Name: UltraChat 200K
Creator: Kaggle
License: https://creativecommons.org/publicdomain/zero/1.0/

About this Dataset

UltraChat 200K

200K Dialogues of Diverse Topics for NLG Research

By Huggingface Hub [source]

About this dataset

UltraChat-200k is an invaluable resource for natural language understanding, generation and dialog system research. With 1.4 million dialogues spanning a variety of topics, this parquet-formatted dataset offers researchers four distinct formats to aid in their studies: test_sft, train_sft, train_gen and test_gen. Each entry follows the same simple format with three essential fields: prompt, prompt_id and messages - making this corpus an ideal choice for anyone looking to advance their work on natural language understanding and generation systems. Whether you're just starting out or already have several years of research experience under your belt, UltraChat-200k will no doubt prove itself a valuable asset!

More Datasets

For more datasets, click here.

Featured Notebooks

🚨 Your notebook can be here! 🚨!

How to use the dataset

First, you'll find three columns within each entry: Promp, Promp_id and Messages. The promp column contains the initial statement or question that starts the dialogue. Then, The messages column is used for compassiong responses to that initial promt.

Next, Familiarizing yourself with the four split dataset's structure and schemas will be beneficial in utilizing this dataset correctly. Of these four splits, Test_sft can be used for evaluating the performance of natural language understanding models while Train_sft holds 1.4 million dialogues to train these models with various topics included in these dialogues (prompts). Then Train_gen is used for natural language generation research which involves building a model that produces its own messages in response to prompts based on training dialogues from Train_sft while Testwart_gen uses thisTraining data as well as other unseen messages for evaluation purposes. Finally ,the parquet-formatted system allows convenient storage of large amounts of structured data into smaller files which takes up significantly less space than traditional file formats suchas JSON or CSV files would require .

With all this information understood ,it is now safe to flexibly use UltraChat-200k :NLP Dataset within your research to develop AI natural conversations systems as well ML algorithms through its wide range ofdat inquiries spread across various domains

Research Ideas

Develop voice-enabled chatbots capable of natural and engaging conversations.

Utilize large dialog language datasets to train AI models on how humans interact naturally and create better, more sophisticated conversational systems.

Create a sentiment analysis system which can identify positive or negative conversation threads in the dataset using NLP techniques such as text classification and topic modeling

Acknowledgements

If you use this dataset in your research, please credit the original authors.
Data Source

License

License: CC0 1.0 Universal (CC0 1.0) - Public Domain Dedication
No Copyright - You can copy, modify, distribute and perform the work, even for commercial purposes, all without asking permission. See Other Information.

Columns

File: test_sft.csv

Column name	Description
prompt	The prompt for the conversation. (String)
messages	The messages in the conversation. (String)

File: train_sft.csv

Column name	Description
prompt	The prompt for the conversation. (String)
messages	The messages in the conversation. (String)

File: train_gen.csv

Column name	Description
prompt	The prompt for the conversation. (String)
messages	The messages in the conversation. (String)

Acknowledgements

If you use this dataset in your research, please credit the original authors.
If you use this dataset in your research, please credit Huggingface Hub.

Tables

Test Gen

@kaggle.thedevastator_ultrachat_200k_nlp_dataset.test_gen

80.71 MB
28,304 rows
3 columns

CREATE TABLE test_gen (
  "prompt" VARCHAR,
  "prompt_id" VARCHAR,
  "messages" VARCHAR
);

Test Sft

@kaggle.thedevastator_ultrachat_200k_nlp_dataset.test_sft

81.57 MB
23,110 rows
3 columns

CREATE TABLE test_sft (
  "prompt" VARCHAR,
  "prompt_id" VARCHAR,
  "messages" VARCHAR
);

Train Gen

@kaggle.thedevastator_ultrachat_200k_nlp_dataset.train_gen

733.54 MB
256,032 rows
3 columns

CREATE TABLE train_gen (
  "prompt" VARCHAR,
  "prompt_id" VARCHAR,
  "messages" VARCHAR
);

Train Sft

@kaggle.thedevastator_ultrachat_200k_nlp_dataset.train_sft

735.66 MB
207,865 rows
3 columns

CREATE TABLE train_sft (
  "prompt" VARCHAR,
  "prompt_id" VARCHAR,
  "messages" VARCHAR
);