I have not looked into it, but two notes - 64GB ram is not much for local AI and AMD is not very common for local inference, most bleeding edge is Nvidia and Apple silicon. You can definitely run models (that fit in memory there), but you don't usually get hw optimized inference.
So prediction - it will run, but not the best open models and it will be slow.
Juraj / Juraj🏴💛🌘
npub1m2…lr8p9
2026-10-01 05:46:59 UTC
in reply to nevent1q…rj0v
Author Public Key
npub1m2mvvpjugwdehtaskrcl7ksvdqnnhnjur9v6g9v266nss504q7mqvlr8p9Seen on
wss://relay.damus.ioPublished at
2026-10-01 05:46:59 UTCEvent JSON
{
"id": "cc259035fa153126f9dd9c4ebc42bc2e2ff9041193cc2310b559052b7e0cd31d",
"pubkey": "dab6c6065c439b9bafb0b0f1ff5a0c68273bce5c1959a4158ad6a70851f507b6",
"created_at": 1790833619,
"kind": 1,
"tags": [
[
"p",
"dab6c6065c439b9bafb0b0f1ff5a0c68273bce5c1959a4158ad6a70851f507b6",
"wss://nos.lol/"
],
[
"p",
"c6ddba5ee3dc641aa5b1b88bfed85ece646153bd9d598866c4122defb9c10a73"
],
[
"p",
"dab6c6065c439b9bafb0b0f1ff5a0c68273bce5c1959a4158ad6a70851f507b6",
"wss://relay.primal.net/"
],
[
"e",
"033e6bd38cf91bbb1482b8e5cec79daae71457205d8898fbb14872a5b64f2da8",
"",
"reply",
"c6ddba5ee3dc641aa5b1b88bfed85ece646153bd9d598866c4122defb9c10a73"
],
[
"e",
"2c3afdb45557314bb13f512ec845725a58ea4194e5afdb88b64f097d2986e116",
"wss://relay.primal.net/",
"root",
"dab6c6065c439b9bafb0b0f1ff5a0c68273bce5c1959a4158ad6a70851f507b6"
],
[
"client",
"Primal Android"
]
],
"content": "I have not looked into it, but two notes - 64GB ram is not much for local AI and AMD is not very common for local inference, most bleeding edge is Nvidia and Apple silicon. You can definitely run models (that fit in memory there), but you don't usually get hw optimized inference. \n\nSo prediction - it will run, but not the best open models and it will be slow. \n\n",
"sig": "6c41e4b9fbcdb7293d5e2ceecd01b6458cdf46197eb3e1456104ede9afa587b93cd3125f19fa4a85fa46941f07e159d0c6423ad5b53bcb17baaf838810e6e2b6"
}