Skip to content

Activity

Showing most recent first

use scalar sums

0cc4mpushed 1 commit to 0cc4m/vulkan-coopmat-int8 • a943515…715ed28 • 
2 hours ago

apply scales inline

0cc4mcreated 0cc4m/vulkan-coopmat-int8 • a943515 • 
9 hours ago

ggml: add GATED_DELTA_NET op (#19504)

Pull request merge
am17anpushed 1 commit to master • 6fce5c6…c5a7788 • 
15 hours ago

fix missing declaration

0cc4mpushed 1 commit to 0cc4m/test-backend-ops-model-load • e2c9674…86c0299 • 
16 hours ago

opencl: add l2_norm (#20160)

Pull request merge
lhezpushed 1 commit to master • c024d85…6fce5c6 • 
21 hours ago

Autoparser: True streaming (#20177)

Pull request merge
pwilkinpushed 1 commit to master • 2f2923f…c024d85 • 
22 hours ago

Autoparser: add optional argument reshuffle capability (#20171)

Pull request merge
pwilkinpushed 1 commit to master • 649f064…2f2923f • 
yesterday

quants : Add memsets and other fixes for IQ quants (#19861)

Pull request merge
ggerganovpushed 1 commit to master • 7463687…649f064 • 
yesterday

Add @pwilkin to CODEOWNERS for autoparser code (#20174)

Pull request merge
pwilkinpushed 1 commit to master • 566059a…7463687 • 
yesterday

Add @pwilkin to CODEOWNERS for autoparser code

pwilkincreated autoparser-codeowner • 7296936 • 
yesterday

Autoparser - complete refactoring of parser architecture (#18675)

Pull request merge
pwilkinpushed 1 commit to master • 34df42f…566059a • 
yesterday

hexagon: add f32 ssm_conv op (#20122)

Pull request merge
max-krasnyanskypushed 1 commit to master • e68f2fb…34df42f • 
yesterday

server : preserve anthropic thinking blocks in conversion (#20120)

Pull request merge
ngxsonpushed 1 commit to master • ba2fd11…e68f2fb • 
yesterday

cpu: skip redudant ROPE cache updates (#20149)

Pull request merge
max-krasnyanskypushed 1 commit to master • d48e876…ba2fd11 • 
yesterday

ggml-cuda: add mem check for fusion (#19916)

Pull request merge
am17anpushed 1 commit to master • ba2ff79…d48e876 • 
yesterday

ggml: update comments for backends which have no memory to report (#2…

Pull request merge
taronaeopushed 1 commit to master • c6980ff…ba2ff79 • 
yesterday

ggml-cpu: Fix gcc 15 ICE on ppc64le (#20083) (#20130)

Pull request merge
taronaeopushed 1 commit to master • 1e38a7a…c6980ff • 
yesterday

CUDA: use shared mem for ssm_conv (#20128)

Pull request merge
am17anpushed 1 commit to master • 388baab…1e38a7a • 
yesterday

test

ggerganovcreated pr/19802-test • 121fe62 • 
yesterday

move llama_graph_reserve function to new llama-ext header, move expor…

0cc4mpushed 1 commit to 0cc4m/test-backend-ops-model-load • 46e824a…e2c9674 • 
yesterday

context: ignore zero scale LoRAs when checking sameness (#20166)

Pull request merge
ggerganovpushed 1 commit to master • f5ddcd1…388baab • 
yesterday

Checkpoint every n tokens: squash (#20087)

Pull request merge
pwilkinpushed 1 commit to master • f6235a4…f5ddcd1 • 
yesterday

webui: Agentic Loop + MCP Client with support for Tools, Resources an…

Pull request merge
allozaurpushed 1 commit to master • 2850bc6…f6235a4 • 
yesterday

ggml-cpu: fix data race for debug asserts (#20148)

Pull request merge
JohannesGaesslerpushed 1 commit to master • 17a4258…2850bc6 • 
yesterday

Deleted branch

ggerganovdeleted gg/kv-cache-fix-mtmd-chkpt • 
yesterday

kv-cache : fix M-RoPE checkpoints (#20132)

Pull request merge
ggerganovpushed 1 commit to master • f7db3f3…17a4258 • 
yesterday

cli : Don't clear system prompt when using '/clear' (#20067)

Pull request merge
danbevpushed 1 commit to master • 6c97bff…f7db3f3 • 
yesterday

opencl: add neg, exp and diag (#20127)

Pull request merge
lhezpushed 1 commit to master • 2b10b62…6c97bff • 
yesterday

hexagon: add fp16 support for binary ops: add,sub,mul,div (#20139)

Pull request merge
max-krasnyanskypushed 1 commit to master • a0ed91a…2b10b62 • 
yesterday

models : kda chunk size = 16 (#19827)

Pull request merge
ggerganovpushed 1 commit to master • 2cd20b7…a0ed91a • 
2 days ago