Inference llama2 model on the AMD GPUs system

• Install the dependencies.
• Clone the repository.
• Run the project.

cd hip_llama.cpp
make

• Instructions on how to use the project.

./build/apps/llama model.bin -m test -f <input_filename> -o <output_filename>

• Examples of how to inference llama2.

./build/apps/llama /shared/erc/getpTA/main/modelbin/stories110M.bin -m test -f assets/in/gen_in_128.txt -o assets/out/gen_out_128.txt

• Not available yet.

• If you have some issues or feature requests, please open issues or let us know by email of contributors bellow.

GPL-3.0

Reference:

• llama2.c
• llama2.c for Dummies
• llama.cpp