pytorch / captum

Model interpretability and understanding for PyTorch
https://captum.ai
BSD 3-Clause "New" or "Revised" License
4.8k stars 485 forks source link

Installing Captum with conda resulted in a `ModuleNotFoundError` #1207

Open aywi opened 9 months ago

aywi commented 9 months ago

🐛 Bug

To Reproduce

Steps to reproduce the behavior:

  1. Install Captum using the latest conda around with the latest PyTorch and Python 3.11 on Linux or Windows.
  2. Importing captum resulted in a ModuleNotFoundError.

Expected behavior

captum is imported without errors.

Environment

Describe the environment used for Captum

 - Captum / PyTorch Version (e.g., 1.0 / 0.4.0): 0.6.0 / 2.1.1
 - OS (e.g., Linux): Linux (Ubuntu 22.04.3 LTS), Linux (CentOS Linux release 7.9.2009), Windows (10 version 22H2)
 - How you installed Captum / PyTorch (`conda`, `pip`, source): conda
 - Build command you used (if compiling from source):
 - Python version: 3.11.5
 - CUDA/cuDNN version: 11.8
 - GPU models and configuration: NVIDIA GeForce GTX 1080, NVIDIA GeForce RTX 2080 Ti, NVIDIA GeForce GTX 1650 with Max-Q Design
 - Any other relevant information:
   - conda version: 23.10.0

Additional context

Installing Captum with pip has no such problem.

BingnanHuo commented 7 months ago

Got the same error, pip install works but conda doesnt

elijahcole commented 6 months ago

Same experience - pip worked, conda did not.

jjbastrape commented 4 months ago

conda install captum -c conda-forge worked for me today but conda install captum -c pytorch did not. Noted that the pytorch channel installed version 0.7 while the conda-forge channel installed 0.6