Model reference · open weights

ESRGAN

Available as managed deployment Image qualcomm Image edit 1 variants 970 dl/mo

ESRGAN is an open-weight image model from qualcomm. AxForge deploys and operates it for you on dedicated EU-owned hardware — with the licence handled where one is required.

Available as managed deployment — configured and operated for you on dedicated EU hardware, quoted per deployment.

What it is

Released byqualcomm
TypeImage models
TaskImage edit
Runs withpytorch
Released2024-02-25
Popularity970 downloads / month
LicenceOpen weights

About

What ESRGAN is

ESRGAN is a machine learning model that upscales an image with minimal loss in quality.

This is based on the implementation of ESRGAN 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.

Read the full model card

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.

RuntimePrecisionChipsetSDK VersionsDownload
ONNXfloatUniversalQAIRT 2.45, ONNX Runtime 1.27.1Download
ONNXw8a16UniversalQAIRT 2.45, ONNX Runtime 1.27.1Download
QNN_DLCfloatUniversalQAIRT 2.45Download
QNN_DLCw8a16UniversalQAIRT 2.45Download
TFLITEfloatUniversalQAIRT 2.45Download

For more device-specific assets and performance metrics, visit ESRGAN 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 ESRGAN on GitHub for usage instructions.

Model Details

Model Type: Model_use_case.super_resolution

Model Stats:

  • Input resolution: 128x128
  • Model checkpoint: ESRGAN_x4
  • Model size (float): 63.9 MB
  • Number of parameters: 16.7M

Performance Summary

ModelRuntimePrecisionChipsetInference Time (ms)Peak Memory Range (MB)Primary Compute Unit
ESRGANONNXfloatSnapdragon® X2 Elite34.318 ms8 - 8 MBNPU
ESRGANONNXfloatSnapdragon® X Elite65.423 ms38 - 38 MBNPU
ESRGANONNXfloatSnapdragon® 8 Gen 3 Mobile51.062 ms0 - 745 MBNPU
ESRGANONNXfloatSnapdragon® 8 Gen 1 Mobile118.049 ms4 - 741 MBNPU
ESRGANONNXfloatQualcomm® Dragonwing™ IQ-8275127.563 ms6 - 10 MBNPU
ESRGANONNXfloatQualcomm® Dragonwing™ QCS8550 (Proxy)63.529 ms0 - 43 MBNPU
ESRGANONNXfloatQualcomm® QCS8450118.049 ms4 - 741 MBNPU
ESRGANONNXfloatQualcomm® Dragonwing™ IQ-9075105.926 ms6 - 9 MBNPU
ESRGANONNXfloatQualcomm® Dragonwing™ IQ-X718165.423 ms38 - 38 MBNPU
ESRGANONNXfloatQualcomm® Dragonwing™ Q-875038.69 ms7 - 358 MBNPU
ESRGANONNXfloatSnapdragon® 8 Elite Mobile38.69 ms7 - 358 MBNPU
ESRGANONNXfloatSnapdragon® 8 Elite Gen 5 Mobile26.12 ms5 - 361 MBNPU
ESRGANONNXw8a16Snapdragon® X2 Elite21.996 ms4 - 4 MBNPU
ESRGANONNXw8a16Snapdragon® X Elite43.397 ms26 - 26 MBNPU
ESRGANONNXw8a16Snapdragon® 8 Gen 3 Mobile31.981 ms3 - 1115 MBNPU
ESRGANONNXw8a16Snapdragon® 8 Gen 1 Mobile83.95 ms4 - 1198 MBNPU
ESRGANONNXw8a16Qualcomm® Dragonwing™ QCS6490323.654 ms4 - 7 MBNPU
ESRGANONNXw8a16Qualcomm® Dragonwing™ IQ-827543.191 ms3 - 7 MBNPU
ESRGANONNXw8a16Qualcomm® Dragonwing™ QCS8550 (Proxy)41.903 ms0 - 725 MBNPU
ESRGANONNXw8a16Qualcomm® QCS845083.95 ms4 - 1198 MBNPU
ESRGANONNXw8a16Qualcomm® Dragonwing™ IQ-907538.006 ms3 - 6 MBNPU
ESRGANONNXw8a16Qualcomm® Dragonwing™ IQ-X718143.397 ms26 - 26 MBNPU
ESRGANONNXw8a16Qualcomm® Dragonwing™ Q-66901128.366 ms0 - 696 MBNPU
ESRGANONNXw8a16Qualcomm® Dragonwing™ Q-779089.102 ms3 - 784 MBNPU
ESRGANONNXw8a16Qualcomm® Dragonwing™ Q-875026.489 ms3 - 907 MBNPU
ESRGANONNXw8a16Snapdragon® 8 Elite Mobile26.489 ms3 - 907 MBNPU
ESRGANONNXw8a16Snapdragon® 8 Elite Gen 5 Mobile16.934 ms3 - 1072 MBNPU
ESRGANONNXw8a16Snapdragon® 7 Gen 4 Mobile89.102 ms3 - 784 MBNPU
ESRGANQNN_DLCfloatSnapdragon® X2 Elite34.371 ms0 - 0 MBNPU
ESRGANQNN_DLCfloatSnapdragon® X Elite65.02 ms0 - 0 MBNPU
ESRGANQNN_DLCfloatSnapdragon® 8 Gen 3 Mobile48.702 ms0 - 706 MBNPU
ESRGANQNN_DLCfloatSnapdragon® 8 Gen 1 Mobile112.305 ms0 - 709 MBNPU
ESRGANQNN_DLCfloatQualcomm® Dragonwing™ IQ-8275127.079 ms0 - 6 MBNPU
ESRGANQNN_DLCfloatQualcomm® Dragonwing™ QCS8550 (Proxy)62.121 ms0 - 4 MBNPU
ESRGAN

From the published model card. Full card on the HuggingFace links in the sidebar.

How it works

How image models work

Text promptwhat to makeText encoderunderstands itDiffusion stepsdenoise to pixelsImagePNG / JPEGA diffusion model starts from noise and denoises it, guided by your prompt, into a finished image.

Using it via the API

Call it like any OpenAI endpoint

Once AxForge deploys esrgan for you, it answers on the OpenAI-compatible API — the same base URL and keys as every other model. (esrgan below is illustrative; you get the exact model name on deployment.)

$ curl -sS https://api.axforge.ai/v1/images/generations \
  -H "Authorization: Bearer $AXFORGE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"esrgan","prompt":"a red bicycle","size":"1024x1024"}'

Create an account — your API key is available in the console. 3M free tokens every 30 days with every new account.

© 2026 AxForge · EU-hosted AI infrastructure Pricing Docs Trust Privacy Terms