datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
MMEvol
Dataset Card for MMEvol-480K
This is the official data collection of the paper "MMEvol: Empowering Multimodal Large Language Models with Evol-Instruct"
Please see paper & website for more information:
arXiv: https://arxiv.org/pdf/2409.05840
website: https://mmevol.github.io/home_page.html
Github: https://github.com/RainBowLuoCS/MMEvol
Overview
The Tongyi-ConvAI generates this dataset for multi-modal supervised fine-tuning. This dataset was used to train our… See the full description on the dataset page: https://huggingface.co/datasets/Tongyi-ConvAI/MMEvol.conv_ai_2ConvAI is a dataset of human-to-bot conversations labelled for quality. This data can be used to train a metric for evaluating dialogue systems. Moreover, it can be used in the development of chatbots themselves: it contains the information on the quality of utterances and entire dialogues, that can guide a dialogue system in search of better answers.OmniCharacterThis is the official data collection for paper "OmniCharacter: Towards Immersive Role-Playing Agents with Seamless Speech-Language Personality Interaction".
Please see paper & code for more information:
paper: https://www.arxiv.org/abs/2505.20277
code: https://github.com/AlibabaResearch/DAMO-ConvAI/tree/main/OmniCharacter
conv_ai_3The Conv AI 3 challenge is organized as part of the Search-oriented Conversational AI (SCAI) EMNLP workshop in 2020. The main aim of the conversational systems is to return an appropriate answer in response to the user requests. However, some user requests might be ambiguous. In Information Retrieval (IR) settings such a situation is handled mainly through the diversification of search result page. It is however much more challenging in dialogue settings. Hence, we aim to study the following situation for dialogue settings:
- a user is asking an ambiguous question (where ambiguous question is a question to which one can return > 1 possible answers)
- the system must identify that the question is ambiguous, and, instead of trying to answer it directly, ask a good clarifying question.conv_aiConvAI is a dataset of human-to-bot conversations labelled for quality. This data can be used to train a metric for evaluating dialogue systems. Moreover, it can be used in the development of chatbots themselves: it contains the information on the quality of utterances and entire dialogues, that can guide a dialogue system in search of better answers.ConvAI2-Qwen-Image-2512ConvAI2-Qwen-Image-2512-enhancedphysics-reasoning-dataset
📚 Flux Physics Reasoning Dataset
This dataset contains detailed physics reasoning scenarios designed to train Small Language Models (SLMs) and Liquid Neural Networks in physical intuition.
📄 Format
The dataset is provided in Parquet format (train.parquet) for efficient loading. Each row contains:
prompt: The physics question or scenario description.
answer: The correct physical explanation and answer.
concept: The underlying physics principle (e.g., "Conservation of… See the full description on the dataset page: https://huggingface.co/datasets/convaiinnovations/physics-reasoning-dataset.task1713_convai3_sentence_generation
Dataset Card for Natural Instructions (https://github.com/allenai/natural-instructions) Task: task1713_convai3_sentence_generation
Additional Information
Citation Information
The following paper introduces the corpus in detail. If you use the corpus in published work, please cite it:
@misc{wang2022supernaturalinstructionsgeneralizationdeclarativeinstructions,
title={Super-NaturalInstructions: Generalization via Declarative Instructions on 1600+ NLP Tasks}… See the full description on the dataset page: https://huggingface.co/datasets/Lots-of-LoRAs/task1713_convai3_sentence_generation.ConvAI2-With-Ids_1k-no-redundancyConvAI2-Mapping_1k-no-redundancyConvAI2-Mapping_1k-ERNIE-enhancedtask1714_convai3_sentence_generation
Dataset Card for Natural Instructions (https://github.com/allenai/natural-instructions) Task: task1714_convai3_sentence_generation
Additional Information
Citation Information
The following paper introduces the corpus in detail. If you use the corpus in published work, please cite it:
@misc{wang2022supernaturalinstructionsgeneralizationdeclarativeinstructions,
title={Super-NaturalInstructions: Generalization via Declarative Instructions on 1600+ NLP Tasks}… See the full description on the dataset page: https://huggingface.co/datasets/Lots-of-LoRAs/task1714_convai3_sentence_generation.ConvAI2-Mapping_1k-ERNIE-enhanced-no-redundancyConvAI2-Mapping_1k-ERNIE-originalConvAI2-Mapping_1k-FLUX-enhancedSDPOThis is the data of our paper SDPO: Segment-Level Direct Preference Optimization for Social Agent
ConvAI2-Qwen-original-Qwen3.5-35B-A3B
Visual Memory Results: convai2-qwen-original
This dataset contains the scored output of a visual-memory perplexity experiment.
Experiment metadata
{
"experiment": {
"model_name": "Qwen/Qwen3.5-35B-A3B",
"hf_results_repo": "visual-memory/ConvAI2-Qwen-original-Qwen3.5-35B-A3B",
"results_jsonl": "results/ConvAI2-Qwen-original-Qwen3.5-35B-A3B.jsonl",
"hf_dataset": "visual-memory/ConvAI2-With-Ids_1k-no-redundancy",
"hf_mapping_dataset":… See the full description on the dataset page: https://huggingface.co/datasets/visual-memory/ConvAI2-Qwen-original-Qwen3.5-35B-A3B.ConvAI2-ERNIE-original-Qwen3.5-35B-A3B
Visual Memory Results: convai2-ernie-original
This dataset contains the scored output of a visual-memory perplexity experiment.
Experiment metadata
{
"experiment": {
"model_name": "Qwen/Qwen3.5-35B-A3B",
"hf_results_repo": "visual-memory/ConvAI2-ERNIE-original-Qwen3.5-35B-A3B",
"results_jsonl": "results/ConvAI2-ERNIE-original-Qwen3.5-35B-A3B.jsonl",
"hf_dataset": "visual-memory/ConvAI2-With-Ids_1k-no-redundancy",
"hf_mapping_dataset":… See the full description on the dataset page: https://huggingface.co/datasets/visual-memory/ConvAI2-ERNIE-original-Qwen3.5-35B-A3B.ConvAI2-Mapping_1k-Qwen-original-no-redundancyConvAI2-Qwen-enhanced-Qwen3.5-35B-A3B
Visual Memory Results: convai2-qwen-enhanced
This dataset contains the scored output of a visual-memory perplexity experiment.
Experiment metadata
{
"experiment": {
"model_name": "Qwen/Qwen3.5-35B-A3B",
"hf_results_repo": "visual-memory/ConvAI2-Qwen-enhanced-Qwen3.5-35B-A3B",
"results_jsonl": "results/ConvAI2-Qwen-enhanced-Qwen3.5-35B-A3B.jsonl",
"hf_dataset": "visual-memory/ConvAI2-With-Ids_1k-no-redundancy",
"hf_mapping_dataset":… See the full description on the dataset page: https://huggingface.co/datasets/visual-memory/ConvAI2-Qwen-enhanced-Qwen3.5-35B-A3B.ConvAI2-Qwen-enhanced-gemma-4-26B-A4B-it
Visual Memory Results: convai2-qwen-enhanced
This dataset contains the scored output of a visual-memory perplexity experiment.
Experiment metadata
{
"experiment": {
"model_name": "google/gemma-4-26B-A4B-it",
"hf_results_repo": "visual-memory/ConvAI2-Qwen-enhanced-gemma-4-26B-A4B-it",
"results_jsonl": "results/ConvAI2-Qwen-enhanced-gemma-4-26B-A4B-it.jsonl",
"hf_dataset": "visual-memory/ConvAI2-With-Ids_1k-no-redundancy",
"hf_mapping_dataset":… See the full description on the dataset page: https://huggingface.co/datasets/visual-memory/ConvAI2-Qwen-enhanced-gemma-4-26B-A4B-it.ConvAI2-FLUX-enhanced-Qwen3.5-35B-A3B
Visual Memory Results: convai2-flux-enhanced
This dataset contains the scored output of a visual-memory perplexity experiment.
Experiment metadata
{
"experiment": {
"model_name": "Qwen/Qwen3.5-35B-A3B",
"hf_results_repo": "visual-memory/ConvAI2-FLUX-enhanced-Qwen3.5-35B-A3B",
"results_jsonl": "results/ConvAI2-FLUX-enhanced-Qwen3.5-35B-A3B.jsonl",
"hf_dataset": "visual-memory/ConvAI2-With-Ids_1k-no-redundancy",
"hf_mapping_dataset":… See the full description on the dataset page: https://huggingface.co/datasets/visual-memory/ConvAI2-FLUX-enhanced-Qwen3.5-35B-A3B.EPO-RL-data
Dataset Card for EPO-RL-data
This is the official data collection for paper "EPO: Explicit Policy Optimization for Strategic Reasoning in LLMs via Reinforcement Learning". Please see paper & code for more information:
paper: https://arxiv.org/abs/2502.12486
code: https://github.com/AlibabaResearch/DAMO-ConvAI/tree/main/EPO
Uses
This dataset was used to train the strategic reasoning model (Llama-3-8B-Instruct) via RL reported in our paper.
Note that dump.rdb was… See the full description on the dataset page: https://huggingface.co/datasets/Tongyi-ConvAI/EPO-RL-data.ConvAI2-FLUX-original-Qwen3.5-35B-A3B
Visual Memory Results: convai2-flux-original
This dataset contains the scored output of a visual-memory perplexity experiment.
Experiment metadata
{
"experiment": {
"model_name": "Qwen/Qwen3.5-35B-A3B",
"hf_results_repo": "visual-memory/ConvAI2-FLUX-original-Qwen3.5-35B-A3B",
"results_jsonl": "results/ConvAI2-FLUX-original-Qwen3.5-35B-A3B.jsonl",
"hf_dataset": "visual-memory/ConvAI2-With-Ids_1k-no-redundancy",
"hf_mapping_dataset":… See the full description on the dataset page: https://huggingface.co/datasets/visual-memory/ConvAI2-FLUX-original-Qwen3.5-35B-A3B.ConvAI2-Qwen-original-gemma-4-26B-A4B-it
Visual Memory Results: convai2-qwen-original
This dataset contains the scored output of a visual-memory perplexity experiment.
Experiment metadata
{
"experiment": {
"model_name": "google/gemma-4-26B-A4B-it",
"hf_results_repo": "visual-memory/ConvAI2-Qwen-original-gemma-4-26B-A4B-it",
"results_jsonl": "results/ConvAI2-Qwen-original-gemma-4-26B-A4B-it.jsonl",
"hf_dataset": "visual-memory/ConvAI2-With-Ids_1k-no-redundancy",
"hf_mapping_dataset":… See the full description on the dataset page: https://huggingface.co/datasets/visual-memory/ConvAI2-Qwen-original-gemma-4-26B-A4B-it.ConvAI2-FLUX-enhanced-gemma-4-26B-A4B-it
Visual Memory Results: convai2-flux-enhanced
This dataset contains the scored output of a visual-memory perplexity experiment.
Experiment metadata
{
"experiment": {
"model_name": "google/gemma-4-26B-A4B-it",
"hf_results_repo": "visual-memory/ConvAI2-FLUX-enhanced-gemma-4-26B-A4B-it",
"results_jsonl": "results/ConvAI2-FLUX-enhanced-gemma-4-26B-A4B-it.jsonl",
"hf_dataset": "visual-memory/ConvAI2-With-Ids_1k-no-redundancy",
"hf_mapping_dataset":… See the full description on the dataset page: https://huggingface.co/datasets/visual-memory/ConvAI2-FLUX-enhanced-gemma-4-26B-A4B-it.ConvAI2-ERNIE-enhanced-Qwen3.5-35B-A3B
Visual Memory Results: convai2-ernie-enhanced
This dataset contains the scored output of a visual-memory perplexity experiment.
Experiment metadata
{
"experiment": {
"model_name": "Qwen/Qwen3.5-35B-A3B",
"hf_results_repo": "visual-memory/ConvAI2-ERNIE-enhanced-Qwen3.5-35B-A3B",
"results_jsonl": "results/ConvAI2-ERNIE-enhanced-Qwen3.5-35B-A3B.jsonl",
"hf_dataset": "visual-memory/ConvAI2-With-Ids_1k-no-redundancy",
"hf_mapping_dataset":… See the full description on the dataset page: https://huggingface.co/datasets/visual-memory/ConvAI2-ERNIE-enhanced-Qwen3.5-35B-A3B.ConvAI2-FLUX-original-gemma-4-26B-A4B-it
Visual Memory Results: convai2-flux-original
This dataset contains the scored output of a visual-memory perplexity experiment.
Experiment metadata
{
"experiment": {
"model_name": "google/gemma-4-26B-A4B-it",
"hf_results_repo": "visual-memory/ConvAI2-FLUX-original-gemma-4-26B-A4B-it",
"results_jsonl": "results/ConvAI2-FLUX-original-gemma-4-26B-A4B-it.jsonl",
"hf_dataset": "visual-memory/ConvAI2-With-Ids_1k-no-redundancy",
"hf_mapping_dataset":… See the full description on the dataset page: https://huggingface.co/datasets/visual-memory/ConvAI2-FLUX-original-gemma-4-26B-A4B-it.ConvAI2-ERNIE-enhanced-gemma-4-26B-A4B-it
Visual Memory Results: convai2-ernie-enhanced
This dataset contains the scored output of a visual-memory perplexity experiment.
Experiment metadata
{
"experiment": {
"model_name": "google/gemma-4-26B-A4B-it",
"hf_results_repo": "visual-memory/ConvAI2-ERNIE-enhanced-gemma-4-26B-A4B-it",
"results_jsonl": "results/ConvAI2-ERNIE-enhanced-gemma-4-26B-A4B-it.jsonl",
"hf_dataset": "visual-memory/ConvAI2-With-Ids_1k-no-redundancy"… See the full description on the dataset page: https://huggingface.co/datasets/visual-memory/ConvAI2-ERNIE-enhanced-gemma-4-26B-A4B-it.