ggml-org/llama.cpp

Pull Requests Commits

common : remove incorrect --model-draft default

cebtenzzre committed 2 years ago

7c87353e

py : open merges file as 'utf-8' (#4566)

howlger committed 2 years ago

Verified 880e3522

cuda : better error message for ggml_get_rows (#4561)

bobqianic committed 2 years ago

Verified 66f35a2f

cuda : replace asserts in wrong architecture checks with __trap (#4556)

slaren committed 2 years ago

Verified 13988239

llama : disable per-tensor info prints on model load (#4562)

JohannesGaessler committed 2 years ago

Verified d3223afd

Fix access violation in ggml_cuda_free_data if tensor->extra is NULL (#4554)

LoganDark committed 2 years ago

Verified 1d7a1912

CUDA: Faster Mixtral prompt processing (#4538)

JohannesGaessler committed 2 years ago

Verified 799fc226

ggml : fixed check for _MSC_VER (#4535)

es0m committed 2 years ago

Verified 328b83de

ggml-cuda: Fix HIP build (#4528)

arlo-phoenix committed 2 years ago

Verified a7aee47b

llama.swiftui : add tinyllama 1.1B F16

ggerganov committed 2 years ago

Verified 0e18b2e7

llama.swiftui : add more models

ggerganov committed 2 years ago

Verified 6ff39b12

llama : add phi-2 + fix NeoX rope + ggml_mul_mat_set_prec (#4490)

ebeyabraham committed 2 years ago

Verified b9e74f9b

llama : fix try_override for bool_value which always return true (#4519)

hankcs committed 2 years ago

Verified 3c04bf6d

decode : fix logits_valid for legacy API (#4516)

cebtenzzre committed 2 years ago

Verified 2994f0c5

readme : update hot topics

ggerganov committed 2 years ago

Verified b1306c43

llama.swiftui : add bench functionality (#4483)

ggerganov committed 2 years ago

Verified 800a489e

gguf-py : fail fast on nonsensical special token IDs (#4489)

cebtenzzre committed 2 years ago

Verified f7f468a9

build : Check the ROCm installation location (#4485)

Su3h7aM committed 2 years ago

Verified 919c4066

finetune : keep allocs alive until all allocations are done (#4486)

slaren committed 2 years ago

Verified 45668633

server : disable llm logs if SERVER_VERBOSE is off (#3792)

olexiyb committed 2 years ago

Verified 0ffc92d2

server : fix grammar being ignored (#4494)

AdithyanI committed 2 years ago

Verified 8edd2b40

server : fix possible ambiguity in content type charset (#4501)

z80maniac committed 2 years ago

Verified eb16dae7

server : allow requests larger than 8K (#4500)

mzcu committed 2 years ago

Verified 62bd52b7

Link to cublas dynamically on Windows even with LLAMA_STATIC (#4506)

bullno1 committed 2 years ago

Verified 5daa5f54

lora : add support for non-llama models (#3333)

slaren committed 2 years ago

Verified c6c4fc08

llama : sanity checks for access to logits (#4274)

cebtenzzre committed 2 years ago

Verified 8a5be3bd

server : add optional API Key Authentication example (#4441)

ShadovvBeast committed 2 years ago

Verified 88ae8952

ggml : group mul_mat_id rows by matrix (cpu only) (#4480)

slaren committed 2 years ago

Verified ee4725a6

ggml : use ggml_row_size where possible (#4472)

slaren committed 2 years ago

Verified 6744dbe9

ggml : remove n_dims from ggml_tensor (#4469)

slaren committed 2 years ago

Verified cafcd4f8

Older