TutorialGuide/blended-skill-talk-fixed

1---2dataset_info:3  features:4  - name: personas5    sequence: string6  - name: additional_context7    dtype: string8  - name: previous_utterance9    sequence: string10  - name: context11    dtype: string12  - name: free_messages13    sequence: string14  - name: guided_messages15    sequence: string16  - name: suggestions17    struct:18    - name: convai219      sequence: string20    - name: empathetic_dialogues21      sequence: string22    - name: wizard_of_wikipedia23      sequence: string24  - name: guided_chosen_suggestions25    sequence: string26  splits:27  - name: train28    num_bytes: 899072029    num_examples: 409630  - name: validation31    num_bytes: 160153632    num_examples: 72333  download_size: 1060044834  dataset_size: 1059225635configs:36- config_name: default37  data_files:38  - split: train39    path: data/train.parquet40  - split: validation41    path: data/validation.parquet42size_categories:43- 1K<n<10K44license: apache-2.045task_categories:46- text-generation47---48 49## Compatibility Update50 51This repository is a compatibility-fixed version of the original **Blended Skill Talk** dataset.52 53The original dataset can be found at:54 55- Original Hugging Face dataset: https://huggingface.co/datasets/anezatra/blended-skill-talk56 57This version was created to maintain compatibility with newer versions of the Hugging Face `datasets` library.58 59### Changes from the Original Dataset60 61The following changes were made:62 63- Removed the unused `label_candidates` column.64- Removed the invalid `label_candidates` feature definition (`sequence: null`) that caused schema parsing failures.65- Re-exported the dataset Parquet files with a corrected schema.66- Updated dataset metadata to match the exported files.67 68No conversation examples were modified.69 70No examples were added or removed.71 72The dataset content remains the same as the original release.73 74---75 76# Blended Skill Talk77 78## Dataset Summary79 80Blended Skill Talk is a conversational dataset designed to train and evaluate dialogue systems capable of combining multiple conversational skills.81 82The dataset combines persona-based conversation, empathetic dialogue, and knowledge-grounded conversation elements to encourage models to produce engaging, context-aware responses.83 84The dataset incorporates information from ConvAI2, EmpatheticDialogues, and Wizard of Wikipedia.85 86---87 88## Data Structure (Updated)89 90### Fields91 92| Field | Description |93|:------|:-------------|94| personas | List of personas participating in the conversation |95| additional_context | Extra context or scenario description |96| previous_utterance | The immediately preceding dialogue turn(s) |97| context | General conversation context |98| free_messages | Free-form user or system messages |99| guided_messages | Messages generated via guided prompts |100| suggestions | Suggested responses from ConvAI2 Empathetic Dialogues and Wizard of Wikipedia |101| guided_chosen_suggestions | Selected guided suggestions associated with the conversation |102 103---104 105### Splits106 107| Split | Examples | Size (bytes) | Description |108|:------|----------|--------------|-------------|109| Train | 4096 | 8990720 | Used for model training |110| Validation | 723 | 1601536 | Used for validation and tuning |111 112**Total dataset size:** 10592256 bytes  113**Total number of dialogues:** 4819114 115---116 117## Loading This Version118 119```python120from datasets import load_dataset121 122dataset = load_dataset(123    "TutorialGuide/blended-skill-talk-fixed"124)125 126print(dataset)127```128 129---130 131## References132 133Original Hugging Face dataset: https://huggingface.co/datasets/anezatra/blended-skill-talk134 135Smith, E. M., Williamson, M., Shuster, K., Weston, J., & Boureau, Y. L. (2020). Can You Put it All Together: Evaluating Conversational Agents' Ability to Blend Skills. *arXiv preprint arXiv:2004.08449*. ([https://arxiv.org/abs/2004.08449](https://arxiv.org/abs/2004.08449))