Mirrors/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-26 06:10:29 +01:00
Code Issues Packages Projects Releases Wiki Activity

scripts: update compare-llama-bench.py (#10319)

Browse Source
This commit is contained in:
Johannes Gäßler 2024-11-15 21:19:03 +01:00 committed by GitHub
parent 883d206fbd
commit 4047be74da
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", "cuda", "vulkan", "kompute", "metal", "sycl", "rpc", "gpu_blas",
"blas", "model_filename", "model_type", "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"
"cpu_info", "gpu_info", "backends", "n_gpu_layers", "model_filename", "model_type", "n_batch", "n_ubatch",
"embeddings", "cpu_mask", "cpu_strict", "poll", "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", "vulkan", "kompute", "metal", "sycl", "gpu_blas", "blas", "embeddings", "use_mmap", "no_kv_offload", "flash_attn"]
BOOL_PROPERTIES = ["embeddings", "cpu_strict", "use_mmap", "no_kv_offload", "flash_attn"]
# Header names for the table:
PRETTY_NAMES = {
"cuda": "CUDA", "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",
"cpu_info": "CPU", "gpu_info": "GPU", "backends": "Backends", "n_gpu_layers": "GPU layers",
"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",
"embeddings": "Embeddings", "cpu_mask": "CPU mask", "cpu_strict": "CPU strict", "poll": "Poll",
"n_threads": "Threads", "type_k": "K type", "type_v": "V type", "split_mode": "Split mode", "main_gpu": "Main GPU",
"no_kv_offload": "NKVO", "flash_attn": "FlashAttention", "tensor_split": "Tensor split", "use_mmap": "Use mmap",
}
DEFAULT_SHOW = ["model_type"] # Always show these properties by default.