HACK++ is a head-aware KV cache compression framework for VAR models that decouples current-scale attention from historical cache under adaptive per-head budgets to achieve near-lossless generation at 30% attention and 10% cache budgets.
hub Baseline reference
Playground v2.5: Three Insights towards Enhancing Aesthetic Quality in Text-to-Image Generation
Baseline reference. 60% of citing Pith papers use this work as a benchmark or comparison.
47
Pith papers citing it
4
external citations · Pith
Baseline
60% of classified citations
abstract
In this work, we share three insights for achieving state-of-the-art aesthetic quality in text-to-image generative models. We focus on three critical aspects for model improvement: enhancing color and contrast, improving generation across multiple aspect ratios, and improving human-centric fine details. First, we delve into the significance of the noise schedule in training a diffusion model, demonstrating its profound impact on realism and visual fidelity. Second, we address the challenge of accommodating various aspect ratios in image generation, emphasizing the importance of preparing a balanced bucketed dataset. Lastly, we investigate the crucial role of aligning model outputs with human preferences, ensuring that generated images resonate with human perceptual expectations. Through extensive analysis and experiments, Playground v2.5 demonstrates state-of-the-art performance in terms of aesthetic quality under various conditions and aspect ratios, outperforming both widely-used open-source models like SDXL and Playground v2, and closed-source commercial systems such as DALLE 3 and Midjourney v5.2. Our model is open-source, and we hope the development of Playground v2.5 provides valuable guidelines for researchers aiming to elevate the aesthetic quality of diffusion-based image generation models.
hub tools
citation-role summary
baseline 5
background 3
dataset 1
method 1
citation-polarity summary
representative citing papers
TwinQuant learns quantization-friendly subspaces for 4-bit LLM weights via manifold optimization and a fused kernel, preserving near-FP16 accuracy with up to 1.8x speedup on LLaMA3 and Qwen3 models.
T2I-FactualBench is a new three-tier benchmark for factuality of knowledge-intensive concepts in T2I models, using multi-round VQA evaluation to show SOTA models need improvement.
Janus decouples visual encoding into task-specific pathways inside a single autoregressive transformer to unify multimodal understanding and generation while outperforming earlier unified models.
Scaled vanilla autoregressive models based on Llama achieve 2.18 FID on ImageNet 256x256 image generation, beating popular diffusion models without visual inductive biases.
Per-position AR prediction entropy jointly drives speculative AR decoding and an anti-symmetric single-step drift decoder, yielding 3.8–5.5× faster hybrid visual generation without distillation.
DiT-Pruning squares the weight term in the pruning saliency metric and adapts per-layer versus per-channel granularity, claiming strong quality retention for diffusion transformers at high sparsity.
Fleet achieves dynamic few-shot adaptation for AIGI detection via avoidance routing in decoupled subspaces, raising accuracy from 20.4% to 73.1% on new generators like Doubao Seedream 4.0 with 10 shots.
BAFIS supplies a new dataset and human-feedback framework demonstrating systematic gender and ethnicity biases in occupational image generation by five text-to-image models, with partial alignment between automated metrics and subjective ratings compared to German employment statistics.
Qwen-RobotWorld is a language-conditioned video world model using Double-Stream MMDiT, an 8.6M-frame embodied corpus, and progressive curriculum training that ranks first on EWMBench and DreamGen Bench.
HierEdit enables efficient 4K image editing via low-resolution proxy localization followed by hierarchical local-window diffusion that reuses unaltered regions as conditioning.
Latent Action Control learns unobserved action trajectories via variational alignment and GRPO to inject reasoning into flow-based image generation, yielding gains on compositional benchmarks.
DiRotQ uses PCA-based rotation-aware activation quantization combined with GPTQ to achieve better FID and PSNR in 4-bit diffusion transformers than prior methods like SVDQuant.
ElasticDiT introduces an elastic DiT architecture with adjustable spatial compression and block depth plus Shift Sparse Block Attention and a distilled VAE to enable a single model to cover multiple fidelity-latency points for high-resolution image generation on mobile devices.
By requiring and using highly discriminative LLM text features, the work enables the first effective one-step text-conditioned image generation with MeanFlow.
IncreFA uses hierarchical constraints with learnable orthogonal priors and a latent memory bank to enable continual adaptation for attributing images to new generative models, reporting SOTA accuracy and 98.93% unseen detection on a 28-model benchmark.
Combining multiple safe visual concepts in one prompt can make text-to-image models produce harmful images, and stronger instruction-following models fail more often.
APEX derives self-adversarial gradients from condition-shifted velocity fields in flow models to achieve high-fidelity one-step generation, outperforming much larger models and multi-step teachers.
A 17B-parameter sparse MoE diffusion transformer activates 2B parameters per pass and reaches competitive quality on image generation benchmarks without post-training.
BiasIG is a multi-dimensional benchmark for social biases in T2I models that shows debiasing interventions frequently cause confounding discrimination effects.
Lotus-2 is a two-stage deterministic adaptation of diffusion priors that achieves state-of-the-art monocular depth estimation with only 59K training samples.
A 6B single-stream diffusion transformer trained with heavily curated data reaches top open-source image-generation quality in 314K H800 GPU hours, releasing Turbo and Edit variants.
PixelDiT generates images in pixel space with a dual-level transformer and reaches 1.61 FID on ImageNet 256, outperforming prior pixel-space models.
VFM-VAE uses a frozen VFM directly as LDM tokenizer via a custom decoder, reaching gFID 2.22 in 80 epochs and 1.62 after 640 epochs.
citing papers explorer
-
MONET: A Massive, Open, Non-redundant and Enriched Text-to-image dataset
MONET is an open 104.9M image-text pair dataset created via safety filtering, deduplication, and multi-VLM recaptioning from 2.9B raw pairs, validated by training a competitive 4B-parameter latent diffusion model.