ggml-org/llama.cpp

Pull Requests Commits

rename to init_from_text

ngxson committed 1 year ago

116b9a16

Add CLI arg to llama-run to adjust the number of threads used (#12370)

ericcurtin committed 1 year ago

Verified 9f2250ba

llama_batch_ext_ptr::from_text/embd

ngxson committed 1 year ago

eaffba0f

main : add -sysf / --system-prompt-file (#12249) (#12250)

CISC committed 1 year ago

Verified 774973b8

Load all MoE experts during warmup (#11571)

fairydreaming committed 1 year ago

Verified 8fcb5636

ngxson committed 1 year ago

8e7714fa

qwen2vl: use llama_batch_ext_set_pos

ngxson committed 1 year ago

a363251f

server: fix "--grammar-file" parameter (#12285)

dodekapod committed 1 year ago

Verified add2a3aa

fix llama_batch_ext_init_from_embd

ngxson committed 1 year ago

ba793696

fix missing n_past in various places

ngxson committed 1 year ago

07d84fa3

ngxson committed 1 year ago

32940369

fix llama-run n_past

ngxson committed 1 year ago

5e6a6d4e

graph : simplify attn input build for unified KV cache (#12381)

ggerganov committed 1 year ago

Verified c522ce41

hparams : add SWA rope parameters (#12374)

ggerganov committed 1 year ago

Verified 081bee8c

bring back mistakenly deleted llama_batch_init/free

ngxson committed 1 year ago

bfdddbc1

correct comment

ngxson committed 1 year ago

54566ad9

rm redundant llama_batch_ext_set_output_last

ngxson committed 1 year ago

04f86418

fix llama_batch_ext_init_from_text

ngxson committed 1 year ago

c3dd7900

ngxson committed 1 year ago

65f01845

fix common_batch missing seq_id

ngxson committed 1 year ago

9fb2d81e

apply to the rest

ngxson committed 1 year ago

47086fa8

llama : fix Gemma3 SWA KV cache shift (#12373)

ggerganov committed 1 year ago

Verified 84d54755

return output ID from llama_batch_ext_add/set

ngxson committed 1 year ago

4aabf4e8

fix merge errors

ngxson committed 1 year ago

86973cb1

Merge branch 'master' into xsn/private_batch_api

ngxson committed 1 year ago

17f954c8

arg : no n_predict = -2 for examples except for main and infill (#12364)

ngxson committed 1 year ago

Verified be7c3034

llama : refactor llama_context, llama_kv_cache, llm_build_context (#12181)

ggerganov committed 1 year ago

Verified e0dbec0b

server : fix crash when using verbose output with input tokens that are not in printable range (#12178) (#12338)

ishaangandhi committed 1 year ago

Verified 2048b591

Update build.yml for Windows Vulkan builder to use Vulkan 1.4.304 SDK for VK_NV_cooperative_matrix2 support (#12301)

oscarbg committed 1 year ago

Verified f08f4b31

llama.swiftui : fix xcframework dir in README [no ci] (#12353)

danbev committed 1 year ago

Verified 80a02aa8

Newer Older