Skip to content

Commit

Permalink
updated llama.cpp
Browse files Browse the repository at this point in the history
  • Loading branch information
@MarcusDunn
Marcus Dunn authored and MarcusDunn committed Jan 30, 2024
1 parent 0405f96 commit dc3716d
Showing 1 changed file with 1 addition and 1 deletion.
2 changes: 1 addition & 1 deletion llama-cpp-sys-2/llama.cpp
Submodule llama.cpp updated 86 files
+32 −0 .devops/server-cuda.Dockerfile
+25 −0 .devops/server-intel.Dockerfile
+45 −0 .devops/server-rocm.Dockerfile
+20 −0 .devops/server.Dockerfile
+1 −0 .ecrc
+61 −1 .github/workflows/build.yml
+4 −0 .github/workflows/docker.yml
+3 −0 .gitmodules
+226 −11 CMakeLists.txt
+13 −0 Makefile
+25 −3 README.md
+252 −0 README_sycl.md
+4 −0 ci/README.md
+11 −0 ci/run.sh
+15 −11 common/common.cpp
+81 −0 convert-hf-to-gguf.py
+4 −1 convert.py
+3 −0 examples/CMakeLists.txt
+8 −3 examples/llama-bench/llama-bench.cpp
+8 −0 examples/server/README.md
+15 −0 examples/server/oai.hpp
+63 −3 examples/server/server.cpp
+9 −0 examples/sycl/CMakeLists.txt
+47 −0 examples/sycl/README.md
+20 −0 examples/sycl/build.sh
+11 −0 examples/sycl/ls-sycl-device.cpp
+19 −0 examples/sycl/run-llama2.sh
+3 −3 flake.lock
+82 −24 ggml-alloc.c
+6 −0 ggml-backend-impl.h
+112 −2 ggml-backend.c
+3 −0 ggml-backend.h
+3 −0 ggml-cuda.cu
+1,990 −0 ggml-kompute.cpp
+46 −0 ggml-kompute.h
+24 −20 ggml-metal.m
+11 −0 ggml-opencl.cpp
+15,199 −0 ggml-sycl.cpp
+27 −0 ggml-sycl.h
+61,420 −0 ggml-vulkan-shaders.hpp
+5,176 −0 ggml-vulkan.cpp
+34 −0 ggml-vulkan.h
+61 −4 ggml.c
+2 −0 ggml.h
+2,362 −0 ggml_vk_generate_shaders.py
+23 −1 gguf-py/gguf/constants.py
+1 −0 kompute
+102 −0 kompute-shaders/common.comp
+58 −0 kompute-shaders/op_add.comp
+25 −0 kompute-shaders/op_addrow.comp
+52 −0 kompute-shaders/op_cpy_f16_f16.comp
+52 −0 kompute-shaders/op_cpy_f16_f32.comp
+52 −0 kompute-shaders/op_cpy_f32_f16.comp
+52 −0 kompute-shaders/op_cpy_f32_f32.comp
+30 −0 kompute-shaders/op_diagmask.comp
+22 −0 kompute-shaders/op_gelu.comp
+17 −0 kompute-shaders/op_getrows.comp
+31 −0 kompute-shaders/op_getrows_f16.comp
+38 −0 kompute-shaders/op_getrows_q4_0.comp
+39 −0 kompute-shaders/op_getrows_q4_1.comp
+44 −0 kompute-shaders/op_getrows_q6_k.comp
+52 −0 kompute-shaders/op_mul.comp
+67 −0 kompute-shaders/op_mul_mat_f16.comp
+51 −0 kompute-shaders/op_mul_mat_mat_f32.comp
+33 −0 kompute-shaders/op_mul_mat_q4_0.comp
+35 −0 kompute-shaders/op_mul_mat_q4_1.comp
+94 −0 kompute-shaders/op_mul_mat_q6_k.comp
+73 −0 kompute-shaders/op_mul_mat_q8_0.comp
+48 −0 kompute-shaders/op_mul_mv_q_n.comp
+22 −0 kompute-shaders/op_mul_mv_q_n_pre.comp
+84 −0 kompute-shaders/op_norm.comp
+21 −0 kompute-shaders/op_relu.comp
+53 −0 kompute-shaders/op_rmsnorm.comp
+73 −0 kompute-shaders/op_rope_f16.comp
+73 −0 kompute-shaders/op_rope_f32.comp
+19 −0 kompute-shaders/op_scale.comp
+23 −0 kompute-shaders/op_scale_8.comp
+22 −0 kompute-shaders/op_silu.comp
+56 −0 kompute-shaders/op_softmax.comp
+67 −0 kompute-shaders/rope_common.comp
+308 −13 llama.cpp
+5 −1 llama.h
+1 −1 scripts/sync-ggml.last
+430 −9 tests/test-backend-ops.cpp
+4 −0 tests/test-c.c
+152 −13 tests/test-sampling.cpp

0 comments on commit dc3716d

Please sign in to comment.