 Gabe Goodhart
|
edc4a29eff
memory : Hybrid recurrent cache (#13979)
|
7 months ago |
 Georgi Gerganov
|
60c666347b
batch : rework llama_batch_allocr (#14153)
|
7 months ago |
 compilade
|
dad5c44398
kv-cache : avoid modifying recurrent cells when setting inputs (#13834)
|
7 months ago |
 Đinh Trọng Huy
|
91a8ee6a6f
add geglu activation function (#14074)
|
7 months ago |
|
Georgi Gerganov
|
7f37b6cf1e
memory : migrate from llama_kv_cache to more generic llama_memory (#14006)
|
7 months ago |
|
Georgi Gerganov
|
12d0188c0d
kv-cache : refactor + add llama_memory_state_i (#13746)
|
7 months ago |
|
Georgi Gerganov
|
e298d2fbd0
kv-cache : add SWA support (#13194)
|
8 months ago |
 Johannes Gäßler
|
10d2af0eaa
llama/ggml: add LLM training support (#10544)
|
8 months ago |
|
Georgi Gerganov
|
c642bc014c
kv-cache : separate recurrent vs non-recurrent impl (#12799)
|
8 months ago |
 Xuan-Son Nguyen
|
d2b2031e5f
llama : (mrope) allow using normal 1D position for text token (#13138)
|
9 months ago |
 Juk Armstrong
|
daa422881a
llama : DeepSeek V2/V3 MLA implementation (#12801)
|
9 months ago |
|
Xuan-Son Nguyen
|
1466621e73
llama : Support llama 4 text-only (#12791)
|
9 months ago |
|
Georgi Gerganov
|
75422e8bc4
graph : normalize Q, K, V shapes + sync cross attention (#12449)
|
10 months ago |
|
Georgi Gerganov
|
c522ce4143
graph : simplify attn input build for unified KV cache (#12381)
|
10 months ago |
|
Georgi Gerganov
|
e0dbec0bc6
llama : refactor llama_context, llama_kv_cache, llm_build_context (#12181)
|
10 months ago |
