Go to file

Jiarui Li a832a45c62 config: Tune hyperparameters for multi-GPU training

Increase model size (n_embd, n_layer, n_head) for the multi-GPU configuration.

Explicitly set AdamW betas to (0.9, 0.99).

2025-10-17 15:37:42 +08:00

.gitattributes

2025-10-16 14:24:56 +08:00

.gitignore

2025-10-17 14:09:34 +08:00

GEMINI.md

2025-10-16 14:21:36 +08:00

LICENSE

Initial commit

2025-10-15 13:54:52 +08:00

models.py

2025-10-17 12:04:50 +08:00

prepare_data.py

2025-10-16 14:21:36 +08:00

README.md

Initial commit

2025-10-15 13:54:52 +08:00

requirements.txt

2025-10-17 10:31:12 +08:00

train_multigpu.py

2025-10-17 15:37:42 +08:00

train.py

2025-10-17 14:09:34 +08:00

ukb_real_train.bin

2025-10-16 14:24:56 +08:00

ukb_real_val.bin

2025-10-16 14:24:56 +08:00

utils.py

2025-10-16 14:21:36 +08:00

DeepHealth