Can i run the cli on google colab?

[Keshav-writes-code]

i dont have a powerfull enough gpu, can i run it on google colab?

[dosubot[bot]]

Hi @Keshav-writes-code! I'm Dosu and I’m helping the upscayl team.

Upscayl CLI requires a Vulkan-compatible GPU to work, and Google Colab's GPUs do not provide Vulkan support—they only support CUDA. This means you can't run Upscayl CLI on Google Colab, even though you can install dependencies and access an NVIDIA GPU. There are no official or community-supported instructions or reports of anyone successfully running Upscayl CLI on Colab, and the lack of Vulkan support is the main blocker for this setup reference.

If you need cloud-based upscaling, you might want to look for services or platforms that offer direct access to a Vulkan-compatible GPU, or consider other upscaling tools that work with CUDA-only environments.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}

   

[Keshav-writes-code]

where can i run it then

[dosubot[bot]]

You can run Upscayl on any computer with a dedicated GPU that supports Vulkan 1.3—this includes most modern desktops and laptops with recent NVIDIA or AMD graphics cards. It's officially supported on Linux, macOS (12+), and Windows (10+) and is available as an installer or portable app for each platform details.

If your hardware doesn't support Vulkan 1.3, you can use the Upscayl online version, which runs on remote servers and doesn't depend on your local GPU reference.

Running Upscayl on cloud VMs is possible if the provider offers a Vulkan-compatible GPU, but this setup isn't officially documented or supported. Google Colab and most cloud notebooks only support CUDA, so they won't work for Upscayl.

If you need a solution for CUDA-only environments like Colab, you might want to look into other AI upscaling tools designed for CUDA GPUs.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}

   

[Keshav-writes-code]

okay, i found a service that matches, how can i exatly run the cli on this linux server?

[dosubot[bot]]

To run the Upscayl CLI on your Linux server:

1. Make sure your server has a Vulkan-compatible GPU and the correct drivers installed. You can check Vulkan support with vulkaninfo or a tool like VulkanCapsViewer. If you run into issues, reinstall your graphics drivers cite.

2. Download the Upscayl CLI binary. The easiest way is to use the update_upscayl_ncnn_binaries.sh script from the repo, which fetches and extracts the latest CLI binary for Linux into resources/linux/bin/upscayl-bin cite.

3. Prepare your input images and create an output directory.

4. In your terminal, navigate to the Upscayl installation directory and run the CLI with a command like:

   ./resources/bin/upscayl-bin -i input_folder -o output_folder -s 4 -m ./resources/models -n upscayl-standard-4x -f png
   

   Replace input_folder and output_folder with your paths, and pick the model you want (e.g., upscayl-standard-4x, ultramix-balanced-4x, etc.). You can also use single image files instead of folders cite.

   Common flags:

   • -i input file or directory
   • -o output file or directory
   • -s scale factor (e.g., 2, 4)
   • -m models path (e.g., ./resources/models)
   • -n model name (e.g., upscayl-standard-4x)
   • -f output format (png, jpg)
   • -g GPU ID (if you have multiple GPUs)
   • -t tile size, -c compression, -x TTA mode (optional advanced options)

5. The CLI will use your GPU directly (unlike the GUI, which sometimes falls back to software rendering). You should see your GPU listed in the logs, not "SwiftShader" cite.

If you need to build from source or want more advanced usage, check the upscayl-ncnn repo for prerequisites and build instructions (cmake, gcc-9, g++-9, Vulkan SDK) cite.

Let me know if you need help with a specific step or run into errors!

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}