ci, tests : use cmake to download models and remove libcurl dependency (#18791)

* ci, tests : use cmake to download models and remove libcurl dependency
* llama_dl_model -> llama_download_model
* use EXPECTED_HASH for robust model downloading
* Move llama_download_model to cmake/common.cmake

Signed-off-by: Adrien Gallouët <angt@huggingface.co>

.github/workflows/build.yml

@@ -92,7 +92,7 @@ jobs:
         id: cmake_test
         run: |
           cd build
-          ctest -L 'main|curl' --verbose --timeout 900
+          ctest -L main --verbose --timeout 900
 
   macOS-latest-cmake-x64:
     runs-on: macos-15-intel
@@ -237,7 +237,7 @@ jobs:
         id: cmake_test
         run: |
           cd build
-          ctest -L 'main|curl' --verbose --timeout 900
+          ctest -L main --verbose --timeout 900
 
       - name: Test llama2c conversion
         id: llama2c_test
@@ -1499,7 +1499,7 @@ jobs:
         id: depends
         run: |
           sudo apt-get update
-          sudo apt-get install build-essential libcurl4-openssl-dev
+          sudo apt-get install build-essential
 
       - name: Test
         id: ggml-ci
@@ -1525,7 +1525,7 @@ jobs:
         id: depends
         run: |
           sudo apt-get update
-          sudo apt-get install build-essential libcurl4-openssl-dev
+          sudo apt-get install build-essential
 
       - name: Test
         id: ggml-ci
@@ -1551,7 +1551,7 @@ jobs:
         id: depends
         run: |
           sudo apt-get update
-          sudo apt-get install build-essential libcurl4-openssl-dev
+          sudo apt-get install build-essential
 
       - name: Test
         id: ggml-ci
@@ -1577,7 +1577,7 @@ jobs:
         id: depends
         run: |
           sudo apt-get update
-          sudo apt-get install build-essential libcurl4-openssl-dev
+          sudo apt-get install build-essential
 
       - name: Test
         id: ggml-ci
@@ -1603,7 +1603,7 @@ jobs:
         id: depends
         run: |
           sudo apt-get update
-          sudo apt-get install build-essential libcurl4-openssl-dev
+          sudo apt-get install build-essential
 
       - name: Test
         id: ggml-ci
@@ -1767,7 +1767,7 @@ jobs:
          id: depends
          run: |
            sudo apt-get update
-           sudo apt-get install -y build-essential libcurl4-openssl-dev
+           sudo apt-get install -y build-essential
 
        - name: Test
          id: ggml-ci
@@ -1853,7 +1853,7 @@ jobs:
         id: cmake_test
         run: |
           cd build
-          ctest -L 'main|curl' --verbose --timeout 900
+          ctest -L main --verbose --timeout 900
 
       - name: Test llama2c conversion
         id: llama2c_test
@@ -2129,7 +2129,6 @@ jobs:
            sudo DEBIAN_FRONTEND=noninteractive NEEDRESTART_MODE=a \
            apt-get install -y \
             build-essential \
-            libcurl4-openssl-dev \
             python3-venv \
             gpg \
             wget \

ci/run.sh

@@ -45,7 +45,7 @@ sd=`dirname $0`
 cd $sd/../
 SRC=`pwd`
 
-CMAKE_EXTRA="-DLLAMA_FATAL_WARNINGS=${LLAMA_FATAL_WARNINGS:-ON} -DLLAMA_CURL=ON -DGGML_SCHED_NO_REALLOC=ON"
+CMAKE_EXTRA="-DLLAMA_FATAL_WARNINGS=${LLAMA_FATAL_WARNINGS:-ON} -DLLAMA_CURL=OFF -DGGML_SCHED_NO_REALLOC=ON"
 
 if [ ! -z ${GG_BUILD_METAL} ]; then
     CMAKE_EXTRA="${CMAKE_EXTRA} -DGGML_METAL=ON"

cmake/common.cmake

@@ -33,3 +33,25 @@ function(llama_add_compile_flags)
         endif()
     endif()
 endfunction()
+
+function(llama_download_model NAME HASH)
+    set(DEST "${CMAKE_BINARY_DIR}/${NAME}")
+    get_filename_component(DEST_DIR "${DEST}" DIRECTORY)
+    file(MAKE_DIRECTORY "${DEST_DIR}")
+    if(NOT EXISTS "${DEST}")
+        message(STATUS "Downloading ${NAME} from ggml-org/models...")
+    endif()
+    file(DOWNLOAD
+        "https://huggingface.co/ggml-org/models/resolve/main/${NAME}?download=true"
+        "${DEST}"
+        TLS_VERIFY ON
+        EXPECTED_HASH ${HASH}
+        STATUS status
+    )
+    list(GET status 0 code)
+    if(NOT code EQUAL 0)
+        list(GET status 1 msg)
+        message(FATAL_ERROR "Failed to download ${NAME}: ${msg}")
+    endif()
+    set(LLAMA_DOWNLOAD_MODEL "${DEST}" PARENT_SCOPE)
+endfunction()

examples/eval-callback/CMakeLists.txt

@@ -6,10 +6,8 @@ target_compile_features(${TARGET} PRIVATE cxx_std_17)
 
 set(TEST_TARGET test-eval-callback)
 if(NOT ${CMAKE_SYSTEM_PROCESSOR} MATCHES "s390x")
-        add_test(NAME ${TEST_TARGET}
-                        COMMAND llama-eval-callback --hf-repo ggml-org/models --hf-file tinyllamas/stories260K.gguf --model stories260K.gguf --prompt hello --seed 42 -ngl 0)
+    llama_download_model("tinyllamas/stories15M-q4_0.gguf" SHA256=66967fbece6dbe97886593fdbb73589584927e29119ec31f08090732d1861739)
 else()
-        add_test(NAME ${TEST_TARGET}
-                        COMMAND llama-eval-callback --hf-repo ggml-org/models --hf-file tinyllamas/stories260K-be.gguf --model stories260K-be.gguf --prompt hello --seed 42 -ngl 0)
+    llama_download_model("tinyllamas/stories15M-be.Q4_0.gguf" SHA256=9aec857937849d976f30397e97eb1cabb53eb9dcb1ce4611ba8247fb5f44c65d)
 endif()
-set_property(TEST ${TEST_TARGET} PROPERTY LABELS eval-callback curl)
+add_test(NAME ${TEST_TARGET} COMMAND llama-eval-callback -m "${LLAMA_DOWNLOAD_MODEL}" --prompt hello --seed 42 -ngl 0)

tests/CMakeLists.txt

@@ -202,15 +202,13 @@ llama_build_and_test(
 llama_build_and_test(test-regex-partial.cpp)
 
 if (NOT ${CMAKE_SYSTEM_PROCESSOR} MATCHES "s390x")
-    llama_build_and_test(test-thread-safety.cpp ARGS -hf ggml-org/models -hff tinyllamas/stories15M-q4_0.gguf -ngl 99 -p "The meaning of life is" -n 128 -c 256 -ub 32 -np 4 -t 2)
+    llama_download_model("tinyllamas/stories15M-q4_0.gguf" SHA256=66967fbece6dbe97886593fdbb73589584927e29119ec31f08090732d1861739)
 else()
-    llama_build_and_test(test-thread-safety.cpp ARGS -hf ggml-org/models -hff tinyllamas/stories15M-be.Q4_0.gguf -ngl 99 -p "The meaning of life is" -n 128 -c 256 -ub 32 -np 4 -t 2)
+    llama_download_model("tinyllamas/stories15M-be.Q4_0.gguf" SHA256=9aec857937849d976f30397e97eb1cabb53eb9dcb1ce4611ba8247fb5f44c65d)
 endif()
+llama_build_and_test(test-thread-safety.cpp ARGS -m "${LLAMA_DOWNLOAD_MODEL}" -ngl 99 -p "The meaning of life is" -n 128 -c 256 -ub 32 -np 4 -t 2)
 
-# this fails on windows (github hosted runner) due to curl DLL not found (exit code 0xc0000135)
-if (NOT WIN32)
-    llama_build_and_test(test-arg-parser.cpp)
-endif()
+llama_build_and_test(test-arg-parser.cpp)
 
 if (NOT LLAMA_SANITIZE_ADDRESS AND NOT GGML_SCHED_NO_REALLOC)
   # TODO: repair known memory leaks
@@ -225,11 +223,7 @@ llama_build_and_test(test-backend-sampler.cpp   LABEL "model")
 
 # Test for state restore with fragmented KV cache
 # Requires a model, uses same args pattern as test-thread-safety
-if (NOT ${CMAKE_SYSTEM_PROCESSOR} MATCHES "s390x")
-    llama_build_and_test(test-state-restore-fragmented.cpp LABEL "model" ARGS -hf ggml-org/models -hff tinyllamas/stories15M-q4_0.gguf)
-else()
-    llama_build_and_test(test-state-restore-fragmented.cpp LABEL "model" ARGS -hf ggml-org/models -hff tinyllamas/stories15M-be.Q4_0.gguf)
-endif()
+llama_build_and_test(test-state-restore-fragmented.cpp LABEL "model" ARGS -m "${LLAMA_DOWNLOAD_MODEL}")
 
 if (NOT GGML_BACKEND_DL)
     # these tests use the backends directly and cannot be built with dynamic loading

tests/test-arg-parser.cpp

@@ -173,7 +173,7 @@ int main(void) {
     assert(params.cpuparams.n_threads == 1010);
 #endif // _WIN32
 
-    printf("test-arg-parser: test curl-related functions\n\n");
+    printf("test-arg-parser: test download functions\n\n");
     const char * GOOD_URL = "http://ggml.ai/";
     const char * BAD_URL  = "http://ggml.ai/404";