datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
transcribed_callsCall recordings and their transcripts, in both words and phonemestool-calls-singleturnsolana-memecoin-calls
Solana memecoin calls — a public record with the misses left in
8,161 pump.fun token calls, each with the market cap we called it at, the peak it reached
afterwards, and the exact second it was posted publicly. The whole file is hashed and the hash is
anchored in a Bitcoin block, so no row can be added, edited or back-dated after the fact.
Every trading channel publishes its winners. This is the same feed with the losers still in it —
about six calls in ten never double, and… See the full description on the dataset page: https://huggingface.co/datasets/Smurfetc/solana-memecoin-calls.earnings-calls-qa
Lamini Earning Calls QA Dataset
Description
This dataset contains transcripts of earning calls for various companies, along with questions and answers related to the companies' financial performance and other relevant topics.
Format
The transcripts, questions, and answers are in the form of jsonlines files, with each json object in the file containing the transcript of an earning call for a single company.
Data Pipeline Code
The entire data pipeline… See the full description on the dataset page: https://huggingface.co/datasets/lamini/earnings-calls-qa.synthetic-tool-calls-v2synthetic-tool-callstool-calls-multiturnaic_test_calls_en
aic_test_calls_en
An evaluation dataset of real far end call recordings for accurate speaker transcription under realistic conditions.
The dataset provides around 35 minutes of genuine team and test calls recorded at ai-coustics, captured under real-world acoustic and transmission conditions — real rooms, real devices, real meeting and telephony channels — together with human-produced ground-truth transcripts. No clean speech reference exists; evaluation is transcript-based.… See the full description on the dataset page: https://huggingface.co/datasets/ai-coustics/aic_test_calls_en.unified-tool-calls
unified-tool-calls
A single consolidated corpus of tool-calling conversations converted from four source datasets into one unified format.
Source datasets
source
repository
raw rows
converted
in final corpus
xlam
dusersad12/xlam-function-calling-60k
100
97
92
toolace
dusersad12/ToolACE
30
30
28
glaive
dusersad12/glaive_toolcall_en
100
97
92
hermes
dusersad12/hermes-tool-calls
18
18
16
Total entries in the merged corpus: 228.… See the full description on the dataset page: https://huggingface.co/datasets/dusersad12/unified-tool-calls.fortivio-graded-stock-calls
Fortivio graded AI stock-prediction calls
Every directional call made by the Fortivio.AI agent
fleet on US-listed stocks, locked before the market open and graded after
the close against that session's open→close move. Nothing is backtested
and nothing is revised after the fact: a call is written once, before 09:30 ET,
and scored once, mechanically, when the session's final price bar arrives.
One row per graded call. One file per month. Monthly summaries with hit rate
by… See the full description on the dataset page: https://huggingface.co/datasets/WhatHui/fortivio-graded-stock-calls.MiMIC_Multi-Modal_Indian_Earnings_Calls_DatasetMiMIC_analysis_code.ipynb : This is the actual code file with results
MiMIC_Multi-Modal_Indian_Earnings_Calls.xlsx : This is the main file having to be used for training. Descripton of columns of this file is given below.
getting_all_texts_together_embeddings_dim128_CPU.pkl : This is the file having embeddings of texts (extracted from transcripts, images) and tables (extracted from images) concatenated one after the other.… See the full description on the dataset page: https://huggingface.co/datasets/sohomghosh/MiMIC_Multi-Modal_Indian_Earnings_Calls_Dataset.tool-calls-sharegptearnings_calls_transcriptsasr_public_phone_calls_2func_calls
retrain-pipelines Function Calling
version 0.237 - 2026-08-25 12:39:39 UTC
Source datasets :
main :
Xlam Function Calling 60k
Salesforce/xlam-function-calling-60k
(26d14eb -
2025-01-24 19:25:58 UTC)
license :
cc-by-4.0
arxiv :
- 2406.18518
data-enrichment :
Natural Questions Clean
lighteval/natural_questions_clean
(a72f7fa -
2023-10-17 20:29:08 UTC)
license :
unknown
The herein dataset has 2 configs :… See the full description on the dataset page: https://huggingface.co/datasets/retrain-pipelines/func_calls.tool-calls-dpofunc_calls_ds
retrain-pipelines Function Calling
version 0.44 - 2026-03-01 12:02:11 UTC
Source datasets :
main :
Xlam Function Calling 60k
Salesforce/xlam-function-calling-60k
(26d14eb -
2025-01-24 19:25:58 UTC)
license :
cc-by-4.0
arxiv :
- 2406.18518
data-enrichment :
Natural Questions Clean
lighteval/natural_questions_clean
(a72f7fa -
2023-10-17 20:29:08 UTC)
license :
unknown
The herein dataset has 2 configs : continued_pre_training and supervised_finetuning.
The former… See the full description on the dataset page: https://huggingface.co/datasets/retrain-pipelines/func_calls_ds.sp500_dataset_earnings_callstool-calls-single-reasoningtool-calls-mini
tool-calls-mini
500 synthetic tool-calling conversations in TRL's conversational format,
for supervised fine-tuning. Built to be coherent: every tool result is a plausible
function of the arguments it was called with, and every final answer reflects that
result — so the set teaches when to call a tool, not just what a call looks like.
Format
Each row has messages and tools. An assistant turn carries tool_calls instead of
content; the tool replies as a tool role… See the full description on the dataset page: https://huggingface.co/datasets/qgallouedec/tool-calls-mini.hermes-tool-calls
hermes-tool-calls
A mirror of the hermes tool-calling conversations. Every line of
data/train.jsonl is one JSON object.
Entry format
messages: list of turns in OpenAI chat style.
Some entries begin with a system message that defines the assistant's
persona. Persona prompts never contain tool definitions or tool-usage
instructions.
Assistant messages that invoke tools carry tool_calls: a list of
{id, type: "function", function: {name, arguments}} objects where… See the full description on the dataset page: https://huggingface.co/datasets/dusersad12/hermes-tool-calls.urdu-emergency-calls
Urdu Emergency Call Conversations Dataset (Pakistan)
Overview
This dataset contains 5,000 curated Urdu emergency call conversation samples from the Pakistan region, designed to support training and evaluation of Urdu Large Language Models (LLMs) for emergency response, command centers, and interpreter-style systems.
The conversations simulate real-world emergency scenarios such as:
Floods
Medical emergencies
Accidents
Crimes
Natural disasters
Public safety… See the full description on the dataset page: https://huggingface.co/datasets/abeeranajam31/urdu-emergency-calls.tool-callsTool calling master dataset
Contains the following:
Query -> Available tools (name + description + schema) -> Tool name
Sources (identified by source column):
subsets of existing tool-calling dataset sources parsed into the above format
synthetic data
Will be parsed into the following two passes:
Query -> List of tool names + descriptions -> Tool name
Tool name + tool schema -> Tool call
synthetic-tool-calls-v2-dpo-pairsasr_calls_2_valformatted-earnings-calls
Dataset Card for Dataset Name
Modified version of the Lamini earnings call dataset that has rolled up the data by stock ticker and
contains a jumbled transcript and list of questions delineated by newline
Dataset Details
Dataset Description
Curated by: [More Information Needed]
Funded by [optional]: [More Information Needed]
Shared by [optional]: [More Information Needed]
Language(s) (NLP): [More Information Needed]
License: [More Information Needed]… See the full description on the dataset page: https://huggingface.co/datasets/JLawI11/formatted-earnings-calls.sharegpt-tool-calls
ShareGPT Tool Calls Dataset
A clean, structured tool-call dataset in ShareGPT format derived from the popular glaiveai/glaive-function-calling-v2 dataset.
Format
Each example follows the canonical ShareGPT schema used by OpenHermes, UltraChat, and many other fine-tuning datasets:
{
"conversations": [
{"from": "system", "value": "You are a helpful assistant with access to the following functions. Use them if required - {...}"},
{"from": "human", "value": "Can… See the full description on the dataset page: https://huggingface.co/datasets/Guilherme34/sharegpt-tool-calls.darija-synthetic-callsxlam-function-callstool-calls-eval