The dataset contains 2231142 cooking recipes (>2 millions). It's processed in more careful way and provides more samples than any other dataset in the area.

Dataset

This dataset was created by Kalikrishna prasanna Yalamati

Released under Other (specified in description)

Contents

Mahimas/recipenlg dataset hosted on Hugging Face and contributed by the HF Datasets community

The RecipeNLG dataset (Poznań University of Technology) is an expansion of the Recipe1M+ dataset, offering a significantly larger selection of recipes. Unlike its predecessor, this dataset does not prioritize linking cooking instructions with their corresponding images and instead emphasizes the recipe text, structure, and logic. This updated collection contains over one million newly preprocessed and deduplicated recipes, making it the largest publicly accessible dataset in its field.

The data could be downloaded from Poznań University of Technology website.

Dataset of all the wonders of the world when it comes to food composition. Dishes ranging from a bread to a Swedish midsommer smorgasbord.

Some intersting things to know before starting: - There are duplicate dishes but the majority has a different recipe. - I have removed all duplicates in "NER" and made them all lowercase - Some of the columns are arrays, use df.apply and import Json to work with them. Check out the code in my notebook

https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F10179217%2Fff93e5e3e0876fbb0298c780e1d4e342%2FWordcloud.png?generation=1687898448282120&alt=media" alt="">

Link to original: https://clickhouse.com/docs/en/getting-started/example-datasets/recipes

Try it with SQL: https://play.clickhouse.com/play?user=play#U0VMRUNUCiAgICBhcnJheUpvaW4oTkVSKSBBUyBrLAogICAgY291bnQoKSBBUyBjCkZST00gcmVjaXBlcwpHUk9VUCBCWSBrCk9SREVSIEJZIGMgREVTQwpMSU1JVCA1MA==

Terms and Conditions

I (the "Researcher") have requested permission to use the RecipeNLG dataset (the "Dataset") at Poznań University of Technology (PUT). In exchange for such permission, Researcher hereby agrees to the following terms and conditions:

1. Researcher shall use the Dataset only for non-commercial research and educational purposes.
2. PUT makes no representations or warranties regarding the Dataset, including but not limited to warranties of non-infringement or fitness for a particular purpose.
3. Researcher accepts full responsibility for his or her use of the Dataset and shall defend and indemnify PUT, including its employees, Trustees, officers and agents, against any and all claims arising from Researcher's use of the Dataset including but not limited to Researcher's use of any copies of copyrighted images or text that he or she may create from the Dataset.
4. Researcher may provide research associates and colleagues with access to the Dataset provided that they first agree to be bound by these terms and conditions.
5. If Researcher is employed by a for-profit, commercial entity, Researcher's employer shall also be bound by these terms and conditions, and Researcher hereby represents that he or she is fully authorized to enter into this agreement on behalf of such employer.

EmTpro01/recipe-nlg-50k dataset hosted on Hugging Face and contributed by the HF Datasets community

A heavily curated dataset from recipe-nlg (source="Gathered" only). A lot of scraping artifacts, typographical errors, unicode, empty and very short recipes were removed. Then it has been formated into Alpaca instruction set with Instructions, Input and Output. The total number of recipes went from ~2M2 (original dataset) to ~500K. Obviously, it's still not perfect (I won't lie, the original dataset was very flawed). To fully fix this would require a very time-consuming manual edition, so you… See the full description on the dataset page: https://huggingface.co/datasets/AdamCodd/recipe-nlg-alpaca.

Context

Please cite the dataset using the BibTex provided in one of the following sections if you are using it in your research, thank you!

A recipe is a piece of writing that outlines the materials required, the cooking procedure, and how to use them while preparing or baking food. Using the knowledge of food specialists and an active learning methodology, the "Assorted, Archetypal, and Annotated Two Million (3A2M) Cooking Recipe Dataset" has two million culinary meals annotated in relevant categories. The RecipeNLG collection serves as the foundation for the 3A2M cooking recipe dataset.

Content

3A2M dataset comprises five features (five columns) in total: 1. title - the name of the food 2. directions - step-by-step description of the food recipes 3. NER - ingredients of cooking recipes 4. genre - assigned category (representation in string format) 5. label - numeric representation of the genres

Data for the title, directions, and NER attributes are directly appended from the RecipeNLG dataset. Three human specialists classify 300,000 random recipes into one of nine categories. The remaining 1900K recipes are automatically classified using active learning and a query-by-committee approach.

Genres

Human experts selected nine genres to categorize this dataset. The nine categories are: - bakery - drinks - non-veg - vegetables - fast food - cereals - meals - sides - fusion

Citation

If you're using this dataset for your work, please cite the following articles:

Citation in text format: N. Sakib, G. Shahariar, M. M. Kabir, M. K. Hasan, and H. Mahmud, “Assorted, archetypal and annotated two million (3a2m) cooking recipes dataset based on active learning.”

Citation in BibTex format: @article{sakibassorted, title={Assorted, Archetypal and Annotated Two Million (3A2M) Cooking Recipes Dataset based on Active Learning}, author={Sakib, Nazmus and Shahariar, GM and Kabir, Md Mohsinul and Hasan, Md Kamrul and Mahmud, Hasan} }

Inspiration

Considering the breadth and arrangement of the information by genre,

1. medical nutritionists may recommend a variety of meals to patients.
2. If a portion of the recipe can be estimated, which can be used to analyze food intake for various types of food analysis or nutrients.
3. This well-annotated dataset might be utilized for NLP tasks such as recipe generation.

Dataset Card for "recipe-nlg-llama2"

More Information needed

Dataset Card for "tokenized-recipe-nlg-gpt2-ners-ingredients-only"

More Information needed

Preprocessed cooking recipes data taken from RecipeNLG

Context

Please cite the dataset using the BibTex provided in one of the following sections if you are using it in your research, thank you!

The "Assorted, Archetypal, and Annotated Two Million Extended (3A2M+) Cooking Recipe Dataset" has 22,31,142 culinary meals annotated in relevant categories. The 3A2M dataset serves as the foundation for the 3A2M+ cooking recipe dataset.

Content

3A2M+ dataset comprises five features (five columns) in total: 1. title - the name of the food 2. directions - step-by-step description of the food recipes 3. NER - ingredients of cooking recipes 4. Extended NER - missing NER from the directions 5. genre - assigned category (representation in string format) 6. label - numeric representation of the genre (in total 9 categories)

Extended NERs are extracted from the text in the "directions" column text considering the following important factors: - Temperature of the food - Cooking Method - Missing Ingredients - Cooking Pot etc.

Data for the title, directions, and NER attributes are directly appended from the RecipeNLG dataset. Three human specialists classify 300,000 random recipes into one of nine categories. The remaining 1900K recipes are automatically classified using active learning and a query-by-committee approach. Named Entities are extracted using popular NER tools using a unique pipeline.

Genres

Human experts selected nine genres to categorize this dataset. The nine categories are: - bakery - drinks - non-veg - vegetables - fast food - cereals - meals - sides - fusion

Citation

If you're using this dataset for your work, please cite the following articles:

Citation in text format: N. Sakib, G. Shahariar, M. M. Kabir, M. K. Hasan, and H. Mahmud, “Assorted, archetypal and annotated two million (3a2m) cooking recipes dataset based on active learning.”

Citation in BibTex format: @article{sakibassorted, title={Assorted, Archetypal and Annotated Two Million (3A2M) Cooking Recipes Dataset based on Active Learning}, author={Sakib, Nazmus and Shahariar, GM and Kabir, Md Mohsinul and Hasan, Md Kamrul and Mahmud, Hasan} }

Inspiration

Considering the breadth and arrangement of the information by genre,

1. If a portion of the recipe can be estimated, which can be used to analyze food intake for various types of food analysis or nutrients.
2. This well-annotated dataset might be utilized for NLP tasks such as recipe generation.

3A2M Cooking Recipe Dataset

https://www.kaggle.com/datasets/nazmussakibrupol/3a2m-cooking-recipe-dataset

3A2M Cooking Recipe Dataset ( Human Annotated Subset)

https://t.ly/qI8rb

Dataset pairing GPT-4 synthesized instructions with outputs from RecipeNLG in Axolotl's "alpaca" jsonl format