We need additional binaries such as tesseract or pretrained models on our ML computes. As those are not preprovisioned in the Azure ML compute instances and we need to restrict external access, we need to load them differently. We try now via Azure artifacts universal packages. Unfortunately this does not work: After having installed the azure devops extension onto the ml compute instance we can download the artifact, but I cannot install it. I get the following error: Encountered an unexpected error.
System.IO.IOException: Hard linking failed!
Status: Failed
Path: tesseract/tesseract-ocr_4.00~git2288-10f4998a-2_amd64.deb
at Microsoft.VisualStudio.Services.BlobStore.WebApi.DedupStoreClient.DownloadToFileAsync(DedupNode node, String fullPath, Uri proxyUri, EdgeCache edgeCache, CancellationToken cancellationToken)
at Microsoft.VisualStudio.Services.BlobStore.WebApi.DedupStoreClientWithDataport.DownloadToFileAsync(IDedupDataPort dataport, DedupNode node, String fullPath, Uri proxyUri, EdgeCache edgeCache, CancellationToken cancellationToken)