 Georgi Gerganov
|
d32e03f449
server : add SWA checkpoints (#15293)
|
vor 5 Monaten |
 l3utterfly
|
7233358d29
memory : handle saving/loading null layers in recurrent memory (#14675)
|
vor 6 Monaten |
|
Georgi Gerganov
|
01612b7409
llama : reuse compute graphs (#14482)
|
vor 6 Monaten |
 compilade
|
4a5686da22
llama : support Jamba hybrid Transformer-Mamba models (#7531)
|
vor 6 Monaten |
|
compilade
|
bb4f7a9e4e
memory : fix broken batch splits for recurrent cache (#14575)
|
vor 6 Monaten |
|
Georgi Gerganov
|
67d1ef23c6
batch : add optional for sequential equal split (#14511)
|
vor 7 Monaten |
|
Georgi Gerganov
|
c79184d2d1
batch : add n_used count (#14512)
|
vor 7 Monaten |
|
Georgi Gerganov
|
745f11fed0
memory : correctly handle failure in apply() (#14438)
|
vor 7 Monaten |
|
Georgi Gerganov
|
43678060c1
recurrent : call balloc split_reset() in init_batch() (#14414)
|
vor 7 Monaten |
|
Georgi Gerganov
|
692e3cdd0a
memory : rename interface to llama_memory_context_i (#14296)
|
vor 7 Monaten |
|
Georgi Gerganov
|
4c9fdfbe15
ubatch : new splitting logic (#14217)
|
vor 7 Monaten |
 Gabe Goodhart
|
edc4a29eff
memory : Hybrid recurrent cache (#13979)
|
vor 7 Monaten |
