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>
This commit is contained in:
@@ -25,7 +25,6 @@ Usage:
|
||||
import argparse
|
||||
import os
|
||||
import sys
|
||||
import math
|
||||
import random
|
||||
import json
|
||||
from pathlib import Path
|
||||
|
||||
Reference in New Issue
Block a user