I noticed that during training the CPU only uses 1 thread. Since my batches are quite small and I’m swapping often I’m wondering if parallelization is possible and if it actually would increase training speed?
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| How to Train Model Using CPU with MultiProcess Each With Some Number of Thread? | 0 | 959 | May 12, 2023 | |
| Trainer.predict in parallel not supported! | 2 | 655 | November 2, 2022 | |
| Seeking Advice on Optimizing Hardware Resources for Model Training | 3 | 114 | August 4, 2024 | |
| Model *and* data parallelism when training on multiple GPUs? | 0 | 27 | January 22, 2025 | |
| Does the CPU speed matter when training using GPU | 1 | 186 | July 2, 2023 |