diff --git a/vall_e/data.py b/vall_e/data.py
index 0772c13..5c1c6ee 100755
--- a/vall_e/data.py
+++ b/vall_e/data.py
@@ -293,6 +293,7 @@ class Dataset(_Dataset):
 		prom_length = 0
 		if cfg.experimental:
 			trim_length = random.randint(75 * 3, 75 * 9) # [3 seconds, 9 seconds]
+			#trim_length =  max(2, int(np.random.normal(loc=5, scale=1.25) * 75))
 		else:
 			trim_length =  int(cfg.dataset.prompt_duration * 75) + random.randint(-75, 75)
 
diff --git a/vall_e/models/ar_nar.py b/vall_e/models/ar_nar.py
index ccd5836..6731261 100644
--- a/vall_e/models/ar_nar.py
+++ b/vall_e/models/ar_nar.py
@@ -10,6 +10,8 @@ from einops import rearrange
 from torch import Tensor
 from tqdm import trange
 
+from ..emb.qnt import trim
+
 class AR_NAR(Base):
 	@property
 	def causal(self):
@@ -113,7 +115,11 @@ class AR_NAR(Base):
 
 				targ_list = [r[..., l] for r, l in zip(resps_list, quant_levels)] # ensures we only have 1 RVQ-bin (our target)
 				resps_list = [r if l == 0 else r[..., :l] for r, l in zip(resps_list, quant_levels)] # yes I can just do min(1, l)
-				quant_levels.to(device=device)
+				
+				if cfg.experimental:
+					proms_list = [ r if l == 0 else trim(r, 75 * 3) for r, l in zip(proms_list, quant_levels) ] # trim input prompt to 3 seconds
+				
+				#quant_levels.to(device=device)
 
 				return super().forward(
 					text_list=text_list,
diff --git a/vall_e/models/nar.py b/vall_e/models/nar.py
index 46a984c..100bdfa 100755
--- a/vall_e/models/nar.py
+++ b/vall_e/models/nar.py
@@ -110,7 +110,7 @@ class NAR(Base):
 			prev_list = [o[..., : l + 1] for o, l in zip(resps_list, quant_levels)]
 			targ_list = [o[..., l + 1] for o, l in zip(resps_list, quant_levels)]
 
-			quant_levels = quant_levels.to(device=device)
+			#quant_levels = quant_levels.to(device=device)
 
 			logits = super().forward(
 				text_list=text_list,