From 9a3908c55266f8f41a7d26dd02d701c75d8c8cc5 Mon Sep 17 00:00:00 2001
From: Fizza-Mukhtar <fizzamukhtar01@gmail.com>
Date: Fri, 2 Jan 2026 08:42:59 -0800
Subject: [PATCH 1/3] Make llama.cpp CURL support optional during CMake builds

---
 unsloth/save.py | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)

diff --git a/unsloth/save.py b/unsloth/save.py
index ceb36854d2..29d9cdcaff 100644
--- a/unsloth/save.py
+++ b/unsloth/save.py
@@ -130,6 +130,10 @@ ALLOWED_QUANTS = {
     "q3_k_xs": "3-bit extra small quantization",
 }
 
+def has_curl():
+    return shutil.which("curl") is not None
+
+CURL_FLAG = "-DLLAMA_CURL=ON" if has_curl() else "-DLLAMA_CURL=OFF"
 
 def print_quantization_methods():
     for key, value in ALLOWED_QUANTS.items():
@@ -879,8 +883,9 @@ def install_llama_cpp_make_non_blocking():
         # Uses new CMAKE
         n_jobs = max(int(psutil.cpu_count()), 1)  # Use less CPUs since 1.5x faster
         check = os.system(
-            "cmake llama.cpp -B llama.cpp/build -DBUILD_SHARED_LIBS=OFF -DGGML_CUDA=OFF -DLLAMA_CURL=ON"
+            f"cmake llama.cpp -B llama.cpp/build -DBUILD_SHARED_LIBS=OFF -DGGML_CUDA=OFF {CURL_FLAG}"
         )
+
         if check != 0:
             raise RuntimeError(
                 f"*** Unsloth: Failed compiling llama.cpp using os.system(...) with error {check}. Please report this ASAP!"
@@ -991,11 +996,12 @@ def install_llama_cpp_old(version = -10):
     if try_execute(commands) == "CMAKE":
         # Instead use CMAKE
         commands = [
-            "cmake llama.cpp -B llama.cpp/build -DBUILD_SHARED_LIBS=OFF -DGGML_CUDA=OFF -DLLAMA_CURL=ON",
+            "cmake llama.cpp -B llama.cpp/build -DBUILD_SHARED_LIBS=OFF -DGGML_CUDA=OFF {CURL_FLAG}",
             f"cmake --build llama.cpp/build --config Release -j{psutil.cpu_count()*2} --clean-first --target {' '.join(LLAMA_CPP_TARGETS)}",
             "cp llama.cpp/build/bin/llama-* llama.cpp",
             "rm -rf llama.cpp/build",
         ]
+
         try_execute(commands)
 
     # Check if successful
@@ -1037,7 +1043,7 @@ def install_llama_cpp_blocking(use_cuda = False):
     if try_execute(commands) == "CMAKE":
         # Instead use CMAKE
         commands = [
-            "cmake llama.cpp -B llama.cpp/build -DBUILD_SHARED_LIBS=OFF -DGGML_CUDA=OFF -DLLAMA_CURL=ON",
+            "cmake llama.cpp -B llama.cpp/build -DBUILD_SHARED_LIBS=OFF -DGGML_CUDA=OFF {CURL_FLAG}",
             f"cmake --build llama.cpp/build --config Release -j{psutil.cpu_count()*2} --clean-first --target {' '.join(LLAMA_CPP_TARGETS)}",
             "cp llama.cpp/build/bin/llama-* llama.cpp",
             "rm -rf llama.cpp/build",

From 8fa3228590a38379851a770a822adce859c7ba26 Mon Sep 17 00:00:00 2001
From: Fizza-Mukhtar <fizzamukhtar01@gmail.com>
Date: Fri, 2 Jan 2026 08:55:58 -0800
Subject: [PATCH 2/3] Make llama.cpp CURL support optional during CMake builds

---
 unsloth/save.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/unsloth/save.py b/unsloth/save.py
index 29d9cdcaff..359010dbe2 100644
--- a/unsloth/save.py
+++ b/unsloth/save.py
@@ -996,7 +996,7 @@ def install_llama_cpp_old(version = -10):
     if try_execute(commands) == "CMAKE":
         # Instead use CMAKE
         commands = [
-            "cmake llama.cpp -B llama.cpp/build -DBUILD_SHARED_LIBS=OFF -DGGML_CUDA=OFF {CURL_FLAG}",
+            f"cmake llama.cpp -B llama.cpp/build -DBUILD_SHARED_LIBS=OFF -DGGML_CUDA=OFF {CURL_FLAG}",
             f"cmake --build llama.cpp/build --config Release -j{psutil.cpu_count()*2} --clean-first --target {' '.join(LLAMA_CPP_TARGETS)}",
             "cp llama.cpp/build/bin/llama-* llama.cpp",
             "rm -rf llama.cpp/build",
@@ -1043,7 +1043,7 @@ def install_llama_cpp_blocking(use_cuda = False):
     if try_execute(commands) == "CMAKE":
         # Instead use CMAKE
         commands = [
-            "cmake llama.cpp -B llama.cpp/build -DBUILD_SHARED_LIBS=OFF -DGGML_CUDA=OFF {CURL_FLAG}",
+            f"cmake llama.cpp -B llama.cpp/build -DBUILD_SHARED_LIBS=OFF -DGGML_CUDA=OFF {CURL_FLAG}",
             f"cmake --build llama.cpp/build --config Release -j{psutil.cpu_count()*2} --clean-first --target {' '.join(LLAMA_CPP_TARGETS)}",
             "cp llama.cpp/build/bin/llama-* llama.cpp",
             "rm -rf llama.cpp/build",

From c50b7499ff2f6f0485c6afb51d6c6aea209ae28e Mon Sep 17 00:00:00 2001
From: "pre-commit-ci[bot]"
 <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Date: Fri, 2 Jan 2026 16:58:04 +0000
Subject: [PATCH 3/3] [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci
---
 unsloth/save.py | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/unsloth/save.py b/unsloth/save.py
index 359010dbe2..714df7682a 100644
--- a/unsloth/save.py
+++ b/unsloth/save.py
@@ -130,11 +130,14 @@ ALLOWED_QUANTS = {
     "q3_k_xs": "3-bit extra small quantization",
 }
 
+
 def has_curl():
     return shutil.which("curl") is not None
 
+
 CURL_FLAG = "-DLLAMA_CURL=ON" if has_curl() else "-DLLAMA_CURL=OFF"
 
+
 def print_quantization_methods():
     for key, value in ALLOWED_QUANTS.items():
         print(f'"{key}"  ==> {value}')