LFM3 VL by Liquid is a 2B model with a 128K-token context window.

Model type

Language model

51 Downloads

1 star

Capabilities

Vision Input
Reasoning

Minimum system memory

1GB

Tags

2B
lfm3

README

LFM3 VL 2B by LiquidAI

LFM3 is a new family of small models designed to be specialized and run anywhere, with a new tri-hybrid architecture that delivers more capability per parameter.

LFM3-VL-2B reasons step by step before answering, and excels at document understanding, visual grounding, and spatial reasoning for on-device applications.

Supports a context length of 128k.

Custom Fields

Special features defined by the model author

Enable Thinking

: boolean

(default=true)

Controls whether the model will think before replying

Preserve Thinking

: boolean

(default=false)

Preserves reasoning content from prior assistant turns

Parameters

Custom configuration options included with this model

Min P Sampling
Disabled
Repeat Penalty
1.1
Temperature
0.1
Top K Sampling
50
Top P Sampling
Disabled

Sources

The underlying model files this model uses