Add distributed end-to-end training

README.md

@@ -83,6 +83,14 @@ torchrun --standalone --nproc_per_node=4 train_exposure_autoencoder.py \
 
 In DDP mode, `--batch_size` is the global batch size and must be divisible by
 the number of processes.
+
+The end-to-end next-step trainer supports the same DDP launch pattern:
+
+```bash
+torchrun --standalone --nproc_per_node=4 train_next_step.py \
+  --exposure_cache_dir ukb_exposure_cache \
+  --batch_size 128
+```
 Training-channel statistics are cached at
 `<exposure_cache_dir>/train_channel_stats.npz`; use
 `--recompute_channel_stats` only when a forced refresh is needed.