ochafik/llama.cpp

Pull Requests Commits

llama.cpp : fix wrong vsnprintf call in MS compiler (#2856)

grahameth committed 2 years ago

Verified be475f60

ggml : tiny ggml_vec_dot_q4_K_q8_K AVX2 improvement (#2819)

hydroo committed 2 years ago

Verified 3af6b863

ggml : sync (mem align to header + conv_transpose_2d fixes + ggml_alloc) (#2852)

ggerganov committed 2 years ago

Verified 35feac65

CUDA: fix RoPE asserts, block sizes (#2833)

JohannesGaessler committed 2 years ago

Verified 92b1bbd2

llama.h : add missing struct keyword for C compat in callback type (#2847)

igarnier committed 2 years ago

Verified dd0dc366

metal : fix memory leak (#2762)

ggerganov committed 2 years ago

Verified f55538c3

quantize : make output filename optional again (#2823)

cebtenzzre committed 2 years ago

Verified ebcee207

devops : added systemd units and set versioning to use date. (#2835)

jboero committed 2 years ago

Verified 3e8ff47a

gguf : fix strings to not be null-terminated (#2839)

ggerganov committed 2 years ago

Verified 103cfafc

llama : fix MPI threads (close #2827)

ggerganov committed 2 years ago

Verified c10704d0

examples : update llama2.c converter to read vocab and write models in GGUF format (#2751)

ochafik committed 2 years ago

Verified 230d46c7

llama : speedup tokenization (#2831)

ikawrakow committed 2 years ago

Verified 463173a6

falcon : fix CUDA inference by making K and Q contiguous (#2830)

ggerganov committed 2 years ago

Verified eaa13a48

readme : fix headings

ggerganov committed 2 years ago

Verified da7455d0

scripts : helper convert script

ggerganov committed 2 years ago

Verified 25423e91

k_quants tuning for Falcon-7b (#2816)

ikawrakow committed 2 years ago

Verified a6d1189f

readme : update hot topics

ggerganov committed 2 years ago

Verified c48c5bb0

gguf : add 64-bit support (GGUF v2) (#2821)

ggerganov committed 2 years ago

Verified d0cee0d3

llama : more tokenizer fixes (#2810)

ggerganov committed 2 years ago

Verified edd4c148

ggml : detect SSSE3 (#2825)

przemoc committed 2 years ago

Verified 1591e2e5

ci : add LoRA test to CI (#2650)

slaren committed 2 years ago

Verified 789c8c94

server : add `/detokenize` endpoint (#2802)

BruceMacD committed 2 years ago

Verified c1ac54b7

convert.py : advanced option (#2753)

KerfuffleV2 committed 2 years ago

Verified 730d9c68

llama : use Unicode Escape Sequence to replace encoded characters (#2814)

drasticactions committed 2 years ago

Verified c7d92e6d

flake.nix : add rocm support and cleanup (#2808)

Tungsten842 committed 2 years ago

Verified 61d1a289

llama : move #includes out of _GNU_SOURCE conditional (#2817)

cebtenzzre committed 2 years ago

Verified 741ca7dd

main : fix bug (penalize_nl=false doesn't work) + suppress warning on mingw (#1528)

tom7 committed 2 years ago

Verified 72f895c9

llama : use std::abs in llama_sample_tail_free (#2800)

cebtenzzre committed 2 years ago

Verified 50526f37

k-quants : remove unnecessary tensor shape restrictions (#2811)

ggerganov committed 2 years ago

Verified 04f4b1eb

Better perplexity for 2- and 3-bit quantization for LLaMA-v2-70B (#2807)

ikawrakow committed 2 years ago

Verified 75923754

Newer Older