[ERROR] Command failed with exit code: -1073741819

[JoergDoellinger]

Dear Carafe developers,

I am trying to test the Carafe GUI on a windows machine. I have installed python and the packages using the install button and received no error. I am using DiaNN 2.5.0.
When I start a Carafe workflow I always receive the same error: [ERROR] Command failed with exit code: -1073741819

What does it mean and how can I fix it?

I have a second question: I have a project with 120 .d files. How many would you use as training LCMS files? I selected 10 %. Does that make sense to you?

The console output is:

Workflow: 3
Date: 2026-06-05 09:15:10

[INFO] Saving parameter panel screenshots to: T:\Doellingerj\ZBS1\Carafe\parameter_screenshots
 - Saved: full_window_capture.png (Full Window)
 - Saved: settings_workflow.png
 - Saved: settings_training_data.png
 - Saved: settings_model_training.png
 - Saved: settings_library_generation.png


========================================
Running: Run DIA-NN search on the training MS data
Command: C:\DIA-NN\2.5.0\diann.exe --f T:\Doellingerj\ZBS1\LCMS\T001_MPXV_Ia_A1_2dpi_dermis_S1-A1_1_10883.d --f T:\Doellingerj\ZBS1\LCMS\T011_MPXV_Ib_A2_2dpi_dermis_S1-C2_1_10827.d --f T:\Doellingerj\ZBS1\LCMS\T021_MPXV_IIa_A3_2dpi_dermis_S1-E3_1_11009.d --f T:\Doellingerj\ZBS1\LCMS\T031_MPXV_IIb_A4_2dpi_dermis_S1-G4_1_10807.d --f T:\Doellingerj\ZBS1\LCMS\T041_MPXV_Ia_B1_5dpi_dermis_S1-A6_1_10837.d --f T:\Doellingerj\ZBS1\LCMS\T051_MPXV_Ib_B2_5dpi_dermis_S1-C7_1_10959.d --f T:\Doellingerj\ZBS1\LCMS\T061_MPXV_IIa_B3_5dpi_dermis_S1-E8_1_10957.d --f T:\Doellingerj\ZBS1\LCMS\T071_MPXV_IIb_B4_5dpi_dermis_S1-G9_1_10823.d --f T:\Doellingerj\ZBS1\LCMS\T081_MPXV_Ia_C1_9dpi_dermis_S1-A11_1_10819.d --f T:\Doellingerj\ZBS1\LCMS\T091_MPXV_Ib_C2_9dpi_dermis_S1-C12_1_11027.d --f T:\Doellingerj\ZBS1\LCMS\T101_MPXV_IIa_C3_9dpi_dermis_S2-E1_1_10811.d --f T:\Doellingerj\ZBS1\LCMS\T111_MPXV_IIb_C4_9dpi_dermis_S2-G2_1_10871.d --lib T:\Doellingerj\ZBS1\Library\lib.predicted.speclib --gen-spec-lib --threads 48 --verbose 1 --out T:\Doellingerj\ZBS1\Carafe\diann_train\report.parquet --out-lib T:\Doellingerj\ZBS1\Carafe\diann_train\report-lib.parquet --unimod4 --cut K*,R* --missed-cleavages 1 --met-excision --min-pep-len 7 --max-pep-len 35 --min-pr-mz 350 --max-pr-mz 1150 --min-pr-charge 2 --max-pr-charge 4 --min-fr-mz 200 --max-fr-mz 1700 --qvalue 0.01 --matrices --reanalyse --mass-acc 15 --mass-acc-ms1 15 --rt-profiling --export-quant
========================================

[DEBUG] DIANN env: OMP_NUM_THREADS=48, MKL_NUM_THREADS=48, KMP_AFFINITY=(unset)

DIA-NN 2.5.0 Academia  (Data-Independent Acquisition by Neural Networks)
Compiled on Apr 12 2026 17:01:59
Current date and time: Fri Jun  5 09:15:14 2026
CPU: GenuineIntel Intel(R) Xeon(R) Platinum 8160 CPU @ 2.10GHz
SIMD instructions: AVX AVX2 AVX512CD AVX512F FMA SSE4.1 SSE4.2 
Logical CPU cores: 48
177Gb out of 190Gb RAM is free
A spectral library will be generated
Thread number set to 48
Cysteine carbamidomethylation enabled as a fixed modification
In silico digest will involve cuts at K*,R*
Maximum number of missed cleavages set to 1
N-terminal methionine excision enabled
Min peptide length set to 7
Max peptide length set to 35
Min precursor m/z set to 350
Max precursor m/z set to 1150
Min precursor charge set to 2
Max precursor charge set to 4
Min fragment m/z set to 200
Max fragment m/z set to 1700
Output will be filtered at 0.01 FDR
Precursor/protein x samples expression level matrices will be saved along with the main report
MBR enabled; .quant files will only be saved to disk during the first pass
The spectral library (if generated) will retain the original spectra but will include empirically-aligned RTs
full quantification information will be exported in .parquet format
Mass accuracy will be fixed to 1.5e-05 (MS2) and 1.5e-05 (MS1)
WARNING: it is recommended to specify a FASTA database


[ERROR] Command failed with exit code: -1073741819

[wenbostar]

Is the disk T a network mapped drive?

[JoergDoellinger]

No, T is not a mapped drive. I have copied all data to C and received the same error.

[wenbostar]

This error is directly from DIA-NN. Could you directly run the DIA-NN search through DIA-NN's GUI to see if you get the same error?

[JoergDoellinger]

DIA-NN 2.5.0 has worked just fine using the same .d, .speclib. and .fasta files located on disk T.

[wenbostar]

This issue has been fixed. Could you try this new version https://github.com/Noble-Lab/Carafe/releases/tag/v2.1.1?

[JoergDoellinger]

Thank you for providing the new release! Currently I have the problem, that whatever I do, carafe is crashing during DiaNN searches without an error, I have attached you the log file. DiaNN itself is working on that PC and on that data.

carafe_log.txt

[wenbostar]

Could you send me your inputs so that I can reproduce the issue? I cannot see any error message from the log file.

[wenbostar]

Hi @JoergDoellinger , the test works well on my side. Could you share one of the .d data files along with the input library file, lib.predicted.speclib, so I can reproduce the issue?