ochafik/llama.cpp

Pull Requests Commits
llava : add struct for FFI bindgen (#12079)
tinglou committed 1 year ago
Verified a800ae46
Refactor gguf scripts to improve metadata handling (#11909)
CISC committed 1 year ago
Verified 69050a11
gguf-py: enable reading non-native endian files (#12081)
AlekseiNikiforovIBM committed 1 year ago
Verified 3567ee3a
readme : update infra list (#9096)
kerthcet committed 1 year ago
Verified 53e4db10
docs: add docs/function-calling.md to lighten server/README.md's plight (#12069)
ochafik committed 1 year ago
Verified d7cfe1ff
vulkan: fix assertion when qy_needs_dequant (#12068)
jeffbolznv committed 1 year ago
Verified a82c9e7c
server: handle echo=false on /v1/completions (#12060)
rhjdvsgsgks committed 1 year ago
Verified 401af80b
add OP sigmoid (#12056)
foldl committed 1 year ago
Verified c132239b
ggml-cpu: Fix build with sve (#12059)
MollySophia committed 1 year ago
Verified 393fca62
vulkan: implement more backpropagation operators (#11914)
remyoudompheng committed 1 year ago
Verified 61d4f39d
server: support add_generation_prompt query param (#12062)
ochafik committed 1 year ago
Verified 0b527456
Add Doc for Converting Granite Vision -> GGUF (#12006)
alex-jw-brooks committed 1 year ago
Verified 4d1051a4
llama : expose llama_model_n_head_kv in the API (#11997)
vlovich committed 1 year ago
Verified 3e9a2860
metal : copy kernels for quant to F32/F16 conversions (#12017)
gcp committed 1 year ago
Verified 58d07a80
opencl: fix for small models (#11950)
lhez committed 1 year ago
Verified 34a846b5
llava : Add Granite Vision Support (#11794)
alex-jw-brooks committed 1 year ago
Verified 7a2c913e
[SYCL] Optimize mul_mat for Q4_0 on Intel GPU (#12035)
NeoZhangJianyu committed 1 year ago
Verified 08d59862
gguf_convert_endian.py: implement byteswapping for q4_k and q6_k (#11349)
AlekseiNikiforovIBM committed 1 year ago
Verified 651adf4b
SYCL: Fix GGML_SYCL_DEBUG macro (#11995)
qnixsynapse committed 1 year ago
Verified 8303e8b0
run: allow to customize prompt by env var LLAMA_PROMPT_PREFIX (#12041)
benoitf committed 1 year ago
Verified 7ad0779f
Some llama-run cleanups (#11973)
ericcurtin committed 1 year ago
Verified f777a73e
ggml-cpu: Support s390x SIMD Instruction Set (#12019)
taronaeo committed 1 year ago
Verified af7747c9
CUDA: app option to compile without FlashAttention (#12025)
JohannesGaessler committed 1 year ago
Verified a28e0d5e
llava: build clip image from pixels (#11999)
tinglou committed 1 year ago
Verified 36c258ee
ci : fix arm upload artifacts (#12024)
ggerganov committed 1 year ago
Verified f3e64859
CUDA: optimize FA for GQA + large batches (#12014)
JohannesGaessler committed 1 year ago
Verified 5fa07c2f
ci : Build on Github-hosted arm64 runners (#12009)
Rohanjames1997 committed 1 year ago
Verified 335eb04a
server : disable Nagle's algorithm (#12020)
ggerganov committed 1 year ago
Verified cf756d6e
cuda: Add Q5_1, Q5_0, Q4_1 and Q4_0 to F32 conversion support. (#12000)
gcp committed 1 year ago
Verified d7090842
llama.swiftui : add "Done" dismiss button to help view (#11998)
danbev committed 1 year ago
Verified de8b5a36
Terms Of Use Privacy Policy Refund Policy Impressum