Qwen3 causal lm (keras-team#2311)

* init

* update

* bug fixes

* add qwen causal lm test

* fix qwen3 tests

Author: kanpuriyanawab

keras_hub/api/models/__init__.py

@@ -322,7 +322,19 @@
 from keras_hub.src.models.qwen.qwen_tokenizer import (
     QwenTokenizer as Qwen2Tokenizer,
 )
+<<<<<<< HEAD
 from keras_hub.src.models.qwen3.qwen3_backbone import Qwen3Backbone
+=======
+from keras_hub.src.models.qwen.qwen_tokenizer import (
+    QwenTokenizer as QwenTokenizer,
+)
+from keras_hub.src.models.qwen3.qwen3_backbone import (
+    Qwen3Backbone as Qwen3Backbone,
+)
+from keras_hub.src.models.qwen3.qwen3_causal_lm import (
+    Qwen3CausalLM as Qwen3CausalLM,
+)
+>>>>>>> 5fd34b15 (Qwen3 causal lm (#2311))
 from keras_hub.src.models.qwen3.qwen3_causal_lm_preprocessor import (
     Qwen3CausalLMPreprocessor,
 )

keras_hub/src/models/qwen3/qwen3_attention.py

@@ -303,7 +303,7 @@ def _compute_attention(
             attention_mask = self._mask_sliding_window(
                 attention_mask,
                 cache_update_index=cache_update_index
-                if cache_update_index
+                if cache_update_index is not None
                 else 0,
             )
         attention_scores = self._masked_softmax(