diff --git a/convert-falcon-hf-to-gguf.py b/convert-falcon-hf-to-gguf.py
index ec786ff67..271e58972 100755
--- a/convert-falcon-hf-to-gguf.py
+++ b/convert-falcon-hf-to-gguf.py
@@ -11,11 +11,14 @@ import sys
 from pathlib import Path
 from typing import Any
 
-import gguf
 import numpy as np
 import torch
 from transformers import AutoTokenizer  # type: ignore[import]
 
+if 'NO_LOCAL_GGUF' not in os.environ:
+    sys.path.insert(1, str(Path(__file__).parent / 'gguf-py' / 'gguf'))
+import gguf
+
 
 def bytes_to_unicode():
     # ref: https://github.com/openai/gpt-2/blob/master/src/encoder.py
diff --git a/convert-gptneox-hf-to-gguf.py b/convert-gptneox-hf-to-gguf.py
index 852123d99..b9c8b4607 100755
--- a/convert-gptneox-hf-to-gguf.py
+++ b/convert-gptneox-hf-to-gguf.py
@@ -11,11 +11,14 @@ import sys
 from pathlib import Path
 from typing import Any
 
-import gguf
 import numpy as np
 import torch
 from transformers import AutoTokenizer  # type: ignore[import]
 
+if 'NO_LOCAL_GGUF' not in os.environ:
+    sys.path.insert(1, str(Path(__file__).parent / 'gguf-py' / 'gguf'))
+import gguf
+
 # ref: https://github.com/openai/gpt-2/blob/master/src/encoder.py
 
 
diff --git a/convert-llama-ggmlv3-to-gguf.py b/convert-llama-ggmlv3-to-gguf.py
index 3f39bc39e..08ba0c490 100755
--- a/convert-llama-ggmlv3-to-gguf.py
+++ b/convert-llama-ggmlv3-to-gguf.py
@@ -7,9 +7,13 @@ import struct
 import sys
 from pathlib import Path
 
-import gguf
 import numpy as np
 
+import os
+if 'NO_LOCAL_GGUF' not in os.environ:
+    sys.path.insert(1, str(Path(__file__).parent / 'gguf-py' / 'gguf'))
+import gguf
+
 # Note: Does not support GGML_QKK_64
 QK_K = 256
 # Items here are (block size, type size)
diff --git a/convert.py b/convert.py
index 9a39edb99..5cc3f6e66 100755
--- a/convert.py
+++ b/convert.py
@@ -25,10 +25,14 @@ from dataclasses import dataclass
 from pathlib import Path
 from typing import IO, TYPE_CHECKING, Any, Callable, Generator, Iterable, Literal, Sequence, TypeVar
 
-import gguf
 import numpy as np
 from sentencepiece import SentencePieceProcessor  # type: ignore[import]
 
+import os
+if 'NO_LOCAL_GGUF' not in os.environ:
+    sys.path.insert(1, str(Path(__file__).parent / 'gguf-py' / 'gguf'))
+import gguf
+
 if TYPE_CHECKING:
     from typing import TypeAlias
 
diff --git a/examples/train-text-from-scratch/convert-train-checkpoint-to-gguf.py b/examples/train-text-from-scratch/convert-train-checkpoint-to-gguf.py
index 01b3ee92a..a527d6153 100644
--- a/examples/train-text-from-scratch/convert-train-checkpoint-to-gguf.py
+++ b/examples/train-text-from-scratch/convert-train-checkpoint-to-gguf.py
@@ -2,13 +2,16 @@
 # train-text-from-scratch checkpoint --> gguf conversion
 
 import argparse
-import gguf
 import os
 import struct
 import sys
 import numpy as np
 from pathlib import Path
 
+if 'NO_LOCAL_GGUF' not in os.environ:
+    sys.path.insert(1, str(Path(__file__).parent / '..' / '..' / 'gguf-py' / 'gguf'))
+import gguf
+
 # gguf constants
 LLM_KV_OPTIMIZER_TYPE = "optimizer.type"
 LLM_KV_OPTIMIZER_TYPE_ADAM  = "adam"