From e3bb7afa1db9a2a8369da02c2a4a094c83f53da5 Mon Sep 17 00:00:00 2001
From: Robert <robert.samoilescu@gmail.com>
Date: Fri, 6 Dec 2024 10:30:24 +0000
Subject: [PATCH] Updated CONTRIBUTING.md

---
 CONTRIBUTING.md | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md
index aecc9b7d7..e542782d2 100644
--- a/CONTRIBUTING.md
+++ b/CONTRIBUTING.md
@@ -34,9 +34,15 @@ errors, the commit will fail and you will see the changes that need to be made.
 We use `pytest` to run tests.
 Because `alibi` uses some TensorFlow 1.x constructs, to run all tests you need to invoke `pytest` twice as follows:
 ```bash
-pytest -m tf1 alibi
-pytest -m "not tf1 alibi"
+TF_USE_LEGACY_KERAS=1 pytest -m "tf1" alibi
+pytest -m "not tf1" alibi
 ```
+
+or run directly:
+```bash
+make test
+```
+
 [see also here](https://github.com/SeldonIO/alibi/blob/4d4f49e07263b20a25f552a8485844dc12281074/.github/workflows/ci.yml#L46-L47).
 It is not necessary to run the whole test suite locally for every PR as this can take a long time, it is enough to run `pytest`
 only on the affected test files or test functions. The whole test suite is run in CI on every PR.