ggerganov/llama.cpp

Pull Requests Commits
run : avoid double tokenization (#14327)
retr0reg committed 1 year ago
Verified 66aba7ac
examples : fix is_first logic for tokenization (#14329)
ggerganov committed 1 year ago
Verified f1f5e82d
HIP: enable vec fattn on RDNA4 (#14323)
IMbackK committed 1 year ago
Verified af3373f1
mtmd : fix Pixtral OOM with large images by capping image_size to 1024 (#14326)
yuiseki committed 1 year ago
Verified 5d5c066d
common : use std::string_view now that we target c++17 (#14319)
CISC committed 1 year ago
Verified 40bfa04c
CUDA: add mean operation (#14313)
am17an committed 1 year ago
Verified aa064b2e
gguf-py : fix Qwen3-Embedding eos token (#14314)
CISC committed 1 year ago
Verified aa0ef5c5
Add support for VK_EXT_debug_utils to add labels to Vulkan objects. (#13792)
mtavenrath committed 1 year ago
Verified bb16041c
gguf-py : fix TemplateProcessing pair when bos/eos is missing (#14312)
CISC committed 1 year ago
Verified 58cba76a
metal : fix thread-safety (#14300)
ggerganov committed 1 year ago
Verified 67ae5312
memory : rename interface to llama_memory_context_i (#14296)
ggerganov committed 1 year ago
Verified 692e3cdd
convert : fix Llama 4 conversion (#14311)
danielhanchen committed 1 year ago
Verified b23fa0b3
sync : ggml
ggerganov committed 1 year ago
06cbedfc
Add `ggml_roll` (ggml/1274)
Acly committed 1 year ago
b7147673
docs : fix the link to llama.h (#14293)
david20571015 committed 1 year ago
Verified d860dd99
CUDA: add conv_2d_transpose (#14287)
am17an committed 1 year ago
Verified c959f462
lint : remove trailing whitepace (#14304)
CISC committed 1 year ago
Verified 22015b20
vocab : prevent tokenizer overflow (#14301)
retr0reg committed 1 year ago
Verified dd6e6d0b
sycl: add usage of enqueue_functions extension (#14244)
s-Nick committed 1 year ago
Verified 8308f98c
Implement GGML_CPU_ALL_VARIANTS for PowerPC (#14286)
ckastner committed 1 year ago
Verified 6369be07
llama : improve sep token handling (#14272)
CISC committed 1 year ago
Verified 88fc854b
cuda : synchronize graph capture and cublas handle destruction (#14288)
slaren committed 1 year ago
Verified e28c1b93
ggml : fix repack work size for mul_mat_id (#14292)
ggerganov committed 1 year ago
Verified d27b3ca1
ggml: Update KleidiAI to v1.9.0 (#14277)
chaxu01 committed 1 year ago
Verified 9230dbe2
model : more uniform output id handling (#14275)
ggerganov committed 1 year ago
Verified 812939a9
ubatch : new splitting logic (#14217)
ggerganov committed 1 year ago
Verified 4c9fdfbe
CUDA: add conv_2d_dw (#14265)
am17an committed 1 year ago
Verified 9eaa51e7
ggml-cpu : remove unnecesary arm feature detection (#14281)
slaren committed 1 year ago
Verified 8f71d0f3
gguf-py : make sentencepiece optional (#14200)
Ahajha committed 1 year ago
Verified 381174bb
server : add server parameters for draft model cache type (#13782)
aa956 committed 1 year ago
Verified d67341dc
Terms Of Use Privacy Policy Refund Policy Impressum