clm_3d08a6b5We show that diffusion models can achieve image sample quality superior to the current state-of-the-art generative models.
arXiv 2105.05233 · cs.LG cs.AI cs.CV
Read on arXiv · arXiv non-exclusive license
Claims were extracted automatically by reducer-1 on 2026-11-12. The paper's license does not allow republishing it, so long claims are shortened to a quotation. Read the full text on arXiv.
clm_3d08a6b5We show that diffusion models can achieve image sample quality superior to the current state-of-the-art generative models.
res_93211d1bWe achieve an FID of 2.97 on ImageNet 128 ×128, 4.59 on ImageNet 256×256, and 7.72 on ImageNet 512 ×512, and we match BigGAN-deep even with as few as 25 … (excerpt)
res_331d7da3Finally, we find that classifier guidance combines well with upsampling diffusion models, further improving FID to 3.94 on ImageNet 256×256 and 3.85 on ImageNet 512×512.
clm_f7f5df81With these improvements, we achieve a new state-of-the-art, surpassing GANs on several different metrics and datasets.
res_35d66ff3In Section 2, we give a brief background of diffusion models based on Ho et al. [ 25] and the improvements from Nichol and Dhariwal [ 43] and Song et … (excerpt)
res_d4773615In Section 3, we introduce simple architecture improvements that give a substantial boost to FID.
res_4291c8d8We find that a single hyperparameter, the scale of the classifier gradients, can be tuned to trade off diversity for fidelity, and we can increase this gradient scale factor by … (excerpt)
res_de01cb50Finally, in Section 5 we show that models with our improved architecture achieve state-of-the-art on unconditional image synthesis tasks, and with classifier guidance achieve state-ofthe-art on conditional image synthesis.
res_f0b54bbfWe also compare our improved models to upsampling stacks, finding that the two approaches give complementary improvements and that combining them gives the best results on ImageNet 256×256 and 512×512.
clm_85ab644eWe show the same result on ImageNet 128×128, finding that architecture can indeed give a substantial boost to sample quality on much larger and more diverse datasets at a higher … (excerpt)
res_055da973We observe in Figure 2 that while increased depth helps performance, it increases training time and takes longer to reach the same performance as a wider model, so we opt … (excerpt)
res_89a26ac2Table 2 shows our results, indicating that more heads or fewer channels per head improves FID.
res_2c2a7ccdIn Table 3, we explicitly ablate this choice, and find that the adaptive group normalization layer indeed improved FID.
res_8bbe7201In the rest of the paper, we use this final improved model architecture as our default: variable width with 2 residual blocks per resolution, multiple heads with 64 channels per … (excerpt)
res_7c3922d8Table 4 shows that the sample quality of both unconditional and conditional models can be greatly improved by classifier guidance.
res_a58bbb70Table 4 also shows that classifier guidance improves precision at the cost of recall, thus introducing a trade-off in sample fidelity versus diversity.
res_887ecf9eTo evaluate our improved model architecture on unconditional image generation, we train separate diffusion models on three LSUN [ 71] classes: bedroom, horse, and cat.
res_34e53196With the improved architecture, we already obtain state-of-the-art image generation on LSUN and ImageNet 64×64.
clm_33a9d30cFor higher resolution ImageNet, we observe that classifier guidance allows our models to substantially outperform the best GANs.
res_0999f08aIn Table 6, we show that guidance and upsampling improve sample quality along different axes.
res_e6abb2d9For upsampling, we use the upsampling stack from Nichol and Dhariwal [43] combined with our architecture improvements, which we refer to as ADM-U.
clm_46095b1bWe have shown that diffusion models, a class of likelihood-based models with a stationary training objective, can obtain better sample quality than state-of-the-art GANs.
clm_c38471b5Our improved architecture is sufficient to achieve this on unconditional image generation tasks, and our classifier guidance technique allows us to do so on class-conditional tasks.
clm_068f92f2Finally, by combining guidance with upsampling, we can further improve sample quality on high-resolution conditional image synthesis.
clm_13a0086dIn this section, we demonstrate that we can achieve results better than StyleGAN2 and BigGAN-deep with the same or lower compute budget.
res_f90a844dIn this section, we show that conditional sampling can be achieved with a transition operator proportional to pθ(xt|xt+1)pφ(y|xt), where pθ(xt|xt+1) approximates q(xt|xt+1) and pφ(y|xt) approximates the label distribution for a … (excerpt)
res_eaed9c1f28 J Using Fewer Sampling Steps on LSUN We initially found that our LSUN models achieved much better results when sampling with 1000 steps rather than 250 steps, contrary to … (excerpt)
Record your result in the Mearc app on your own machine, then use Publish this result with this paper's id (2105.05233) and the claim id shown next to the claim. Mearc creates a file that holds only what you choose to publish. Add it to the public claim data by pull request; a maintainer other than you reviews it before it counts.