|
|
|
|
|
|
by NitpickLawyer
511 days ago
|
|
|
You can't run the big R1 in any useful quant, but can use the distilled models with your setup. They've released (MIT) versions of qwen (1.5,7,14 and 32b) and llama3 (8 and 70b) distilled on 800k samples from R1. They are pretty impressive, so you can try them out. |
|
|