100K"],
+        "recommended": true
+      },
+      {
+        "id": "imagenet-1k",
+        "name": "ImageNet-1k",
+        "description": "Large-scale image classification",
+        "configs": [],
+        "splits": ["train", "validation"],
+        "image_column": "image",
+        "label_column": "label",
+        "size_categories": [">10M"],
+        "recommended": false
+      }
+    ]
+  },
+  "models": {
+    "causal-lm": {
+      "small": [
+        {"id": "gpt2", "params": "124M", "recommended": true},
+        {"id": "distilgpt2", "params": "82M", "recommended": true},
+        {"id": "EleutherAI/gpt-neo-125M", "params": "125M", "recommended": true},
+        {"id": "bigscience/bloom-560m", "params": "560M", "recommended": true}
+      ],
+      "medium": [
+        {"id": "gpt2-medium", "params": "355M", "recommended": true},
+        {"id": "gpt2-large", "params": "774M", "recommended": true},
+        {"id": "EleutherAI/gpt-neo-1.3B", "params": "1.3B", "recommended": true},
+        {"id": "EleutherAI/gpt-j-6b", "params": "6B", "recommended": false},
+        {"id": "bigscience/bloom-1b7", "params": "1.7B", "recommended": true},
+        {"id": "meta-llama/Llama-2-7b-hf", "params": "7B", "recommended": true},
+        {"id": "mistralai/Mistral-7B-v0.1", "params": "7B", "recommended": true}
+      ],
+      "large": [
+        {"id": "EleutherAI/gpt-neox-20b", "params": "20B", "recommended": false},
+        {"id": "bigscience/bloom", "params": "176B", "recommended": false},
+        {"id": "meta-llama/Llama-2-13b-hf", "params": "13B", "recommended": false},
+        {"id": "meta-llama/Llama-2-70b-hf", "params": "70B", "recommended": false}
+      ]
+    },
+    "seq2seq": {
+      "small": [
+        {"id": "google-t5/t5-small", "params": "60M", "recommended": true},
+        {"id": "facebook/bart-base", "params": "140M", "recommended": true},
+        {"id": "google/flan-t5-small", "params": "80M", "recommended": true}
+      ],
+      "medium": [
+        {"id": "google-t5/t5-base", "params": "220M", "recommended": true},
+        {"id": "facebook/bart-large", "params": "400M", "recommended": true},
+        {"id": "google/flan-t5-base", "params": "250M", "recommended": true},
+        {"id": "google/flan-t5-large", "params": "780M", "recommended": true},
+        {"id": "google-t5/t5-large", "params": "770M", "recommended": true}
+      ],
+      "large": [
+        {"id": "google-t5/t5-3b", "params": "3B", "recommended": false},
+        {"id": "google/flan-t5-xl", "params": "3B", "recommended": false},
+        {"id": "facebook/bart-large-cnn", "params": "400M", "recommended": true}
+      ]
+    },
+    "token-classification": {
+      "small": [
+        {"id": "dslim/bert-base-NER", "params": "110M", "recommended": true},
+        {"id": "dslim/distilbert-NER", "params": "66M", "recommended": true},
+        {"id": "dbmdz/bert-large-cased-finetuned-conll03-english", "params": "340M", "recommended": true}
+      ],
+      "medium": [
+        {"id": "dslim/bert-base-NER", "params": "110M", "recommended": true},
+        {"id": "elastic/distilbert-base-uncased-finetuned-conll03-english", "params": "66M", "recommended": true}
+      ]
+    },
+    "text-classification": {
+      "small": [
+        {"id": "distilbert/distilbert-base-uncased", "params": "66M", "recommended": true},
+        {"id": "google-bert/bert-base-uncased", "params": "110M", "recommended": true},
+        {"id": "roberta-base", "params": "125M", "recommended": true}
+      ],
+      "medium": [
+        {"id": "google-bert/bert-large-uncased", "params": "340M", "recommended": true},
+        {"id": "roberta-large", "params": "355M", "recommended": true},
+        {"id": "microsoft/deberta-v3-base", "params": "184M", "recommended": true}
+      ]
+    },
+    "question-answering": {
+      "small": [
+        {"id": "distilbert/distilbert-base-uncased-distilled-squad", "params": "66M", "recommended": true},
+        {"id": "deepset/minilm-uncased-squad2", "params": "33M", "recommended": true}
+      ],
+      "medium": [
+        {"id": "deepset/roberta-base-squad2", "params": "125M", "recommended": true},
+        {"id": "google-bert/bert-large-uncased-whole-word-masking-finetuned-squad", "params": "340M", "recommended": true}
+      ]
+    }
+  },
+  "task_metadata": {
+    "causal-lm": {
+      "display_name": "Causal Language Modeling",
+      "description": "Generate text, autocomplete, story writing",
+      "icon": "text_fields",
+      "metrics": ["perplexity", "accuracy", "f1"],
+      "requires_decoder_only": true
+    },
+    "seq2seq": {
+      "display_name": "Sequence-to-Sequence",
+      "description": "Summarization, translation, paraphrase",
+      "icon": "compare_arrows",
+      "metrics": ["rouge1", "rouge2", "rougeL", "bleu", "meteor"],
+      "requires_encoder_decoder": true
+    },
+    "token-classification": {
+      "display_name": "Token Classification",
+      "description": "Named entity recognition, POS tagging",
+      "icon": "label",
+      "metrics": ["precision", "recall", "f1", "accuracy"],
+      "requires_encoder": true
+    },
+    "text-classification": {
+      "display_name": "Text Classification",
+      "description": "Sentiment analysis, topic classification",
+      "icon": "category",
+      "metrics": ["accuracy", "f1", "precision", "recall"],
+      "requires_encoder": true
+    },
+    "question-answering": {
+      "display_name": "Question Answering",
+      "description": "Extractive and generative QA",
+      "icon": "help",
+      "metrics": ["exact_match", "f1"],
+      "requires_encoder": true
+    },
+    "translation": {
+      "display_name": "Translation",
+      "description": "Machine translation between languages",
+      "icon": "translate",
+      "metrics": ["bleu", "meteor", "chrf"],
+      "requires_encoder_decoder": true
+    }
+  }
+}