| .. |
|
baby-llama
|
42c76d1358
Threadpool: take 2 (#8672)
|
hai 1 ano |
|
batched
|
6262d13e0b
common : reimplement logging (#9418)
|
hai 1 ano |
|
batched-bench
|
6262d13e0b
common : reimplement logging (#9418)
|
hai 1 ano |
|
batched.swift
|
0abc6a2c25
llama : llama_perf + option to disable timings during decode (#9355)
|
hai 1 ano |
|
convert-llama2c-to-ggml
|
6102037bbb
vocab : refactor tokenizer to reduce init overhead (#9449)
|
hai 1 ano |
|
cvector-generator
|
cad341d889
metal : reduce command encoding overhead (#9698)
|
hai 1 ano |
|
deprecation-warning
|
be6d7c0791
examples : remove `finetune` and `train-text-from-scratch` (#8669)
|
hai 1 ano |
|
embedding
|
f4d2b8846a
llama : add reranking support (#9510)
|
hai 1 ano |
|
eval-callback
|
6262d13e0b
common : reimplement logging (#9418)
|
hai 1 ano |
|
export-lora
|
6262d13e0b
common : reimplement logging (#9418)
|
hai 1 ano |
|
gbnf-validator
|
df270ef745
llama : refactor sampling v2 (#9294)
|
hai 1 ano |
|
gen-docs
|
afbbfaa537
server : add more env vars, improve gen-docs (#9635)
|
hai 1 ano |
|
gguf
|
07283b1a90
gguf : handle null name during init (#8587)
|
hai 1 ano |
|
gguf-hash
|
1666f92dcd
gguf-hash : update clib.json to point to original xxhash repo (#8491)
|
hai 1 ano |
|
gguf-split
|
76b37d1541
gguf-split : improve --split and --merge logic (#9619)
|
hai 1 ano |
|
gritlm
|
6262d13e0b
common : reimplement logging (#9418)
|
hai 1 ano |
|
imatrix
|
eca0fab44e
imatrix : disable prompt escape by default (#9543)
|
hai 1 ano |
|
infill
|
cea1486ecf
log : add CONT level for continuing previous log entry (#9610)
|
hai 1 ano |
|
jeopardy
|
1c641e6aac
`build`: rename main → llama-cli, server → llama-server, llava-cli → llama-llava-cli, etc... (#7809)
|
hai 1 ano |
|
llama-bench
|
7be099fa81
llama-bench: correct argument parsing error message (#9524)
|
hai 1 ano |
|
llama.android
|
5fb5e24811
llama : minor sampling refactor (2) (#9386)
|
hai 1 ano |
|
llama.swiftui
|
5fb5e24811
llama : minor sampling refactor (2) (#9386)
|
hai 1 ano |
|
llava
|
cad341d889
metal : reduce command encoding overhead (#9698)
|
hai 1 ano |
|
lookahead
|
6262d13e0b
common : reimplement logging (#9418)
|
hai 1 ano |
|
lookup
|
6262d13e0b
common : reimplement logging (#9418)
|
hai 1 ano |
|
main
|
cea1486ecf
log : add CONT level for continuing previous log entry (#9610)
|
hai 1 ano |
|
main-cmake-pkg
|
07a3fc0608
Removes multiple newlines at the end of files that is breaking the editorconfig step of CI. (#8258)
|
hai 1 ano |
|
parallel
|
6262d13e0b
common : reimplement logging (#9418)
|
hai 1 ano |
|
passkey
|
6262d13e0b
common : reimplement logging (#9418)
|
hai 1 ano |
|
perplexity
|
37f8c7b4c9
perplexity : remove extra new lines after chunks (#9596)
|
hai 1 ano |
|
quantize
|
63351143b2
quantize : improve type name parsing (#9570)
|
hai 1 ano |
|
quantize-stats
|
df270ef745
llama : refactor sampling v2 (#9294)
|
hai 1 ano |
|
retrieval
|
6262d13e0b
common : reimplement logging (#9418)
|
hai 1 ano |
|
rpc
|
841713e1e4
rpc : enable vulkan (#9714)
|
hai 1 ano |
|
save-load-state
|
bfe76d4a17
common : move arg parser code to `arg.cpp` (#9388)
|
hai 1 ano |
|
server
|
8c475b97b8
rerank : use [SEP] token instead of [BOS] (#9737)
|
hai 1 ano |
|
simple
|
6262d13e0b
common : reimplement logging (#9418)
|
hai 1 ano |
|
speculative
|
b0f27361f3
sampling : avoid expensive softmax during greedy sampling (#9605)
|
hai 1 ano |
|
sycl
|
faf67b3de4
[SYCL]set context default value to avoid memory issue, update guide (#9476)
|
hai 1 ano |
|
tokenize
|
6262d13e0b
common : reimplement logging (#9418)
|
hai 1 ano |
|
CMakeLists.txt
|
148844fe97
examples : remove benchmark (#9704)
|
hai 1 ano |
|
Miku.sh
|
1c641e6aac
`build`: rename main → llama-cli, server → llama-server, llava-cli → llama-llava-cli, etc... (#7809)
|
hai 1 ano |
|
base-translate.sh
|
1c641e6aac
`build`: rename main → llama-cli, server → llama-server, llava-cli → llama-llava-cli, etc... (#7809)
|
hai 1 ano |
|
chat-13B.bat
|
d9ad104440
Create chat-13B.bat (#592)
|
%!s(int64=2) %!d(string=hai) anos |
|
chat-13B.sh
|
1c641e6aac
`build`: rename main → llama-cli, server → llama-server, llava-cli → llama-llava-cli, etc... (#7809)
|
hai 1 ano |
|
chat-persistent.sh
|
1c641e6aac
`build`: rename main → llama-cli, server → llama-server, llava-cli → llama-llava-cli, etc... (#7809)
|
hai 1 ano |
|
chat-vicuna.sh
|
1c641e6aac
`build`: rename main → llama-cli, server → llama-server, llava-cli → llama-llava-cli, etc... (#7809)
|
hai 1 ano |
|
chat.sh
|
1c641e6aac
`build`: rename main → llama-cli, server → llama-server, llava-cli → llama-llava-cli, etc... (#7809)
|
hai 1 ano |
|
convert_legacy_llama.py
|
672a6f1018
convert-*.py: GGUF Naming Convention Refactor and Metadata Override Refactor (#7499)
|
hai 1 ano |
|
json_schema_pydantic_example.py
|
3fd62a6b1c
py : type-check all Python scripts with Pyright (#8341)
|
hai 1 ano |
|
json_schema_to_grammar.py
|
3fd62a6b1c
py : type-check all Python scripts with Pyright (#8341)
|
hai 1 ano |
|
llama.vim
|
125d03a503
llama.vim : added api key support (#5090)
|
%!s(int64=2) %!d(string=hai) anos |
|
llm.vim
|
ad9ddcff6e
llm.vim : stop generation at multiple linebreaks, bind to <F2> (#2879)
|
%!s(int64=2) %!d(string=hai) anos |
|
pydantic_models_to_grammar.py
|
090fca7a07
pydantic : replace uses of __annotations__ with get_type_hints (#8474)
|
hai 1 ano |
|
pydantic_models_to_grammar_examples.py
|
22f281aa16
examples : Rewrite pydantic_models_to_grammar_examples.py (#8493)
|
hai 1 ano |
|
reason-act.sh
|
1c641e6aac
`build`: rename main → llama-cli, server → llama-server, llava-cli → llama-llava-cli, etc... (#7809)
|
hai 1 ano |
|
regex_to_grammar.py
|
e235b267a2
py : switch to snake_case (#8305)
|
hai 1 ano |
|
server-llama2-13B.sh
|
1c641e6aac
`build`: rename main → llama-cli, server → llama-server, llava-cli → llama-llava-cli, etc... (#7809)
|
hai 1 ano |
|
server_embd.py
|
3fd62a6b1c
py : type-check all Python scripts with Pyright (#8341)
|
hai 1 ano |
|
ts-type-to-grammar.sh
|
ab9a3240a9
JSON schema conversion: ⚡️ faster repetitions, min/maxLength for strings, cap number length (#6555)
|
hai 1 ano |
Georgi Gerganov
8c475b97b8
rerank : use [SEP] token instead of [BOS] (#9737)