Problem with installing dependancies

[raoden1]

I cannot install dependancies:

**conda env create -f env.yml**

Solving environment: failed

ResolvePackageNotFound: 
  - python==3.6.5=hc167b69_0
  - xz==5.2.3=h0278029_2
  - numpy==1.12.1=py36h8871d66_1
  - zlib==1.2.11=hf3cbc9b_2
  - six==1.11.0=py36h0e22d5e_1
  - readline==7.0=hc1231fa_4
  - sqlite==3.23.1=hf1716c9_0
  - libgfortran==3.0.1=h93005f0_2
  - libedit==3.1=hb4e282d_0
  - pyyaml==3.12=py36h2ba1e63_1
  - libffi==3.2.1=h475c297_4
  - hdf5==1.10.1=ha036c08_1
  - protobuf==3.5.2=py36h0a44026_0
  - libprotobuf==3.5.2=h2cd40f5_0
  - libcxx==4.0.1=h579ed51_0
  - scipy==1.0.1=py36hcaad992_0
  - openssl==1.0.2o=h26aff7b_0
  - libcxxabi==4.0.1=hebd6815_0
  - yaml==0.1.7=hc338f04_2
  - pandas==0.22.0=py36h0a44026_0
  - tk==8.6.7=h35a86e2_3
  - h5py==2.7.1=py36h39cdac5_0
  - ncurses==6.0=hd04f020_2a

[OsandaG]

Seems to same the same problem, with anaconda python 3.7 on Ubuntu 18.04, so changed the export file as follows:

`name: drlts
channels:

• conda-forge
• bioconda
• bokeh
• defaults
  dependencies:
  • ca-certificates=2018.03.07
  • certifi=2018.4.16
  • h5py=2.7.1
  • hdf5=1.10.1
  • intel-openmp=2018.0.0
  • keras=2.1.5
  • libcxx
  • libedit=3.1
  • libffi=3.2.1
  • libgfortran
  • libprotobuf=3.5.2
  • mkl=2018.0.2
  • ncurses=6.0
  • numpy=1.12.1
  • openssl=1.0.2o
  • pandas=0.22.0
  • pip=9.0.3
  • protobuf=3.5.2
  • python=3.6.5
  • python-dateutil=2.7.2
  • pytz=2018.4
  • pyyaml=3.12
  • readline=7.0
  • scipy=1.0.1
  • setuptools=39.0.1
  • six=1.11.0
  • sqlite=3.23.1
  • tensorflow=1.1.0
  • tk=8.6.7
  • werkzeug=0.14.1
  • wheel=0.31.0
  • xz=5.2.3
  • yaml=0.1.7
  • zlib=1.2.11
    `

[OsandaG]

Think Theano will need to be complied on your local system

[greg2paris]

I created a simple google colab file from this implementation, so you can easily run it with your own google drive and colab.

1. Just copy the entire directory from this github to your google drive directory.
2. Copy the google colab file I created.
3. Change the "google_drive_directory_path" value to your own google drive directory.
4. Use python 3 and GPU accelerator.
5. Run the code.

https://colab.research.google.com/drive/1GMxAUdpU38DFOsSL4r0AglpdIe00YpAP

Tell me if you have any problem running the code,
Greg

[zhongkeshuanghe]

I didn't find your train/ test/ folder used in your notebook