ggerganov/llama.cpp

Pull Requests Commits

convert : update help string

cebtenzzre committed 2 years ago

f8ab5391

docs : s/LLaMa/LLaMA/

cebtenzzre committed 2 years ago

767aef90

convert : automatically fall back to HfVocab if needed

cebtenzzre committed 2 years ago

17d22efa

gemma : fix bfloat16 -> float16 conversion issue (#5810)

kunal-vaishnavi committed 2 years ago

Verified e7433867

common : fix flag `--logits-all` to `--all-logits` (#5805)

ensan-hcl committed 2 years ago

Verified f49a5356

llama : cleanup unused mmq flags (#5772)

phymbert committed 2 years ago

Verified 3ab8b3a9

unicode : switch to multimap based nfd_map (#5799)

iamlemec committed 2 years ago

Verified 9600d59e

server: allow to override threads server pool with --threads-http (#5794)

phymbert committed 2 years ago

Verified 5cb02b4a

ci : add Ubuntu 22 Vulkan CI run (#5789)

netrunnereve committed 2 years ago

Verified 6ea0f010

server : fix newlines in help (#5785)

ggerganov committed 2 years ago

Verified f105471e

[SYCL] Use batched mul_mat pathway (#5591)

AidanBeltonS committed 2 years ago

Verified 38d15216

Server: normalize naming (#5779)

ngxson committed 2 years ago

Verified 052051d8

llama : constified `llama_set_state_data`'s `src` (#5774)

MarcusDunn committed 2 years ago

Verified d5ab2975

ci : reduce 3b ppl chunks to 1 to avoid timeout (#5771)

ggerganov committed 2 years ago

Verified 87c91c07

make portability_enumeration_ext apple only (#5757)

netrunnereve committed 2 years ago

Verified 317709b2

llama : remove deprecated API (#5770)

ggerganov committed 2 years ago

Verified 08c5ee87

awq-py : remove (#5768)

ggerganov committed 2 years ago

Verified 78aacf36

ggerganov committed 2 years ago

Verified 8c0e8f4e

add google magika inference example (ggml/748)

slaren committed 2 years ago

Verified 2774b0c9

Introduce backend GUIDs (ggml/743)

uextm committed 2 years ago

Verified 5f706718

server : hit Ctrl+C twice to exit (#5734)

ngxson committed 2 years ago

Verified a693bea1

llama : fix non-quantization of expert gating tensors (#5754)

compilade committed 2 years ago

Verified adcb12a9

llama : improve BERT tokenization (#5740)

iamlemec committed 2 years ago

Verified 177628bf

readme : add link to LLaVA 1.6 models (#5758)

danbev committed 2 years ago

Verified 6c441686

server : add "/chat/completions" alias for "/v1/...` (#5722)

jorgealias committed 2 years ago

Verified efc72253

ggml : make i-quants work with super-blocks of 64 (CPU,Metal) (#5760)

ikawrakow committed 2 years ago

Verified 7c4263d4

Attempt to fix android build (#5752)

ikawrakow committed 2 years ago

Verified cb49e0f8

IQ4_XS: a 4.25 bpw quantization (#5747)

ikawrakow committed 2 years ago

Verified 0becb22a

cuda : replace remaining shfl_xor with calls to warp_reduce functions (#5744)

Engininja2 committed 2 years ago

Verified c24a2a6e

ggml-quants : fix avx2 iq1_s vec_dot when compiled with gcc (#5742)

Engininja2 committed 2 years ago

Verified 1f30b7a9

Older