mlbio-epfl / turtle

[ICML 2024] Let Go of Your Labels with Unsupervised Transfer
https://brbiclab.epfl.ch/projects/turtle/
45 stars 5 forks source link

Dinov2 killed issue #2

Closed coporteer closed 5 months ago

coporteer commented 5 months ago

Hello, i am trying replicate papers result with github resource. But i cannnot help but raise issue in precompute_representations.py

When i am trying to do:

python3 precompute_representations.py --dataset mnist --phis dinov2 --batch_size 100

i got killed without error

i setup my environment with github guide

# Pull git
git clone https://github.com/mlbio-epfl/turtle.git

# Install dependency
pip3 install -U xformers --index-url https://download.pytorch.org/whl/cu121
pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121
pip install numpy scipy scikit-learn clip tqdm
pip install \
    --extra-index-url=https://pypi.nvidia.com \
    cudf-cu12==24.6.* dask-cudf-cu12==24.6.* cuml-cu12==24.6.* \
    cugraph-cu12==24.6.* cuspatial-cu12==24.6.* cuproj-cu12==24.6.* \
    cuxfilter-cu12==24.6.* cucim-cu12==24.6.* pylibraft-cu12==24.6.* \
    raft-dask-cu12==24.6.* cuvs-cu12==24.6.*

thanks in advances

coporteer commented 5 months ago

For someone in concern, I solved this issue by using 12G VRAM GPU