summaryrefslogtreecommitdiff
path: root/src/freq_ablation.py
diff options
context:
space:
mode:
Diffstat (limited to 'src/freq_ablation.py')
-rw-r--r--src/freq_ablation.py2
1 files changed, 1 insertions, 1 deletions
diff --git a/src/freq_ablation.py b/src/freq_ablation.py
index c011140..1227ae6 100644
--- a/src/freq_ablation.py
+++ b/src/freq_ablation.py
@@ -33,7 +33,7 @@ def train(data_dir, out_dir, max_iters=5000):
model = GPT(config).to(device)
print(f" Params: {sum(p.numel() for p in model.parameters())/1e6:.2f}M")
- optimizer = model.configure_optimizers(weight_decay=0.1, lr=1e-3,
+ optimizer = model.configure_optimizers(weight_decay=0.1, learning_rate=1e-3,
betas=(0.9, 0.99), device_type='cuda')
os.makedirs(out_dir, exist_ok=True)
best_val = 1e9