From-scratch 1.6B LLM trained solo on a TPU grant. Matches Llama 3.2 1B with 18x less data. Apache 2.0. Report + story: 0arch.io/core
Joseph Rodriguez
0arch-io
AI & ML interests
None yet
Recent Activity
updated a collection 6 days ago
Kisoku 1.6B updated a collection 6 days ago
Kisoku 1.6B updated a collection 6 days ago
Kisoku 1.6BOrganizations
None yet
models 10
0arch-io/kisoku-1.6b-gguf
Text Generation • 2B • Updated • 184
0arch-io/kisoku-1.6b-chat
Text Generation • 2B • Updated • 22
0arch-io/kisoku-1.6b
Text Generation • 2B • Updated • 24
0arch-io/kisoku-1.6b-preview
Text Generation • 2B • Updated • 222
0arch-io/kisoku-1.6b-preview-GGUF
Text Generation • 2B • Updated • 179
0arch-io/kisoku-3b
Text Generation • 4B • Updated • 13 • 1
0arch-io/kisoku-3b-sft
4B • Updated • 10 • 1
0arch-io/kisoku-3b-base
4B • Updated • 21 • 2
0arch-io/dolphin-v2-8b-abliterated
Text Generation • 8B • Updated • 1.01k • • 4
0arch-io/kisoku-3.2b-base
Text Generation • Updated • 2