SetFit dependency >= 1.0.0 (#4467)

Closes #4435.

Updating SetFit to bump the version >= 1.0.0

**Type of change**
- [ ] Bug fix (non-breaking change which fixes an issue)
- [ ] New feature (non-breaking change which adds functionality)
- [ ] Breaking change (fix or feature that would cause existing
functionality to not work as expected)
- [ ] Refactor (change restructuring the codebase without changing
functionality)
- [X] Improvement (change adding some improvement to an existing
functionality)
- [ ] Documentation update

**How Has This Been Tested**

(Please describe the tests that you ran to verify your changes. And
ideally, reference `tests`)

- [x] Running the SetFit tutorial
- [x] Running SetFit tests on the suite

environment_dev.yml

@@ -52,7 +52,7 @@ dependencies:
       - transformers[torch]>=4.30.0 # <- required for DPO with TRL
       - evaluate
       - seqeval
-      - setfit>=0.7.0,<1.0.0
+      - setfit>=1.0.0
       - span_marker
       - openai>=0.27.10,<1.0.0
       - peft

pyproject.toml

@@ -109,7 +109,7 @@ integrations = [
     "evaluate",
     "seqeval",
     "sentence-transformers",
-    "setfit>=0.7.0,<1.0.0",
+    "setfit>=1.0.0",
     "span_marker",
     "sentence-transformers>=2.0.0,<3.0.0",
     "textdescriptives>=2.7.0,<3.0.0",

src/argilla/training/setfit.py

@@ -27,7 +27,7 @@ class ArgillaSetFitTrainer(ArgillaTransformersTrainer):
     _logger.setLevel(logging.INFO)
 
     def __init__(self, *args, **kwargs):
-        require_dependencies(["torch", "datasets", "transformers", "setfit>=0.6"])
+        require_dependencies(["torch", "datasets", "transformers", "setfit>=1.0.0"])
         if kwargs.get("model") is None and "model" in kwargs:
             kwargs["model"] = "all-MiniLM-L6-v2"
             self._logger.warning(f"No model defined. Using the default model {kwargs['model']}.")
@@ -39,6 +39,11 @@ def __init__(self, *args, **kwargs):
             raise NotImplementedError("SetFit only supports the `TextClassification` task.")
 
         if self._multi_label:
+            # We shall rename binarized_label as label, we need to remove the column that was previously called label.
+            # This change is due to SetFit version >=1.0.0
+            self._dataset = self._dataset.remove_columns("label")
+            self._eval_dataset = self._eval_dataset.remove_columns("label")
+            self._train_dataset = self._train_dataset.remove_columns("label")
             self._column_mapping = {"text": "text", "binarized_label": "label"}
             self.multi_target_strategy = "one-vs-rest"
         else: