From 34a9ef6c6c6ba28fa755eee55ee5049b7399eb6d Mon Sep 17 00:00:00 2001 From: vodkaslime <646329483@qq.com> Date: Sun, 19 Nov 2023 17:20:18 +0800 Subject: [PATCH 1/3] fix: readme --- README.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/README.md b/README.md index 4de06476569f9..b5a6a52f1dd23 100644 --- a/README.md +++ b/README.md @@ -214,7 +214,7 @@ cd llama.cpp ### Build -In order to build llama.cpp you have three different options. +In order to build llama.cpp you have different options. - Using `make`: - On Linux or MacOS: @@ -320,7 +320,7 @@ mpirun -hostfile hostfile -n 3 ./main -m ./models/7B/ggml-model-q4_0.gguf -n 128 ### BLAS Build -Building the program with BLAS support may lead to some performance improvements in prompt processing using batch sizes higher than 32 (the default is 512). BLAS doesn't affect the normal generation performance. There are currently three different implementations of it: +Building the program with BLAS support may lead to some performance improvements in prompt processing using batch sizes higher than 32 (the default is 512). BLAS doesn't affect the normal generation performance. There are currently various different implementations of it: - #### Accelerate Framework: From 5d21a827869789b091b4ba4c2a55a65169e2312a Mon Sep 17 00:00:00 2001 From: vodkaslime <646329483@qq.com> Date: Mon, 20 Nov 2023 00:40:45 +0800 Subject: [PATCH 2/3] chore: resolve comments --- README.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/README.md b/README.md index b5a6a52f1dd23..03ae2f7533b8d 100644 --- a/README.md +++ b/README.md @@ -214,7 +214,7 @@ cd llama.cpp ### Build -In order to build llama.cpp you have different options. +In order to build llama.cpp you have three different options. - Using `make`: - On Linux or MacOS: @@ -320,7 +320,7 @@ mpirun -hostfile hostfile -n 3 ./main -m ./models/7B/ggml-model-q4_0.gguf -n 128 ### BLAS Build -Building the program with BLAS support may lead to some performance improvements in prompt processing using batch sizes higher than 32 (the default is 512). BLAS doesn't affect the normal generation performance. There are currently various different implementations of it: +Building the program with BLAS support may lead to some performance improvements in prompt processing using batch sizes higher than 32 (the default is 512). BLAS doesn't affect the normal generation performance. There are currently several different implementations of it: - #### Accelerate Framework: From ee66c69dbac3daff23f503d4695d2109553e2da2 Mon Sep 17 00:00:00 2001 From: vodkaslime <646329483@qq.com> Date: Mon, 20 Nov 2023 01:03:24 +0800 Subject: [PATCH 3/3] chore: resolve comments --- README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README.md b/README.md index 03ae2f7533b8d..40f2661004a30 100644 --- a/README.md +++ b/README.md @@ -320,7 +320,7 @@ mpirun -hostfile hostfile -n 3 ./main -m ./models/7B/ggml-model-q4_0.gguf -n 128 ### BLAS Build -Building the program with BLAS support may lead to some performance improvements in prompt processing using batch sizes higher than 32 (the default is 512). BLAS doesn't affect the normal generation performance. There are currently several different implementations of it: +Building the program with BLAS support may lead to some performance improvements in prompt processing using batch sizes higher than 32 (the default is 512). Support with CPU-only BLAS implementations doesn't affect the normal generation performance. We may see generation performance improvements with GPU-involved BLAS implementations, e.g. cuBLAS, hipBLAS and CLBlast. There are currently several different BLAS implementations available for build and use: - #### Accelerate Framework: