wehub-resource-sync
73 lines
2.7 KiB
Markdown
73 lines
2.7 KiB
Markdown
## Using Captum and MLflow to interpret Pytorch models
|
|
|
|
In this example, we will demonstrate the basic features of the [Captum](https://captum.ai/) interpretability,and logging those features using mlflow library through an example model trained on the Titanic survival data.
|
|
We will first train a deep neural network on the data using PyTorch and use Captum to understand which of the features were most important and how the network reached its prediction.
|
|
|
|
you can get more details about used attributions methods used in this example
|
|
|
|
1. [Titanic_Basic_Interpret](https://captum.ai/tutorials/Titanic_Basic_Interpret)
|
|
2. [integrated-gradients](https://captum.ai/docs/algorithms#primary-attribution)
|
|
3. [layer-attributions](https://captum.ai/docs/algorithms#layer-attribution)
|
|
|
|
### Running the code
|
|
|
|
To run the example via MLflow, navigate to the `mlflow/examples/pytorch/CaptumExample` directory and run the command
|
|
|
|
```
|
|
mlflow run .
|
|
```
|
|
|
|
This will run `Titanic_Captum_Interpret.py` with default parameter values, e.g. `--max_epochs=100` and `--use_pretrained_model False`. You can see the full set of parameters in the `MLproject` file within this directory.
|
|
|
|
In order to run the file with custom parameters, run the command
|
|
|
|
```
|
|
mlflow run . -P max_epochs=X
|
|
```
|
|
|
|
where `X` is your desired value for `max_epochs`.
|
|
|
|
If you have the required modules for the file and would like to skip the creation of a conda environment, add the argument `--env-manager=local`.
|
|
|
|
```
|
|
mlflow run . --env-manager=local
|
|
```
|
|
|
|
### Viewing results in the MLflow UI
|
|
|
|
Once the code is finished executing, you can view the run's metrics, parameters, and details by running the command
|
|
|
|
```
|
|
mlflow server
|
|
```
|
|
|
|
and navigating to [http://localhost:5000](http://localhost:5000).
|
|
|
|
For more details on MLflow tracking, see [the docs](https://www.mlflow.org/docs/latest/tracking.html#mlflow-tracking).
|
|
|
|
### Passing custom training parameters
|
|
|
|
The parameters can be overridden via the command line:
|
|
|
|
1. max_epochs - Number of epochs to train model. Training can be interrupted early via Ctrl+C
|
|
2. lr - Learning rate
|
|
3. use_pretrained_model - If want to use pretrained model
|
|
|
|
For example:
|
|
|
|
```
|
|
mlflow run . -P max_epochs=5 -P learning_rate=0.01 -P use_pretrained_model=True
|
|
```
|
|
|
|
Or to run the training script directly with custom parameters:
|
|
|
|
```sh
|
|
python Titanic_Captum_Interpret.py \
|
|
--max_epochs 50 \
|
|
--lr 0.1
|
|
```
|
|
|
|
## Logging to a custom tracking server
|
|
|
|
To configure MLflow to log to a custom (non-default) tracking location, set the MLFLOW_TRACKING_URI environment variable, e.g. via export MLFLOW_TRACKING_URI=http://localhost:5000/. For more details, see [the docs](https://mlflow.org/docs/latest/tracking.html#where-runs-are-recorded).
|