diff --git a/nodes/selva_lora_trainer.py b/nodes/selva_lora_trainer.py
index b1a8253..bce709f 100644
--- a/nodes/selva_lora_trainer.py
+++ b/nodes/selva_lora_trainer.py
@@ -271,11 +271,11 @@ class SelvaLoraTrainer:
                     "tooltip": "Path to a step checkpoint (.pt) to resume training from.",
                 }),
                 "seed": ("INT", {"default": 42}),
-                "timestep_mode": (["logit_normal", "uniform"], {
-                    "default": "logit_normal",
+                "timestep_mode": (["uniform", "logit_normal"], {
+                    "default": "uniform",
                     "tooltip": "How to sample training timesteps. "
-                               "logit_normal concentrates steps near t=0.5 (recommended — reduces white noise artifacts). "
-                               "uniform samples all timesteps equally (original behavior).",
+                               "uniform samples all timesteps equally (default, matches original MMAudio training). "
+                               "logit_normal concentrates steps near t=0.5 — reaches lower loss but perceptual improvement is dataset-dependent.",
                 }),
                 "logit_normal_sigma": ("FLOAT", {
                     "default": 1.0, "min": 0.1, "max": 3.0, "step": 0.1,
@@ -305,7 +305,7 @@ class SelvaLoraTrainer:
     def train(self, model, data_dir, output_dir, steps, rank, lr,
               alpha=0.0, target="attn.qkv", batch_size=4, warmup_steps=100,
               grad_accum=1, save_every=500, resume_path="", seed=42,
-              timestep_mode="logit_normal", logit_normal_sigma=1.0):
+              timestep_mode="uniform", logit_normal_sigma=1.0):
 
         torch.manual_seed(seed)
         random.seed(seed)
@@ -451,7 +451,7 @@ class SelvaLoraTrainer:
         data_dir, output_dir, steps, rank, lr,
         alpha_val, target_suffixes, batch_size, warmup_steps,
         grad_accum, save_every, resume_path, seed,
-        timestep_mode="logit_normal", logit_normal_sigma=1.0,
+        timestep_mode="uniform", logit_normal_sigma=1.0,
     ):
         # --- Prepare generator copy with LoRA ---
         generator = copy.deepcopy(model["generator"]).to(device, dtype)
diff --git a/train_lora.py b/train_lora.py
index a603470..8203d67 100644
--- a/train_lora.py
+++ b/train_lora.py
@@ -167,8 +167,8 @@ def main():
                         help="Path to a step checkpoint (.pt) to resume training from.")
     parser.add_argument("--precision",          default="bf16", choices=["bf16", "fp16", "fp32"])
     parser.add_argument("--seed",               type=int,   default=42)
-    parser.add_argument("--timestep_mode",      default="logit_normal", choices=["logit_normal", "uniform"],
-                        help="Timestep sampling distribution. logit_normal reduces white noise artifacts.")
+    parser.add_argument("--timestep_mode",      default="uniform", choices=["uniform", "logit_normal"],
+                        help="Timestep sampling distribution. uniform matches original MMAudio training. logit_normal reaches lower loss but perceptual improvement is dataset-dependent.")
     parser.add_argument("--logit_normal_sigma", type=float, default=1.0,
                         help="Spread of logit-normal distribution (only used with --timestep_mode logit_normal).")
     args = parser.parse_args()