Pith. sign in

REVIEW 5 cited by

Backpropagating through Fr\'echet Inception Distance

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2009.14075 v2 pith:JDEKC3BB submitted 2020-09-29 cs.LG stat.ML

Alexander Mathiasen , Frederik Hvilsh{o}j This is my paper
classification cs.LGstat.ML
keywords generativedistanceechetinceptionlossmodelsadditionaladversarial
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

The Fr\'echet Inception Distance (FID) has been used to evaluate hundreds of generative models. We introduce FastFID, which can efficiently train generative models with FID as a loss function. Using FID as an additional loss for Generative Adversarial Networks improves their FID.

Discussion (0). Sign in to comment.

Forward citations

Cited by 5 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Beyond Token-Level Cross-Entropy: Fr\'echet Distributional Post-Training for Autoregressive Image Generation

    cs.CV 2026-08 conditional novelty 6.0 of 10

    FD-loss post-training with detached rollout replay and a probability-level straight-through estimator improves FID and FD_r6 across eight ImageNet configurations.

  2. Amortized Moment Matching for Visual Generation

    cs.LG 2026-07 accept novelty 6.0 of 10

    Amortized Fréchet Distance uses neural nets to match conditional means and covariances, yielding stronger one-step visual generators than explicit FD-loss or multi-step teachers.

  3. Joint Training of Image Generator and Detector for Road Defect Detection

    cs.CV 2025-09 conditional novelty 6.0 of 10

    JTGD trains a CycleGAN to synthesize road defects and jointly hardens an InternImage-T detector, reporting 63.13 average F1 on RDD2022 with 49M parameters versus 59.20 F1 and 253M parameters for Faster Swin.

  4. Improving atomic force microscopy structure discovery via style-translation

    cond-mat.mtrl-sci 2025-09 conditional novelty 6.0 of 10

    Style-translated simulated AFM images improve machine-learning structure discovery on experimental AFM data, as judged by agreement with simulation-derived structural distributions.

  5. Improving Medical Image Generative Models with Fr\'echet Distance Loss

    cs.CV 2026-07 conditional novelty 5.0 of 10

    Adding a Fréchet distance loss during generative model finetuning improves realism of synthetic medical images and downstream tumor segmentation.

Pith tools