fix: correct LoRALinear dtype and remove unused import

- LoRALinear now creates lora_A/lora_B with dtype matching the base
  linear's weight, preventing a float32/bf16 mismatch at forward time
  when the generator is loaded in bf16 or fp16.
- Remove unused `import math` from train_lora.py.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

selva_core/model/lora.py

@@ -41,8 +41,9 @@ class LoRALinear(nn.Module):
         if linear.bias is not None:
             linear.bias.requires_grad_(False)
 
-        self.lora_A = nn.Parameter(torch.empty(rank, in_f))
+        ref_dtype   = linear.weight.dtype
-        self.lora_B = nn.Parameter(torch.zeros(out_f, rank))
+        self.lora_A = nn.Parameter(torch.empty(rank, in_f, dtype=ref_dtype))
+        self.lora_B = nn.Parameter(torch.zeros(out_f, rank, dtype=ref_dtype))
         self.scale  = alpha / rank
 
         nn.init.kaiming_uniform_(self.lora_A, a=math.sqrt(5))

train_lora.py

@@ -25,7 +25,6 @@ Usage:
 import argparse
 import os
 import sys
-import math
 import random
 import json
 from pathlib import Path