Drop-Upcycling
updated
9B • Updated • 30
9B • Updated • 33
19B • Updated • 29
9B • Updated • 34
9B • Updated • 31
0.4B • Updated • 33
0.4B • Updated • 29
0.4B • Updated • 39
0.4B • Updated • 40
9B • Updated • 52
19B • Updated • 29
0.4B • Updated • 33
9B • Updated • 29
0.4B • Updated • 29
• 1
2B • Updated • 29
0.2B • Updated • 39
4B • Updated • 30
14B • Updated • 34
llm-jp/Dense-btx-code-expert-152M
0.2B • Updated • 35
• 1
llm-jp/Dense-btx-english-expert-1.5B
2B • Updated • 32
llm-jp/Dense-btx-code-expert-1.5B
2B • Updated • 36
• 1
llm-jp/Dense-btx-japanese-expert-1.5B
2B • Updated • 26
• 1
llm-jp/Dense-btx-english-expert-152M
0.2B • Updated • 26
llm-jp/Dense-btx-japanese-expert-152M
0.2B • Updated • 28
Drop-Upcycling: Training Sparse Mixture of Experts with Partial
Re-initialization
Paper
• 2502.19261
• Published • 6
Text Generation
• 73B • Updated • 37
llm-jp/llm-jp-3-8x13b-instruct3
Text Generation
• 73B • Updated • 140
• 8
llm-jp/llm-jp-3-8x1.8b-instruct3
Text Generation
• 9B • Updated • 46
• 4
Text Generation
• 9B • Updated • 97
llm-jp/llm-jp-3-8x13b-instruct2
Text Generation
• 73B • Updated • 129
llm-jp/llm-jp-3-8x1.8b-instruct2
Text Generation
• 9B • Updated • 40
llm-jp/llm-jp-3.1-8x13b-instruct4
Text Generation
• 73B • Updated • 135
• 4
Text Generation
• 73B • Updated • 36