Upload merge_lora.py

merge_lora.py

@@ -0,0 +1,56 @@
+import os
+from safetensors.torch import load_file, save_file
+from transformers import AutoTokenizer, AutoModel
+from diffusers import StableDiffusionPipeline
+import torch
+
+file1_path = r"D:\hansika lora\hansika lora.safetensors"
+file2_path = r"D:\first lora\first lora.safetensors"
+file3_path = r"D:\NSFW_master.safetensors"
+merged_file_path = r"D:\new lora\new lora.safetensors"
+
+print("File 1 Path:", file1_path)
+print("File exists:", os.path.exists(file1_path))
+print("File size:", os.path.getsize(file1_path))
+
+def load_tensors(file_path):
+    return load_file(file_path)
+
+try:
+    tensors1 = load_tensors(file1_path)
+    tensors2 = load_tensors(file2_path)
+    tensors3 = load_tensors(file3_path)
+
+    # Merge tensors
+    merged_tensors = {**tensors1, **tensors2, **tensors3}
+
+    # Save merged tensors
+    save_file(merged_tensors, merged_file_path)
+
+    print(f"Merged file saved at: {merged_file_path}")
+
+    # Validate the merged file
+    merged_load = load_file(merged_file_path)
+    print(merged_load.keys())
+
+    try:
+        # Load the tokenizer and model
+        tokenizer = AutoTokenizer.from_pretrained(merged_file_path)
+        model = AutoModel.from_pretrained(merged_file_path, from_tf=False, from_safetensors=True)
+        pipeline = StableDiffusionPipeline.from_pretrained(merged_file_path, torch_dtype=torch.float16)
+        print("It is successfully loaded!")
+    except Exception as e:
+        print(f"Error: {e}")
+except Exception as e:
+    print(f"An error occurred: {e}")
+
+
+
+
+
+
+
+
+
+
+