AMD compilation error

[taikai-zz]

import torch torch.cuda.is_available() True device_name = torch.cuda.get_device_name(0) print("device_name:", device_name) device_name: AMD Radeon PRO W6800 print(torch.version) 2.1.0+rocm5.6 print(sys.version) 3.10.12 (main, Jun 11 2023, 05:26:28) [GCC 11.4.0] cd fairseq2/native cmake -GNinja -DCMAKE_HIP_ARCHITECTURES="gfx1030" -DFAIRSEQ2N_USE_HIP=ON -B build

CMake Error at third-party/fmt.cmake:11 (add_subdirectory): The source directory

/data/AI/seamless-m4t-v2-large/fairseq2/native/third-party/fmt

does not contain a CMakeLists.txt file. Call Stack (most recent call first): CMakeLists.txt:193 (fairseq2n_add_fmt)

CMake Error at third-party/fmt.cmake:13 (target_compile_features): Cannot specify compile features for target "fmt" which is not built by this project. Call Stack (most recent call first): CMakeLists.txt:193 (fairseq2n_add_fmt)

CMake Error at third-party/fmt.cmake:15 (set_target_properties): set_target_properties Can not find target to add properties to: fmt Call Stack (most recent call first): CMakeLists.txt:193 (fairseq2n_add_fmt)

CMake Error at third-party/fmt.cmake:24 (target_link_libraries): Cannot specify link libraries for target "fmt" which is not built by this project. Call Stack (most recent call first): CMakeLists.txt:193 (fairseq2n_add_fmt)

-- Configuring incomplete, errors occurred!

May I ask what the problem is.？

[cbalioglu]

@taikai-zz please make sure that all git submodules are initialized: git submodule update --init --recursive.

[cbalioglu]

closing, but feel free to reopen it if the issue persists.