From d89082ac059111f4cb8445a88084d3f743378f66 Mon Sep 17 00:00:00 2001
From: Erland366 <erland.pg366@gmail.com>
Date: Fri, 14 Mar 2025 22:55:05 +0000
Subject: [PATCH 1/2] Enhance gradient checkpointing and add original model ID
 retrieval in saving utilities

---
 unsloth_zoo/peft_utils.py   |  3 ++-
 unsloth_zoo/saving_utils.py | 25 ++++++++++++++++++++++++-
 2 files changed, 26 insertions(+), 2 deletions(-)

diff --git a/unsloth_zoo/peft_utils.py b/unsloth_zoo/peft_utils.py
index babae8671..7db605fad 100644
--- a/unsloth_zoo/peft_utils.py
+++ b/unsloth_zoo/peft_utils.py
@@ -208,7 +208,8 @@ def requires_grad_pre_hook(module, input):
                 raise RuntimeError("Unsloth: Failed to make input require gradients!")
                 # print(f"  WARNING: Empty list input to {module.__class__.__name__}!") # 
                 # return
-            input[0].requires_grad_(True)
+            if torch.is_floating_point(input[0]):
+                input[0].requires_grad_(True)
         else:
             raise RuntimeError("Unsloth: Failed to make input require gradients!")
     pass
diff --git a/unsloth_zoo/saving_utils.py b/unsloth_zoo/saving_utils.py
index abe99634a..d4b2e4809 100644
--- a/unsloth_zoo/saving_utils.py
+++ b/unsloth_zoo/saving_utils.py
@@ -61,6 +61,7 @@
 pass
 from transformers.modeling_utils import PushToHubMixin
 import json
+import os
 from pathlib import Path
 import tempfile
 from peft import PeftModelForCausalLM
@@ -540,7 +541,13 @@ def merge_and_overwrite_lora(
         model_name = model.config._name_or_path
 
     # Find repository's max shard size and total size of everything
-    file_list = HfFileSystem(token = token).ls(model_name, detail = True)
+    try:
+        file_list = HfFileSystem(token = token).ls(model_name, detail = True)
+    except:
+        original_model_id = get_original_model_id(model_name)
+        model_name = original_model_id
+        file_list = HfFileSystem(token = token).ls(model_name, detail = True)
+
     safetensors_list = []
     max_size_in_bytes = 0
     total_size_in_bytes = 0
@@ -909,6 +916,22 @@ def merge_lora_weights(state_dict, name):
     pass
 pass
 
+def get_original_model_id(local_path: str):
+    import json
+    import os
+    
+    config_path = os.path.join(local_path, "config.json")
+    if os.path.exists(config_path):
+        with open(config_path, "r") as f:
+            config = json.load(f)
+            
+        # Check for _name_or_path that's not a local path
+        # When we load using AutoConfig, the _name_or_path changed into the local path instead
+        if "_name_or_path" in config and not os.path.exists(config["_name_or_path"]):
+            return config["_name_or_path"]
+    
+    return None
+
 # Unsloth Zoo - Utilities for Unsloth
 # Copyright 2023-present Daniel Han-Chen, Michael Han-Chen & the Unsloth team. All rights reserved.
 #

From ae6b1e3ecc57acfc33c85f977fceccd4bdc452a9 Mon Sep 17 00:00:00 2001
From: Erland366 <erland.pg366@gmail.com>
Date: Fri, 14 Mar 2025 23:01:19 +0000
Subject: [PATCH 2/2] In case adapter_config.json as well

---
 unsloth_zoo/saving_utils.py | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/unsloth_zoo/saving_utils.py b/unsloth_zoo/saving_utils.py
index d4b2e4809..1e72fbb5f 100644
--- a/unsloth_zoo/saving_utils.py
+++ b/unsloth_zoo/saving_utils.py
@@ -927,8 +927,16 @@ def get_original_model_id(local_path: str):
             
         # Check for _name_or_path that's not a local path
         # When we load using AutoConfig, the _name_or_path changed into the local path instead
-        if "_name_or_path" in config and not os.path.exists(config["_name_or_path"]):
+        if "_name_or_path" in config:
             return config["_name_or_path"]
+        
+    config_path = os.path.join(local_path, "adapter_config.json")
+    if os.path.exists(config_path):
+        with open(config_path, "r") as f:
+            config = json.load(f)
+            
+        if "base_model_name_or_path" in config:
+            return config["base_model_name_or_path"]
     
     return None