CoolFace
Datasetpublic

llama-lang-adapt/EmotionAnalysis

Dataset Card for EmotionAnalysisFinal EmotionAnalysisFinal is the official dataset for SemEval-2025 Task 11, Track C: Cross-lingual Emotion Detection in Social Media Text. This dataset comprises multilingual social media posts annotated for six basic emotions: anger, disgust, fear, joy, sadness, and surprise. The annotation schema is multi-label. Each language-specific configuration (subset) contains validation and test splits. Split Original Source Notes validation… See the full description on the dataset page: https://huggingface.co/datasets/llama-lang-adapt/EmotionAnalysis.

sourceHugging Facecc-by-sa-4.0updated 1y agoView on Hugging Face
0likes23downloads
Dataset Card

Dataset Card for EmotionAnalysisFinal

EmotionAnalysisFinal is the official dataset for SemEval-2025 Task 11, Track C: Cross-lingual Emotion Detection in Social Media Text. This dataset comprises multilingual social media posts annotated for six basic emotions: anger, disgust, fear, joy, sadness, and surprise. The annotation schema is multi-label.

Each language-specific configuration (subset) contains validation and test splits.

SplitOriginal SourceNotes
validationdev CSV filesUsed as development set during the task
testtest CSV filesGold labels released after task evaluation

Languages and Configuration Names

The dataset supports the following languages, accessible via their ISO 639-3 or custom codes as configuration names (which correspond to the name parameter in load_dataset if using the accompanying Python loading script, or the second argument if loading via metadata only):

afr, amh, arq, ary, chn, deu, eng, esp, hau, hin, ibo, ind, jav, kin, mar, orm, pcm, ron, rus, som, sun, swa, swe, tat, tir, ukr, vmw, xho, yor, zul

How to Use

Load a specific language and split using load_dataset:

python
from datasets import load_dataset

# Example: Load the Yoruba test split
repo_id = "llama-lang-adapt/EmotionAnalysisFinal"
lang_code = "yor" # Replace with desired language code
split_name = "test" # Or "validation"

# If a .py loading script (e.g., emotion_analysis_final.py) is present in the repo,
# 'load_dataset' will use it. The 'name' parameter corresponds to the config name.
ds = load_dataset(
    repo_id,
    name=lang_code,  # Configuration name (language code)
    split=split_name
)

# Print the first example
print(f"First example from {lang_code} {split_name} split:")
if len(ds) > 0:
    print(ds[0])
else:
    print("Dataset is empty or loading failed for the split.")

# Example: Load the Afrikaans validation split
ds_afr_val = load_dataset(repo_id, name="afr", split="validation")
print(f"\nFirst example from afr validation split:")
if len(ds_afr_val) > 0:
    print(ds_afr_val[0])
else:
    print("Dataset is empty or loading failed for the split.")