Stop requiring Torch for our TF examples! (#21997)

* Stop requiring Torch for our TF examples!

* Slight tweak to logging in the example itself

examples/tensorflow/contrastive-image-text/run_clip.py

@@ -273,9 +273,8 @@ def main():
         handlers=[logging.StreamHandler(sys.stdout)],
     )
 
-    if training_args.should_log:
-        # The default of training_args.log_level is passive, so we set log level at info here to have that default.
-        transformers.utils.logging.set_verbosity_info()
+    # The default of training_args.log_level is passive, so we set log level at info here to have that default.
+    transformers.utils.logging.set_verbosity_info()
 
     log_level = training_args.get_process_log_level()
     logger.setLevel(log_level)

src/transformers/training_args_tf.py

@@ -249,6 +249,13 @@ class TFTrainingArguments(TrainingArguments):
         requires_backends(self, ["tf"])
         return self._setup_strategy.num_replicas_in_sync
 
+    @property
+    def should_log(self):
+        """
+        Whether or not the current process should produce log.
+        """
+        return False  # TF Logging is handled by Keras not the Trainer
+
     @property
     def train_batch_size(self) -> int:
         """