Hugging Face's logo

qualcomm
/
LFM3-1.7B

Text Generation
PyTorch
llm
generative_ai
android
Model card Files
xet
 Community

LFM3-1.7B: Optimized for Qualcomm Devices

LFM3 is a family of hybrid language models from Liquid AI designed for fast, high-quality on-device deployment. The 1.7B model is tuned for instruction following, chat, and tool use.

This is based on the implementation of LFM3-1.7B found here. This repository contains pre-exported model files optimized for Qualcomm® devices. You can use the Qualcomm® AI Hub Models library to export with custom configurations. More details on model performance across various devices, can be found here.

Qualcomm AI Hub Models uses Qualcomm AI Hub Workbench to compile, profile, and evaluate this model. Sign up to run these models on a hosted Qualcomm® device.

Deploying LFM3-1.7B on-device

Follow the GenieX quickstart to install GenieX and deploy the model on a target device.

Getting Started

There are two ways to deploy this model on your device:

Option 1: Download Pre-Exported Models

Below are pre-exported model assets ready for deployment.

Runtime Precision Chipset SDK Versions Download
GENIEX_LLAMACPP q4_0 Universal Download

For more device-specific assets and performance metrics, visit LFM3-1.7B on Qualcomm® AI Hub.

Option 2: Export with Custom Configurations

Use the Qualcomm® AI Hub Models Python library to compile and export the model with your own:

  • Custom weights (e.g., fine-tuned checkpoints)
  • Custom input shapes
  • Target device and runtime configurations

This option is ideal if you need to customize the model beyond the default configuration provided here.

See our repository for LFM3-1.7B on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.text_generation

Model Stats:

  • Response Rate: Rate of response generation after the first response token.

Performance Summary

Model Runtime Precision Chipset Context Length Response Rate (tokens per second) Time To First Token (range, seconds)
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 512 51.151488 0.29636 - 1.18544
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 512 44.87391 0.17198766666666665 - 0.6879506666666666
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 512 42.830536 0.039266999999999996 - 0.15706799999999999
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 4096 39.757529 0.6073130322580645 - 19.434017032258065
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 4096 40.111837 0.21321906451612904 - 6.823010064516129
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® 8 Elite Gen 5 For Galaxy Mobile 4096 46.016466 0.051927999999999995 - 1.6616959999999998
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® 8 Elite For Galaxy Mobile 512 53.408016 0.3475826666666667 - 1.3903306666666668
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® 8 Elite For Galaxy Mobile 512 45.206583 0.19607233333333332 - 0.7842893333333333
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® 8 Elite For Galaxy Mobile 512 40.891133 0.04947366666666666 - 0.19789466666666664
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® 8 Elite For Galaxy Mobile 4096 35.956955 0.6050833548387097 - 19.36266735483871
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® 8 Elite For Galaxy Mobile 4096 36.058034 0.2817183870967742 - 9.014988387096775
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® 8 Elite For Galaxy Mobile 4096 35.951138 0.055328225806451616 - 1.7705032258064517
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® X2 Elite 512 86.380845 0.08983066666666667 - 0.3593226666666667
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® X2 Elite 512 57.465794 0.08886166666666664 - 0.3554466666666666
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® X2 Elite 512 52.561447 0.039210999999999996 - 0.15684399999999998
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® X2 Elite 4096 72.499887 0.11773925806451613 - 3.767656258064516
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® X2 Elite 4096 55.279231 0.10467622580645161 - 3.3496392258064516
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® X2 Elite 4096 42.133176 0.046599935483870966 - 1.491197935483871
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® X Elite 512 72.70765 0.15710833333333332 - 0.6284333333333333
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® X Elite 512 48.97021 0.18853899999999998 - 0.7541559999999999
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® X Elite 512 33.555912 0.08885599999999999 - 0.35542399999999996
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® X Elite 4096 55.831076 0.25588406451612905 - 8.18829006451613
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® X Elite 4096 39.675086 0.236493 - 7.567776
LFM3-1.7B GENIEX_LLAMACPP q4_0 Snapdragon® X Elite 4096 31.213026 0.09012703225806452 - 2.8840650322580648
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ IQ-8275 512 18.546421 1.4220739999999998 - 5.688295999999999
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ IQ-8275 512 7.684679 2.1180676666666667 - 8.472270666666667
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ IQ-8275 512 29.403475 0.15746266666666664 - 0.6298506666666666
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ IQ-8275 4096 9.572464 1.7507801612903224 - 56.02496516129032
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ IQ-8275 4096 7.094126 2.378988354838709 - 76.1276273548387
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ IQ-8275 4096 25.862583 0.1589815806451613 - 5.0874105806451615
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ IQ-9075 512 41.047678 0.5884536666666667 - 2.3538146666666666
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ IQ-9075 512 18.527265 0.7198109999999999 - 2.8792439999999995
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ IQ-9075 512 42.031735 0.100036 - 0.400144
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ IQ-X7181 512 72.70765 0.15710833333333332 - 0.6284333333333333
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ IQ-X7181 512 48.97021 0.18853899999999998 - 0.7541559999999999
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ IQ-X7181 512 33.555912 0.08885599999999999 - 0.35542399999999996
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ Q-8750 512 53.408016 0.3475826666666667 - 1.3903306666666668
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ Q-8750 512 45.206583 0.19607233333333332 - 0.7842893333333333
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ Q-8750 512 40.891133 0.04947366666666666 - 0.19789466666666664
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ IQ-9075 4096 35.366033 0.6821005806451612 - 21.82721858064516
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ IQ-9075 4096 17.162926 0.8255575483870967 - 26.417841548387095
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ IQ-9075 4096 39.191288 0.10069006451612902 - 3.2220820645161288
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ IQ-X7181 4096 55.831076 0.25588406451612905 - 8.18829006451613
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ IQ-X7181 4096 39.675086 0.236493 - 7.567776
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ IQ-X7181 4096 31.213026 0.09012703225806452 - 2.8840650322580648
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ Q-8750 4096 35.956955 0.6050833548387097 - 19.36266735483871
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ Q-8750 4096 36.058034 0.2817183870967742 - 9.014988387096775
LFM3-1.7B GENIEX_LLAMACPP q4_0 Qualcomm® Dragonwing™ Q-8750 4096 35.951138 0.055328225806451616 - 1.7705032258064517

License

  • The license for the original implementation of LFM3-1.7B can be found here.

References

Community

Usage and Limitations

This model may not be used for or in connection with any of the following applications:

  • Accessing essential private and public services and benefits;
  • Administration of justice and democratic processes;
  • Assessing or recognizing the emotional state of a person;
  • Biometric and biometrics-based systems, including categorization of persons based on sensitive characteristics;
  • Education and vocational training;
  • Employment and workers management;
  • Exploitation of the vulnerabilities of persons resulting in harmful behavior;
  • General purpose social scoring;
  • Law enforcement;
  • Management and operation of critical infrastructure;
  • Migration, asylum and border control management;
  • Predictive policing;
  • Real-time remote biometric identification in public spaces;
  • Recommender systems of social media platforms;
  • Scraping of facial images (from the internet or otherwise); and/or
  • Subliminal manipulation
Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
Text Generation
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support