Sep 23, 2026 · Artificial Intelligence

GPU Starved? DataLoader Optimizations Cut Embodied Model Training Time 2x

Baidu Baige optimizes DataLoader across sampling, indexing, decoding, and post-processing for embodied model training, reducing FastWAM single-sample video processing from 133ms to 63ms, stabilizing indexing at 0.61ms across 27,500 files, and boosting Qwen3-VL-30B throughput 1.68x via load-balanced sampling.

Baidu BaigeDataLoader optimizationFastWAM

0 likes · 20 min read

GPU Starved? DataLoader Optimizations Cut Embodied Model Training Time 2x