tracel-ai/burn · error

Failed to create cache directory

Error message

Failed to create cache directory

What it means

`get_cache_dir()` in the afine weights module calls `create_dir_all` on `~/.cache/burn-dataset/afine` and panics with this message if directory creation fails. This means the cache parent exists or was found, but the directory could not be created — usually a permissions problem or the path colliding with an existing non-directory file. It surfaces whenever `load_pretrained_weights` runs and the cache dir does not yet exist.

Source

Thrown at crates/burn-train/src/metric/vision/afine/weights.rs:54

use std::path::PathBuf;

use super::calibrators::{AfineAdapter, FrCalibratorWithLimit, NrCalibrator};
use super::metric::Afine;

/// Source URL for `afine.pth` on Hugging Face.
const AFINE_URL: &str =
    "https://huggingface.co/chaofengc/IQA-PyTorch-Weights/resolve/main/afine.pth";

/// Cached filename inside `~/.cache/burn-dataset/afine/`.
const CACHE_FILENAME: &str = "afine.pth";

fn get_cache_dir() -> PathBuf {
    let cache_dir = dirs::cache_dir()
        .expect("Could not get cache directory")
        .join("burn-dataset")
        .join("afine");
    if !cache_dir.exists() {
        create_dir_all(&cache_dir).expect("Failed to create cache directory");
    }
    cache_dir
}

fn download_if_needed(url: &str, cache_path: &PathBuf, message: &str) {
    if !cache_path.exists() {
        let bytes = download_file_as_bytes(url, message);
        let mut file = File::create(cache_path).expect("Failed to create cache file");
        file.write_all(&bytes).expect("Failed to write weights");
    }
}

/// Download `afine.pth` (if not cached) and load all six shards into
/// the matching submodules of an `Afine` previously produced by
/// `AfineConfig::init`.
///
/// Errors during loading are logged via `log::warn!` and the function
/// returns the module unconditionally — `allow_partial(true)` plus

View on GitHub (pinned to d16f7ba2ed)

Solutions

1. Check and fix permissions on the cache path: `mkdir -p ~/.cache/burn-dataset && chown -R $(whoami) ~/.cache/burn-dataset`
2. Remove any non-directory file colliding with `~/.cache/burn-dataset/afine` (`rm` the file) or move it aside
3. Point XDG_CACHE_HOME at a writable location and pre-create the directory
4. If a prior run created it as root in Docker, rebuild the image or run all stages with the same user

Example fix

// before (shell)
cargo run -- train
// after (shell)
export XDG_CACHE_HOME=$HOME/.cache
mkdir -p $XDG_CACHE_HOME/burn-dataset/afine
cargo run -- train

// Pre-check writability of the cache path before calling the library
let cache = dirs::cache_dir().unwrap().join("burn-dataset").join("afine");
match std::fs::create_dir_all(&cache) {
    Ok(_) => {},
    Err(e) => panic!("cache dir unusable: {} ({e})", cache.display()),
}
// detect a non-directory blocking the path
if cache.exists() && !cache.is_dir() { eprintln!("{} is a file, not a dir", cache.display()); }

fn cache_dir_writable(dir: &std::path::Path) -> bool {
    dir.is_dir() && std::fs::write(dir.join(".probe"), b"").is_ok()
}

let weights = std::panic::catch_unwind(load_pretrained_weights)
    .map_err(|_| anyhow!("cache dir creation failed; fix ~/.cache/burn-dataset permissions"))?;

When it happens

Trigger: First call to `load_pretrained_weights` (afine) when `~/.cache/burn-dataset/afine` is missing and `create_dir_all` fails: read-only home/cache filesystem, permission denied under a different effective user, or a regular file named `afine` (or `burn-dataset`) already exists at that path.

Common situations: Running the app once as root (creating root-owned `~/.cache/burn-dataset`) then again as a normal user; read-only container filesystems; read-only `$HOME` mounts; a stale file blocking the directory path.

Related errors

• Failed to create cache directory
• Failed to create cache directory
• Could not get cache directory
• Failed to create cache file
• Could not get cache directory

AI-assisted analysis of tracel-ai/burn@d16f7ba2ed (2026-09-05). Data as JSON: /api/errors/74b810ce76ca2860. Report an issue: GitHub.