[P] Liger Kernel: One line to make LLM Training +20% faster and -60% memory
https://github.com/linkedin/Liger-Kernel
Discussions: https://discu.eu/q/https://github.com/linkedin/Liger-Kernel
#compsci #machinelearning
Compsci Weekly /
npub12y…mwm90
2024-08-24 19:25:02
Author Public Key
npub12yxz5pw6kxawd4wshrvzkrdzyp772te52nzdeujcdjsjxngcqulsqmwm90Published at
2024-08-24 19:25:02Event JSON
{
"id": "ee504d6b7c3916d5a5d084173e38956be85c4503ec40845153b93526226360ef",
"pubkey": "510c2a05dab1bae6d5d0b8d82b0da2207de52f3454c4dcf2586ca1234d18073f",
"created_at": 1724527502,
"kind": 1,
"tags": [
[
"t",
"compsci"
],
[
"t",
"machinelearning"
],
[
"proxy",
"https://mastodon.social/users/compsci_discussions/statuses/113018634421629587",
"activitypub"
]
],
"content": "[P] Liger Kernel: One line to make LLM Training +20% faster and -60% memory\n\nhttps://github.com/linkedin/Liger-Kernel\n\nDiscussions: https://discu.eu/q/https://github.com/linkedin/Liger-Kernel\n\n#compsci #machinelearning",
"sig": "945d42fde318c4cf41d9ea55ef56a2711c9ff7a78a25bd7b74ff4f477239912e2547264bad277da5622e38730088405ea196880a59ff906063960d5a538a74d9"
}