Dependencies from environment

[rongithubbed]

Creating an environment from the yaml file results in:

Could not solve for environment specs
The following packages are incompatible
├─ numpy 1.15.0** is installable with the potential options
│ ├─ numpy 1.15.0 would require
│ │ └─ numpy-base 1.15.0 py36h3dfced4_0, which requires
│ │ └─ mkl >=2018.0.3,<2019.0a0 , which can be installed;
│ ├─ numpy 1.15.0 would require
│ │ └─ python >=2.7,<2.8.0a0 , which can be installed;
│ ├─ numpy 1.15.0 would require
│ │ └─ python >=3.5,<3.6.0a0 , which can be installed;
│ ├─ numpy 1.15.0 would require
│ │ └─ numpy-base 1.15.0 py36h7cdd4dd_0, which requires
│ │ └─ blas 1.0 openblas, which can be installed;
│ └─ numpy 1.15.0 would require
│ └─ python >=3.7,<3.8.0a0 , which can be installed;
├─ python 3.6.8** is not installable because it conflicts with any installable versions previously reported;
└─ pytorch 1.2.0** is installable with the potential options
├─ pytorch 1.2.0 would require
│ ├─ blas 1.0 mkl, which conflicts with any installable versions previously reported;
│ └─ mkl >=2019.4,<2021.0a0 , which conflicts with any installable versions previously reported;
├─ pytorch 1.2.0 would require
│ └─ python >=2.7,<2.8.0a0 , which can be installed;
└─ pytorch 1.2.0 would require
└─ python >=3.7,<3.8.0a0 , which can be installed.

  it seems like having numpy==1.12.0 is more important than having pytorch 1.2.0. I will try out different compatible configurations and upload a modified yaml here.  

[rongithubbed]

these seem to work for me so far:
ame: meshcnn
channels:

• defaults
• conda-forge
• pytorch
  dependencies:
• python=3.7
• cython
• numpy=1.17.4
• matplotlib=3.1.3
• pytorch=1.3.1
• torchvision=0.4.2
• pip
• pip:
  • git+https://github.com/lanpa/tensorboardX.git
  • pytest==5.1.1