43 datasets found

A large native one-to-many text dataset for text generation tasks, constructed to accelerate the research of diverse text generation.

A vision-language pre-training dataset, BLIP2, which consists of 100 million image-text pairs.

The E2E dataset contains restaurant reviews labeled by 8 fields including food type, price, and customer rating.

A large-scale dataset for generating prompts that can be used in diffusion models for text-to-text generation tasks.

The dataset used for the text-to-image task consists of 20 context tokens, to which three fixed tokens have been concatenated, representing the static context "the picture of".

The dataset used in this paper is not explicitly described. However, it is mentioned that the authors used the Wikitext-2 dataset for text generation tasks.

Diffusion models have emerged as a power-ful paradigm for generation, obtaining strong performance in various continuous domains. However, applying continuous diffusion models...

MME: A comprehensive evaluation benchmark for multimodal large language models

Mmbench: Is your multi-modal model an all-around player?

A language model that demonstrates capabilities in processing and generating human-like text.

Mmicl: Empowering vision-language model with multi-modal in-context learning

Prompt Highlighter is a novel paradigm for user-model interactions in multi-modal LLMs, offering output control through a token-level highlighting mechanism.

The dataset used for pre-training language models, containing a large collection of text documents.

Commonsense generation aims to generate a realistic sentence describing a daily scene under the given concepts, which is very challenging, since it requires models to have...

Semi-autoregressive simplex-based diffusion language model for text generation and modular control

Word2Vec is a distributed word embedding generator that uses an artificial neural network to learn dense vector representations of words.

The STC dataset is a short text conversation dataset used for evaluating the performance of conversation response generation models.

The dataset used in this paper is Wikitext-103, a general English language corpus containing good and featured Wikipedia articles.

The dataset used in this work is a custom synthetic dataset generated using the liquid-dsp library, containing 600000 examples of each of 13.8 million examples, with SNRs...

The dataset used in the SeqDiffuSeq paper for sequence-to-sequence text generation.