Other
ConTextTab
tabular
foundation-model
deep-learning
in-context
marekpolewczyk commited on
Commit
a19ec95
·
verified ·
1 Parent(s): 8067157

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +19 -3
README.md CHANGED
@@ -16,7 +16,7 @@ tags:
16
 
17
  ## Description
18
 
19
- Implementation of the deep learning model with the inference pipeline described in the paper "ConTextTab: A Semantics-Aware Tabular In-Context Learner".
20
 
21
  ![logo](./ConTextTab_architecture.png)
22
  ## Abstract
@@ -25,6 +25,8 @@ Tabular in-context learning (ICL) has recently achieved state-of-the-art (SOTA)
25
 
26
  ## Requirements
27
 
 
 
28
  The requirements are detailed in the `requirements.txt` file for Python 3.11 version.
29
 
30
  Local development installation:
@@ -53,7 +55,7 @@ X, y = load_breast_cancer(return_X_y=True)
53
  X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.5, random_state=42)
54
 
55
  # Initialize a classifier
56
- clf = ConTextTabClassifier(bagging=1, max_context_size=2048)
57
 
58
  clf.fit(X_train, y_train)
59
 
@@ -82,7 +84,7 @@ y = df.target.astype(float)
82
  X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.5, random_state=42)
83
 
84
  # Initialize the regressor
85
- regressor = ConTextTabRegressor(bagging=1, max_context_size=2048)
86
 
87
  regressor.fit(X_train, y_train)
88
 
@@ -93,6 +95,20 @@ r2 = r2_score(y_test, predictions)
93
  print("R² Score:", r2)
94
  ```
95
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
96
  ## Known Issues
97
  No known issues
98
 
 
16
 
17
  ## Description
18
 
19
+ Implementation of the deep learning model with the inference pipeline described in the paper ["ConTextTab: A Semantics-Aware Tabular In-Context Learner"](https://arxiv.org/abs/2506.10707).
20
 
21
  ![logo](./ConTextTab_architecture.png)
22
  ## Abstract
 
25
 
26
  ## Requirements
27
 
28
+ This project uses model checkpoints available on https://huggingface.co/sap-ai-research/contexttab that are automatically downloaded when running the model.
29
+
30
  The requirements are detailed in the `requirements.txt` file for Python 3.11 version.
31
 
32
  Local development installation:
 
55
  X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.5, random_state=42)
56
 
57
  # Initialize a classifier
58
+ clf = ConTextTabClassifier(bagging=1, max_context_size=2048, test_chunk_size=1000)
59
 
60
  clf.fit(X_train, y_train)
61
 
 
84
  X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.5, random_state=42)
85
 
86
  # Initialize the regressor
87
+ regressor = ConTextTabRegressor(bagging=1, max_context_size=2048, test_chunk_size=1000)
88
 
89
  regressor.fit(X_train, y_train)
90
 
 
95
  print("R² Score:", r2)
96
  ```
97
 
98
+ ## Citations
99
+
100
+ If you use this model in your research or want to refer to our work, please cite:
101
+ ```
102
+ @inproceedings{
103
+ spinaci2025contexttab,
104
+ title={ConTextTab: A Semantics-Aware Tabular In-Context Learner},
105
+ author={Marco Spinaci and Marek Polewczyk and Maximilian Schambach and Sam Thelin},
106
+ booktitle={1st ICML Workshop on Foundation Models for Structured Data},
107
+ year={2025},
108
+ url={https://openreview.net/forum?id=MmKuX9ZvM3}
109
+ }
110
+ ```
111
+
112
  ## Known Issues
113
  No known issues
114