FIM-Midtraining - a TIGER-Lab Collection

TIGER-Lab 's Collections

FIM-Midtraining

RationalRewards

General-Reasoner

VisualWebInstruct

CritiqueFineTuning

FIM-Midtraining

updated about 5 hours ago

Function-aware FIM mid-training for coding agents: 6 checkpoints (7B/8B/14B) + the 400K/2.6B-token corpus. github.com/TIGER-AI-Lab/FIM-Midtraining

TIGER-Lab/FIM-14B

Text Generation • 841k • Updated about 4 hours ago

Note FIM-Midtrain + R2E-Gym 14B
TIGER-Lab/FIM-7B

Text Generation • 333k • Updated about 4 hours ago

Note FIM-Midtrain + R2E-Gym 7B
TIGER-Lab/FIM-8B

Text Generation • 8B • Updated about 4 hours ago

Note FIM-Midtrain + SWE-LEGO 8B
TIGER-Lab/FIM-Mid-7B

Text Generation • 333k • Updated about 4 hours ago

Note Only FIM-Midtrain 7B
TIGER-Lab/FIM-Mid-14B

Text Generation • 15B • Updated about 4 hours ago

Note Only FIM-Midtrain 14B
TIGER-Lab/FIM-Mid-8B

Text Generation • 308k • Updated about 4 hours ago

Note Only FIM-Midtrain 8B
TIGER-Lab/FIM-Midtraining-400K

Viewer • Updated about 4 hours ago • 524k • 32

Note FIM mid-training corpus: 400K function-aware FIM samples / 2.6B tokens from 968 repos