MELD resource

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1079 stars · 234 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

A research repository from declare-lab containing implementations of neural architectures (DialogueRNN, DialogueGCN, COSMIC, bcLSTM) for emotion recognition in conversations, along with the MELD multimodal dataset. It serves as both a benchmark dataset and reference codebase for conversational emotion recognition research.

Use cases

• recognize emotions in conversations
• train emotion recognition models on the MELD dataset
• benchmark conversational emotion recognition architectures
• run COSMIC for state-of-the-art ERC
• extract emotion causes from dialogue
• research multimodal sentiment analysis in conversations

When to choose

• you need the MELD dataset or baselines for emotion recognition in conversation research
• you want reference implementations of DialogueRNN, DialogueGCN, or COSMIC
• you are doing academic work on conversational emotion or sentiment analysis

When to avoid

• you need a production-ready emotion detection API or service
• you want a maintained library with stable APIs rather than research code
• your task is not conversation-level emotion recognition

Facets

dataset · maturity maintenance

nlp machine-learning deep-learning machine-learning artificial-intelligence python emotion-recognition conversational-ai sentiment-analysis pytorch multimodal dialogue-systems meld-dataset natural-language-processing

Member repositories

For agents

markdown · JSON · MCP: product_card(name="declare-lab/MELD")