root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-24 10:24:35 +00:00
Code Issues Actions 5 Packages Projects Releases Wiki Activity

scripts: update compare_llama_bench.py [no ci] (#7673)

Browse Source
This commit is contained in:
Johannes Gäßler 2024-05-31 16:26:21 +02:00 committed by GitHub
parent 30e238b246
commit c8047d538f
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 10 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

20
scripts/compare-llama-bench.py
View File

@ -19,22 +19,22 @@ logger = logging.getLogger("compare-llama-bench")
# Properties by which to differentiate results per commit:
KEY_PROPERTIES = [
"cpu_info", "gpu_info", "n_gpu_layers", "main_gpu", "cuda", "opencl", "metal", "gpu_blas",
"blas", "model_filename", "model_type", "model_size", "model_n_params", "n_batch", "n_threads",
"type_k", "type_v", "no_kv_offload", "tensor_split", "n_prompt", "n_gen"
"cpu_info", "gpu_info", "n_gpu_layers", "cuda", "opencl", "vulkan", "kompute", "metal", "sycl", "rpc", "gpu_blas",
"blas", "model_filename", "model_type", "model_size", "model_n_params", "n_batch", "n_ubatch", "embeddings", "n_threads",
"type_k", "type_v", "use_mmap", "no_kv_offload", "split_mode", "main_gpu", "tensor_split", "flash_attn", "n_prompt", "n_gen"
]
# Properties that are boolean and are converted to Yes/No for the table:
BOOL_PROPERTIES = ["cuda", "opencl", "metal", "gpu_blas", "blas"]
BOOL_PROPERTIES = ["cuda", "opencl", "vulkan", "kompute", "metal", "sycl", "gpu_blas", "blas", "embeddings", "use_mmap", "no_kv_offload", "flash_attn"]
# Header names for the table:
PRETTY_NAMES = {
"cuda": "CUDA", "opencl": "OpenCL", "metal": "Metal", "gpu_blas": "GPU BLAS", "blas": "BLAS",
"cpu_info": "CPU", "gpu_info": "GPU", "model_filename": "File", "model_type": "Model",
"model_size": "Model Size [GiB]", "model_n_params": "Num. of Parameters",
"n_batch": "Batch size", "n_threads": "Threads", "type_k": "K type", "type_v": "V type",
"n_gpu_layers": "GPU layers", "main_gpu": "Main GPU", "no_kv_offload": "NKVO",
"tensor_split": "Tensor split"
"cuda": "CUDA", "opencl": "OpenCL", "vulkan": "Vulkan", "kompute": "Kompute", "metal": "Metal", "sycl": "SYCL", "rpc": "RPC",
"gpu_blas": "GPU BLAS", "blas": "BLAS", "cpu_info": "CPU", "gpu_info": "GPU", "model_filename": "File", "model_type": "Model",
"model_size": "Model Size [GiB]", "model_n_params": "Num. of Par.", "n_batch": "Batch size", "n_ubatch": "Microbatch size",
"n_threads": "Threads", "type_k": "K type", "type_v": "V type", "n_gpu_layers": "GPU layers", "split_mode": "Split mode",
"main_gpu": "Main GPU", "no_kv_offload": "NKVO", "flash_attn": "FlashAttention", "tensor_split": "Tensor split",
"use_mmap": "Use mmap", "embeddings": "Embeddings",
}
DEFAULT_SHOW = ["model_type"] # Always show these properties by default.