juglab / hdn Goto Github PK

Hi,
I am getting a Solving environment: failed error with a long list of packages that can not be found. Is there another way to import/install the environment? Thanks

ResolvePackageNotFound:

• imagecodecs==2020.5.30=py37hfa7d478_2
• libzopfli==1.0.3=he1b5a44_0
• libffi==3.3=he6710b0_2
• scikit-image==0.17.2=py37hdf5156a_0
• libaec==1.0.4=he1b5a44_1
• pip==20.3.3=py37h06a4308_0
• libgcc-ng==9.1.0=hdf63c60_0
• lz4-c==1.9.2=heb0550a_3
• mkl_fft==1.2.0=py37h23d657b_0
• scipy==1.5.2=py37h0b6359f_0
• readline==8.0=h7b6447c_0
• jpeg==9b=h024ee3a_2
• mkl_random==1.1.1=py37h0573a6f_0
• ld_impl_linux-64==2.33.1=h53a641e_7
• lcms2==2.11=h396b838_0
• bzip2==1.0.8=h516909a_3
• brotli==1.0.9=he1b5a44_3
• charls==2.1.0=he1b5a44_2
• giflib==5.1.7=h516909a_1
• matplotlib-base==3.3.1=py37h817c723_0
• libtiff==4.1.0=h2733197_1
• numpy-base==1.19.2=py37hfa32c7d_0
• xz==5.2.5=h7b6447c_0
• blosc==1.20.1=he1b5a44_0
• libgfortran-ng==7.3.0=hdf63c60_0
• mkl-service==2.3.0=py37he8ac12f_0
• libpng==1.6.37=hbc83047_0
• freetype==2.10.4=h5ab3b9f_0
• libwebp==1.0.1=h8e7db2f_0
• scikit-learn==0.23.2=py37h0573a6f_0
• zlib==1.2.11=h7b6447c_3
• yaml==0.2.5=h7b6447c_0
• pyyaml==5.3.1=py37h7b6447c_1
• cytoolz==0.11.0=py37h7b6447c_0
• zstd==1.4.5=h9ceee32_0
• openjpeg==2.3.1=h981e76c_3
• openssl==1.1.1h=h7b6447c_0
• sqlite==3.33.0=h62c20be_0
• jxrlib==1.1=h516909a_2
• cudatoolkit==10.1.243=h6bb024c_0
• libedit==3.1.20191231=h14c3975_1
• setuptools==51.0.0=py37h06a4308_2
• six==1.15.0=py37h06a4308_0
• python==3.7.9=h7579374_0
• pytorch==1.7.1=py3.7_cuda10.1.243_cudnn7.6.3_0
• libuv==1.40.0=h7b6447c_0
• tk==8.6.10=hbc83047_0
• pywavelets==1.1.1=py37h7b6447c_2
• libstdcxx-ng==9.1.0=hdf63c60_0
• ninja==1.10.2=py37hff7bd54_0
• snappy==1.1.8=he1b5a44_3
• ncurses==6.2=he6710b0_1
• numpy==1.19.2=py37h54aff64_0
• pillow==8.0.1=py37he98fc37_0

I tried installing via .yml but got a "ResolvePackageNotFound" error on a bunch of libraries (file attached). Here's what I did:

1. on the .yml link -> open this file in GitHub Desktop (the folder is cloned in my system)
2. Anaconda prompt -> cd "....\HDN"
3. conda env create -f HDN.yml
4. Collecting package metadata (repodata.json): done -> Solving environment: failed
5. Below the packages returned as not found

ResolvePackageNotFound:

• libgfortran-ng==7.3.0=hdf63c60_0
• lz4-c==1.9.2=heb0550a_3
• zstd==1.4.5=h9ceee32_0
• snappy==1.1.8=he1b5a44_3
• ncurses==6.2=he6710b0_1
• giflib==5.1.7=h516909a_1
• openjpeg==2.3.1=h981e76c_3
• libffi==3.3=he6710b0_2
• mkl_random==1.1.1=py37h0573a6f_0
• jpeg==9b=h024ee3a_2
• python==3.7.9=h7579374_0
• bzip2==1.0.8=h516909a_3
• cudatoolkit==10.1.243=h6bb024c_0
• freetype==2.10.4=h5ab3b9f_0
• mkl-service==2.3.0=py37he8ac12f_0
• pytorch==1.7.1=py3.7_cuda10.1.243_cudnn7.6.3_0
• sqlite==3.33.0=h62c20be_0
• libzopfli==1.0.3=he1b5a44_0
• scikit-image==0.17.2=py37hdf5156a_0
• ninja==1.10.2=py37hff7bd54_0
• pyyaml==5.3.1=py37h7b6447c_1
• libedit==3.1.20191231=h14c3975_1
• pywavelets==1.1.1=py37h7b6447c_2
• libgcc-ng==9.1.0=hdf63c60_0
• libuv==1.40.0=h7b6447c_0
• numpy==1.19.2=py37h54aff64_0
• cytoolz==0.11.0=py37h7b6447c_0
• mkl_fft==1.2.0=py37h23d657b_0
• setuptools==51.0.0=py37h06a4308_2
• pillow==8.0.1=py37he98fc37_0
• openssl==1.1.1h=h7b6447c_0
• libaec==1.0.4=he1b5a44_1
• tk==8.6.10=hbc83047_0
• ld_impl_linux-64==2.33.1=h53a641e_7
• scikit-learn==0.23.2=py37h0573a6f_0
• numpy-base==1.19.2=py37hfa32c7d_0
• lcms2==2.11=h396b838_0
• matplotlib-base==3.3.1=py37h817c723_0
• blosc==1.20.1=he1b5a44_0
• jxrlib==1.1=h516909a_2
• zlib==1.2.11=h7b6447c_3
• readline==8.0=h7b6447c_0
• libpng==1.6.37=hbc83047_0
• charls==2.1.0=he1b5a44_2
• libtiff==4.1.0=h2733197_1
• imagecodecs==2020.5.30=py37hfa7d478_2
• xz==5.2.5=h7b6447c_0
• scipy==1.5.2=py37h0b6359f_0
• six==1.15.0=py37h06a4308_0
• pip==20.3.3=py37h06a4308_0
• libstdcxx-ng==9.1.0=hdf63c60_0
• brotli==1.0.9=he1b5a44_3
• libwebp==1.0.1=h8e7db2f_0
• yaml==0.2.5=h7b6447c_0

Hello everyone,

scikit-learn v1.0.2 breaks the code here:

Downgraded to v0.24.1 like it is written in the requirements.txt and everything works fine. Just leaving this here for documentation.

Hi! Thank you for offering the code.
When I try to train the HDN network on my own microscopy data (data type: int16 data range: -179 to 20351), at first everything goes well. But when it comes to 32 epochs, it reports 'Value Error':

Using the same training parameters, the Convallaria dataset didn't encounter this problem. I would appreciate it if you could help me!