Konic Labs' LFM2.5 work: encoders, multimodal vision extensions, adapters, compression, deployment experiments, and research artifacts.
-
konic-labs/LFM2.5-multimodal-encoder-230M
Feature Extraction • 0.2B • Updated • 16 -
konic-labs/LFM2.5-8B-A1B-REAP-50
Text Generation • 5B • Updated • 18 • 1 -
konic-labs/LFM2.5-8B-A1B-REAP-50-AWQ-INT4
Text Generation • 5B • Updated • 8 -
konic-labs/LFM2.5-VL-3B-ffn8192-int4-MLX-optimized
Image-Text-to-Text • 3B • Updated • 27 • 1