Dataset · Salesforce
Unified DialogStudio dialogue dataset corpus
Salesforce/dialogstudio
Combined dialogue data for QA, summarization, generation, classification, and table-to-text.
What's inside
DialogStudio unites various dialogue datasets for conversational AI across multiple tasks: question-answering, summarization, text-generation, table-to-text, and classification. The language is English. The card indicates that examples of each dataset can be viewed on GitHub; instruction-aware models and an update for AI Agent/xLAM are also mentioned.
What you can do with it
- Training a dialogue model on unified data
- SFT instruction-aware conversational AI
- Testing dialogue QA and summarization
- Tasks
- Question answering, Summarization, Text generation, Table → text, Text classification, Table question answering
- Languages
- English
- Downloads this month
- 1.9K
- Likes
- 230
- Access
- on request on Hugging Face
Models trained on it
- sergioburdisso/dialog2flow-joint-dse-baseSentence similarity 1.2K
- sergioburdisso/dialog2flow-joint-bert-baseSentence similarity 512
- mahimairaja/tweet-summarization-llama-2-finetunedText generation 181
- Salesforce/dialogstudio-t5-3b-v1.0 159
- Salesforce/dialogstudio-t5-base-v1.0 95
- Salesforce/dialogstudio-t5-large-v1.0 86
- sergioburdisso/dialog2flow-single-bert-baseSentence similarity 52
- sergioburdisso/dialog2flow-single-dse-baseSentence similarity 37
How to load it
from datasets import load_dataset
ds = load_dataset("Salesforce/dialogstudio")