Fix: replace FP4 Inf with 24 (avoid NaN in dequant)

single_shot_inference.py

@@ -28,7 +28,7 @@ PROMPT = "The capital of France is"
 # FP4 E2M1 lookup table: index → float value (unsigned)
 # E2M1: 1-bit sign, 2-bit exp (bias=1), 1-bit mantissa
 # Values: 0, 2, 3, 4, 6, 8, 12, Inf  (for exp 00,01,10,11 × mantissa 0,1)
-FP4_LUT = torch.tensor([0., 2., 3., 4., 6., 8., 12., float('inf')])
+FP4_LUT = torch.tensor([0., 2., 3., 4., 6., 8., 12., 24.])  # E2M1: last value (0b111) is nominally Inf, use 24 (2^(2+1) * 1.5)
 
 def dequant_nvfp4_weight(
     weight: torch.Tensor,           # (out, in/2) uint8