Sensitivity Analysis of Gradient Sparsification Ratio k on CIFAR-10, VGG11

5.5 KB of experimental results analyzing the sensitivity of the gradient sparsification ratio k. The dataset was authored by Jie Guo and last updated on May 18, 2026. It contains results from 20 training rounds on the CIFAR-10 image classification task using the VGG11 model architecture.

Use Cases

• Analyzing the impact of gradient sparsification on model convergence based on the described ratio k.
• Comparing communication efficiency across different sparsification settings for the VGG11 architecture.
• Benchmarking sparsification techniques on the CIFAR-10 image classification task.
• Reproducing sensitivity analysis experiments for gradient compression methods.

Strengths

• Data is shared under a permissive CC-BY-4.0 license.
• The experiment is clearly scoped to 20 training rounds on a standard benchmark (CIFAR-10/VGG11).
• File size is 5.5 KB, indicating a focused and likely well-structured result set.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset's small scale (5.5 KB) suggests it contains summary results, not raw training data.

Provenance

Source

figshare

Collection Method

Likely contains results from a computational experiment measuring the effect of gradient sparsification.

Freshness

Last updated 2026-05-18 17:42:01; freshness should be verified.