{"success":true,"data":{"author":"Miguel Del Rio, Natalie Delworth, Ryan Westerman et al. (Rev.com)","bucketAccessType":"public","bucketName":"Revai/earnings21","bucketPrefix":"","category":"Speech & Audio","cloudProvider":"huggingface","createdAt":"2026-11-14T14:15:18.298Z","description":"Earnings-21 is a 39-hour corpus of English earnings calls containing entity-dense speech from nine financial sectors, released by Rev to benchmark automatic speech recognition systems in the wild with special attention to named entity recognition. Calls are long-form, spontaneous and full of company names, people, numbers, dates and financial terms, which makes them a hard test for ASR and for entity-aware word error rate. The release includes reference transcripts in Rev's .nlp token format with entity labels, normalisation files that list acceptable alternative spellings, and per-file WER tags, designed for Rev's fstalign scoring tool. This repository, published by Rev, provides the audio transcoded to 16 kHz in a wav folder, with nlp_references, normalizations and wer_tags folders (44 files each) and a metadata.jsonl index (measured 2026-11-14; the Hub reports about 5.3 GB). Read the wav files with soundfile or torchaudio and the .nlp and JSON references with pandas. The original GitHub repository holds the original audio and full entity tags. Rev releases the Earnings datasets under CC BY-SA 4.0; cite Del Rio et al., Interspeech 2021.","displayTitle":"Earnings-21: entity-dense earnings call ASR benchmark","id":"3225f7df-0868-55bc-ad93-6cfa94866758","license":"CC-BY-SA-4.0","licenseEvidenceUrl":"https://github.com/revdotcom/speech-datasets/blob/main/earnings21/README.md","objectCountEstimate":179,"probeStatus":"ok","referenceUrl":"https://arxiv.org/abs/2104.11348","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"Revai/earnings21\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"Revai/earnings21\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=Revai, license:cc-by-sa-4.0, downloads=742"],"format":"mixed","generatedAt":"2026-11-14T14:15:23.438Z","layout":{"example":"metadata.jsonl","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"{\"file_name\": \"wav/4320211.wav\", \"audio_length\": 3285.848, \"original_sample_rate\": 24000, \"company_name\": \"Monro Inc\", \"financial_quarter\": 3, \"sector\": \"Consumer Goods\", \"speaker_switches\": 82, \"unique_speakers\": 10, \"curator_id\": \"1\", \"text\": \"Good morning ladies and gentlemen, and welcome to the Monro Inc Earnings Conference Call for the third quarter Fiscal 2020. At this time all participants are in listen only mode. Later, we will conduct a question and answer session and instructions will follow at th","encoding":"utf8","key":"metadata.jsonl","truncatedAt":512}],"sampleObjects":["metadata.jsonl","nlp_references/4320211.nlp","nlp_references/4330115.nlp","nlp_references/4341191.nlp","nlp_references/4344338.nlp"],"sidecars":[{"bytes":"---\nlicense: cc-by-sa-4.0\nlanguage:\n- en\ntags:\n- earnings\n- earnings-calls\n- speech\n- audio\nconfigs:\n- config_name: default\n  data_files:\n  - split: test\n    path:\n    - metadata.jsonl\n    - wav/*.wav\n---\n\n# Earnings 21\n\nThe Earnings 21 dataset ( also referred to as earnings21 ) is a 39-hour corpus of earnings calls containing entity dense speech from nine different financial sectors. This corpus is intended to benchmark automatic speech recognition (ASR) systems in the wild with special attention towards named entity recognition (NER).\n\nIn this repo, we provided the transcoded files to 16KHz with the formatted text. This limits the utility of the original dataset due to the restrictions on how we can identify entities and perform normalization for more fair evaluations.\nFor more advanced usage, usage of the entity tags, original audios, and more checkout the [original Github Repository](https://github.com/revdotcom/speech-datasets/tree/main/earnings21) as well as our WER calculation tool, [fstalign](https://github.com/revdotcom/fstalign/tree/develop).\n\nRead the paper on ISCA's page [here](https://www.isca-archive.org/interspeech_2021/delrio21_interspeech.pdf) or on ArXiv [here](https://arxiv.org/abs/2104.11348)!\n\n# Table of Contents\n\n* [File Format Overview](#file-format-overview)\n  + [nlp Files](#nlp-files)\n    - [Example](#example-nlp-file)\n  + [wer_tag JSON](#wer_tag-json)\n    - [Example](#example-wer_tag-json)\n* [Entity Labels](#entity-labels)\n* [WER Calculation](#wer-calculation)\n* [Results](#results)\n* [Cite this Dataset](#cite-this-dataset)\n\n# File Format Overview\nIn the following section, we provide an overview of the file formats we provide with this dataset.\n\n## nlp Files\nNLP files are `.csv` inspired, pipe-separated text files that contain token and metadata information of a transcript. Each line of a file represents a single transcript token and the metadata associated with it.\n\n|Column Title|Description\n|--|--|\n| Column 1: `token` | A single token in the transcript. These are typically single words or multiple words with hyphens in between. |\n| Column 2: `speaker` | A unique integer that associates this token to a specific speaker in an audio |\nColumn 3: `ts`          |     A float representing the start time of the token, in seconds |\nColumn 4: `endTs`       |     A float representing the end time of the token, in seconds |\nColumn 5: `punctuation` |     A punctuation character that is included at the end of a token that is used when reconstructing the transcript. Example punctuation: `\",\", \";\", \".\", \"!\"`. |\nColumn 6: `case`  | A two letter code to denominate the which of four possible casings for this token: <ul><li>`UC` - Denotes a token that has the first character in uppercase and every other character lowercase.</li><li>`LC` - Denotes a token that has every character in lowercase.</li><li>`CA` - Denotes a token that has every character in uppercase.</li><li>`MC` - Denotes a token that doesnâ€™t follow the previous rules. This is the case when upper- and lowercase characters are mixed throughout the token</li></ul> |\nColumn 7: `tags`        |     Displays one of the several entity tags that are listed in wer_tags in long form - such that the displayed entity here is in the form `ID:ENTITY_CLASS`. |\nColumn 8: `wer_tags`    |     A list of entity tags that are associated with this token. In this field, only entity IDs should be present. The specific ENTITY_CLASS for each ID can be extracted from an accompanying wer_tags sidecar json. |\n\n_**Note that each entity ID is unique to that specific entity. Entities can be comprised of single and multiple tokens. Within a file there can be several entities of the same ENTITY_CLASS but only one entity can be labeled with any given ID.**_\n\n\n### Example nlp File\n`example.nlp`\n\n```\ntoken|speaker|ts|endTs|punctuation|case|tags|wer_tags\nGood|0||||UC|[]|[]\nmorning|0||||LC|['5:TIME']|['5']\nand|0||||LC|[]|[]\nwelcome|0||||LC|[]|[]\nto|0||||LC|[]|[]\nthe|0||||LC|['6:DATE']|['6']\nfirst|0||||LC|['6:DATE']|['6']\nquarter|0||||LC|['6:DATE']|['6']\n2020|0||||CA|['0:YEAR']|['0', ","encoding":"utf8","key":"README.md","name":"README.md","size":11241,"truncated":true}],"sourceObjectKey":"metadata.jsonl","starterCell":"from datasets import load_dataset\nds = load_dataset(\"Revai/earnings21\")\nprint(ds)","starterSamplePath":"metadata.jsonl","totalObjects":179},"summary":"Earnings-21 is a 39-hour corpus of English earnings calls containing entity-dense speech from nine financial sectors, released by Rev to benchmark automatic speech recognition…","title":"Earnings-21: entity-dense earnings call ASR benchmark","updatedAt":"2026-11-14T17:49:14.603Z","visibility":"public","websiteUrl":"https://github.com/revdotcom/speech-datasets"}}