Adding Forest-Flow: Generating and Imputing Tabular Data via Diffusion and Flow-based Gradient-Boosted Trees #69
Conversation
Add the ForestFlow notebook to the forest-flow branch.
|
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
Codecov ReportAttention:
Additional details and impacted files@@ Coverage Diff @@
## main #69 +/- ##
==========================================
+ Coverage 35.63% 35.81% +0.17%
==========================================
Files 67 67
Lines 7417 7419 +2
==========================================
+ Hits 2643 2657 +14
+ Misses 4774 4762 -12 ☔ View full report in Codecov by Sentry. |
kilianFatras
changed the title
|
@atong01 it seems that only the tests in runner are run and not the new one. Can you check that, please? Thank you |
|
@kilianFatras I had a similar issue, the directory name should be |
examples/tabular/README.md
Outdated
| ```bash | ||
| cd ../../ | ||
|
|
||
| # install requirements |
There was a problem hiding this comment.
Don't need to do this, it will be done automatically with the next line
requirements.txt
Outdated
| @@ -13,3 +13,9 @@ pot | |||
| torchdiffeq | |||
| absl-py | |||
| clean-fid | |||
| pytest | |||
|
|
|||
| # Forest-flow example | |||
There was a problem hiding this comment.
I don't think we need these here as they are in the extra requires
runner-requirements.txt
Outdated
| @@ -50,3 +50,6 @@ pot | |||
| # --------- notebook reqs -------- # | |||
| seaborn>=0.12.2 | |||
| pandas | |||
| xgboost | |||
| @@ -36,4 +36,5 @@ | |||
| long_description=readme, | |||
| long_description_content_type="text/markdown", | |||
| packages=find_packages(), | |||
| extras_require={"forest-flow": ["xgboost", "scikit-learn", "ForestDiffusion"]}, | |||
| @@ -164,6 +164,9 @@ def sample_location_and_conditional_flow(self, x0, x1, return_noise=False): | |||
| represents the source minibatch | |||
| x1 : Tensor, shape (bs, *dim) | |||
| represents the target minibatch | |||
| (optionally) t : Tensor, shape (bs) | |||
There was a problem hiding this comment.
T here has to be batch size, but I think we should allow t as float.
There was a problem hiding this comment.
Maybe better input checking would be nice, but not necessary, good for now
What does this PR do?
This PR adds a notebook on the Forest-Flow method to TorchCFM. Forest-Flow's purpose is to generate tabular data with Flow Matching methods! We have added a notebook showing how to use XGBoost to train the vector field of I-CFM and generate tabular data. This has required the addition of a parameter 't' within the 'sample_location_and_conditional_flow' function within each class. As we have modified classes, we have also added our first test function.
Before submitting
pytestcommand?pre-commit run -acommand?Did you have fun?
Make sure you had fun coding 🙃