ggml-org/llama.cpp

Pull Requests Commits
quantize : use `map` to assign quantization type from `string` (#1191)
prusnak committed 3 years ago
Verified 859fee6d
Update SHA256SUMS after quantization change (#1181)
sw committed 3 years ago
Verified 4afcc378
py : cast lora_alpha to int in convert-lora-to-ggml (#1170)
ostix360 committed 3 years ago
Verified 667c5013
nix: use convert.py instead of legacy wrapper convert-pth-to-ggml.py (#981)
prusnak committed 3 years ago
Verified bb98e77b
ggml : add Q8_0 quantization format (rename the old one to Q8_1) (ARM NEON) (#1179)
ggerganov committed 3 years ago
Verified 7a32fcb3
ggml : use full range for Q4_0 and Q4_2 quantization (#729)
unbounded committed 3 years ago
Verified dd0eabc0
ggml : fix bug in ggml_compute_forward_sum_f32 (#1162)
xaedes committed 3 years ago
Verified 54bb60e2
ggml : export symbols (#1155)
ggerganov committed 3 years ago
Verified 8a0f8673
examples : add save_load_state example (#1150)
xaedes committed 3 years ago
Verified 0c569234
llama : increase scratch buffer size for 65B (ref #1152)
ggerganov committed 3 years ago
Verified 957c8ae2
examples/main README improvements and some light refactoring (#1131)
mgroeber9110 committed 3 years ago
Verified 9b0a4d42
Fix build for gcc 8 and test in CI (#1154)
sw committed 3 years ago
Verified 2ec83428
Fix cuda compilation (#1128)
slaren committed 3 years ago
Verified e4cf982e
llama : refactor get / set state + remove redundant kv cache API (#1143)
ggerganov committed 3 years ago
Verified c4fe84fb
Fix LoRA acronym (#1145)
slaren committed 3 years ago
Verified 1d78fecd
scripts : add helper scripts to synch ggml repo
ggerganov committed 3 years ago
Verified 284685f1
Added README.md for main with examples and explanations (#1139)
DannyDaemonic committed 3 years ago
Verified edce63ba
ggml : do not print perf ops that have not been used at all
ggerganov committed 3 years ago
Verified ec9cdb67
ggml : better PERF prints + support "LLAMA_PERF=1 make"
ggerganov committed 3 years ago
Verified e4422e29
Improve AVX2 for vec_dot_q4_3_q8_0 (#1138)
sw committed 3 years ago
Verified 53c84343
readme : update gpt4all instructions (#980)
prusnak committed 3 years ago
Verified c6524f46
A better `packNibbles` and `mul_sum_i8_pairs_float` implementation using AVX512 (#1119)
MeouSker77 committed 3 years ago
Verified c9e2c26f
ggml : fix Q4_3 cuBLAS
ggerganov committed 3 years ago
Verified 0e018fe0
ci : trigger CI for drafts, but not most PR actions (#1125)
sw committed 3 years ago
Verified 857308d1
Fix CI: ARM NEON, quantization unit tests, editorconfig (#1122)
sw committed 3 years ago
Verified c50b6288
ggml : unit test for quantization functions (#953)
unbounded committed 3 years ago
Verified 5f939498
llama : print timings on ctrl+c exit (#1021)
wbpxre150 committed 3 years ago
Verified 36b4f7e0
llama : have n_batch default to 512 (#1091)
eiery committed 3 years ago
Verified 10f19c11
cmake : fix build under Windows when enable BUILD_SHARED_LIBS (#1100)
howard0su committed 3 years ago
Verified 7e312f16
ggml : fix AVX build + update to new Q8_0 format
ggerganov committed 3 years ago
872c365a
Terms Of Use Privacy Policy Refund Policy Impressum