tpu-inference 0.11.1.dev202511220812__py3-none-any.whl → 0.13.2.dev20251230__py3-none-any.whl

tests/{test_quantization.py → layers/jax/test_qwix.py}

@@ -11,9 +11,9 @@ from jax.sharding import Mesh, NamedSharding
 from jax.sharding import PartitionSpec as P
 from qwix._src.providers import ptq
 
-import tpu_inference.models.jax.utils.quantization.quantization_utils as quantize_qwix  # noqa: E402
+import tpu_inference.models.jax.utils.qwix.qwix_utils as quantize_qwix  # noqa: E402
 from tpu_inference.models.common.model_loader import apply_qwix_quantization
-from tpu_inference.models.jax.utils.quantization.quantization_utils import (
+from tpu_inference.models.jax.utils.qwix.qwix_utils import (
     DEFAULT_MAX_NUM_BLOCKS_PER_REQ, DEFAULT_MAX_NUM_SEQS_FOR_MODEL_INPUTS,
     DEFAULT_NUM_TOKENS_FOR_MODEL_INPUTS)
 
@@ -29,8 +29,7 @@ module_mocks = {
     'vllm.config': MagicMock(),
     'tpu_inference': MagicMock(),
     'tpu_inference.logger': MagicMock(init_logger=lambda name: MagicMock()),
-    'tpu_inference.models.jax.utils.quantization.quantization_utils':
-    MagicMock(),
+    'tpu_inference.models.jax.utils.qwix.qwix_utils': MagicMock(),
 }
 
 
@@ -112,6 +111,8 @@ class TestQwixQuantizeNnxModel(unittest.TestCase):
         self.mesh = Mesh(jax.devices(), ('model', ))
         self.rng = jax.random.PRNGKey(0)
         self.model = SimpleModel(rngs=nnx.Rngs(0))
+        self.model.vllm_config = MagicMock()
+        self.model.vllm_config.model_config.use_mla = False
 
         self.qwix_config = [
             {
@@ -131,18 +132,19 @@ class TestQwixQuantizeNnxModel(unittest.TestCase):
         """Test that qwix.quantize_model is called with the correct arguments."""
         quantized_model_mock = MagicMock(spec=nnx.Module)
         mock_quantize_model.return_value = quantized_model_mock
+        self.model.vllm_config.sharding_config.total_dp_size = 1
 
         with patch(
-                "tpu_inference.models.jax.utils.quantization.quantization_utils.init_logger",
+                "tpu_inference.models.jax.utils.qwix.qwix_utils.init_logger",
                 return_value=MagicMock()
         ), patch(
                 "tpu_inference.utils.hbm_usage_gb",
                 return_value=[(0.0, 0.0), (0.0, 0.0)]
         ), patch(
-                "tpu_inference.models.jax.utils.quantization.quantization_utils.create_kv_caches",
+                "tpu_inference.models.jax.utils.qwix.qwix_utils.create_kv_caches",
                 return_value=self.mock_kv_caches
         ), patch(
-                "tpu_inference.models.jax.utils.quantization.quantization_utils.quantization_config_file_path_to_dict",
+                "tpu_inference.models.jax.utils.qwix.qwix_utils.quantization_config_file_path_to_dict",
                 return_value=self.qwix_config):
             returned_model = quantize_qwix.qwix_quantize_nnx_model(
                 model=self.model,
@@ -317,10 +319,9 @@ class TestApplyQwixQuantizationLogic(unittest.TestCase):
         self.assertIs(result2, self.mock_model)
 
     @patch(
-        'tpu_inference.models.jax.utils.quantization.quantization_utils.qwix_quantize_nnx_model'
+        'tpu_inference.models.jax.utils.qwix.qwix_utils.qwix_quantize_nnx_model'
     )
-    @patch(
-        'tpu_inference.models.jax.utils.quantization.quantization_utils.utils')
+    @patch('tpu_inference.models.jax.utils.qwix.qwix_utils.utils')
     def test_apply_to_abstract_model(self, mock_utils, mock_quantize_func):
         """Test quantization is correctly applied to an abstract model factory."""
         mock_utils.get_padded_num_heads.return_value = 8
@@ -357,10 +358,9 @@ class TestApplyQwixQuantizationLogic(unittest.TestCase):
         self.assertIs(result_model, quantized_model)
 
     @patch(
-        'tpu_inference.models.jax.utils.quantization.quantization_utils.qwix_quantize_nnx_model'
+        'tpu_inference.models.jax.utils.qwix.qwix_utils.qwix_quantize_nnx_model'
     )
-    @patch(
-        'tpu_inference.models.jax.utils.quantization.quantization_utils.utils')
+    @patch('tpu_inference.models.jax.utils.qwix.qwix_utils.utils')
     def test_apply_to_abstract_model_with_initialize_cache(
             self, mock_utils, mock_quantize_func):
         """Test abstract model quantization with 'initialize_cache' method."""
@@ -461,15 +461,13 @@ class TestLoadRandomWeightsIntoQwixAbstractModel(unittest.TestCase):
         # Mock model structure
         self.model = MagicMock(spec=['weight_loader', 'initialize_cache'])
         self.model.weight_loader = MagicMock(
-            spec=['scale_dtype', 'scale_shap_map_for_random_weight_loading'])
+            spec=['scale_dtype', 'scale_shape_map_for_random_weight_loading'])
         self.model.weight_loader.scale_dtype = jnp.float16
-        self.model.weight_loader.scale_shap_map_for_random_weight_loading = {}
+        self.model.weight_loader.scale_shape_map_for_random_weight_loading = {}
 
+    @patch('tpu_inference.models.jax.utils.qwix.qwix_utils.nnx.iter_graph')
     @patch(
-        'tpu_inference.models.jax.utils.quantization.quantization_utils.nnx.iter_graph'
-    )
-    @patch(
-        'tpu_inference.models.jax.utils.quantization.quantization_utils.get_random_sharded_array'
+        'tpu_inference.models.jax.utils.qwix.qwix_utils.get_random_sharded_array'
     )
     def test_successful_initialization(self, mock_get_random_array,
                                        mock_iter_graph):
@@ -482,6 +480,10 @@ class TestLoadRandomWeightsIntoQwixAbstractModel(unittest.TestCase):
         mock_random_array = jax.numpy.ones(1)
         mock_get_random_array.return_value = mock_random_array
 
+        self.model.weight_loader.scale_shape_map_for_random_weight_loading = {
+            'attention.wq': (1, 1)
+        }
+
         mock_iter_graph.return_value = [
             (('layers', '0', 'attention', 'wq', 'kernel'), mock_weight_param),
             (('layers', '0', 'attention', 'wq', 'array', 'scale'),
@@ -509,9 +511,7 @@ class TestLoadRandomWeightsIntoQwixAbstractModel(unittest.TestCase):
             quantize_qwix.load_random_weights_into_qwix_abstract_model(
                 self.rng, self.model, self.mesh, invalid_config)
 
-    @patch(
-        'tpu_inference.models.jax.utils.quantization.quantization_utils.nnx.iter_graph'
-    )
+    @patch('tpu_inference.models.jax.utils.qwix.qwix_utils.nnx.iter_graph')
     def test_param_shape_setting_no_scale_map(self, mock_iter_graph):
         """Test correct scale shape calculation when not in the map."""
         old_weight_param_val = jnp.empty((128, 64))
@@ -525,26 +525,11 @@ class TestLoadRandomWeightsIntoQwixAbstractModel(unittest.TestCase):
              mock_scale_var),
         ]
 
-        quantize_qwix.load_random_weights_into_qwix_abstract_model(
-            self.rng, self.model, self.mesh, self.quantization_config)
-
-        new_weight_param_val = mock_weight_param.value
-        new_scale_var_val = mock_scale_var.value
-
-        expected_scale_shape = (128 // 64, 64 // 64)
-        actual_scale_shape = new_scale_var_val.shape
-
-        expected_weight_shape = (128, 64)
-        actual_weight_shape = new_weight_param_val.shape
-
-        self.assertEqual(expected_scale_shape, actual_scale_shape)
-        self.assertEqual(expected_weight_shape, actual_weight_shape)
-        self.assertNotEqual(old_scale_var_val.shape, new_scale_var_val.shape)
-        assert jnp.not_equal(old_weight_param_val, new_weight_param_val).all()
+        with self.assertRaises(ValueError):
+            quantize_qwix.load_random_weights_into_qwix_abstract_model(
+                self.rng, self.model, self.mesh, self.quantization_config)
 
-    @patch(
-        'tpu_inference.models.jax.utils.quantization.quantization_utils.nnx.iter_graph'
-    )
+    @patch('tpu_inference.models.jax.utils.qwix.qwix_utils.nnx.iter_graph')
     def test_param_shape_setting_with_scale_map(self, mock_iter_graph):
         """Test correct scale shape calculation when in the map."""
         old_weight_param_val = jnp.empty((128, 64))
@@ -554,8 +539,8 @@ class TestLoadRandomWeightsIntoQwixAbstractModel(unittest.TestCase):
 
         expected_scale_shape = (55, 34)
 
-        self.model.weight_loader.scale_shap_map_for_random_weight_loading = {
-            'wq': expected_scale_shape
+        self.model.weight_loader.scale_shape_map_for_random_weight_loading = {
+            'attention.wq': expected_scale_shape
         }
 
         mock_iter_graph.return_value = [
@@ -604,9 +589,7 @@ class TestLoadRandomWeightsIntoQwixAbstractModel(unittest.TestCase):
         mock_randint.assert_not_called()
         mock_normal.assert_called_once()
 
-    @patch(
-        "tpu_inference.models.jax.utils.quantization.quantization_utils.logger.warning"
-    )
+    @patch("tpu_inference.models.jax.utils.qwix.qwix_utils.logger.warning")
     @patch("jax.make_array_from_callback")
     def test_get_random_sharded_array_sharding_fallback(
             self, mock_make_array, mock_logger_warning):
@@ -648,7 +631,7 @@ class TestManualQwixQuantization(unittest.TestCase):
         self.calibration_method = 'max'
 
     @patch(
-        'tpu_inference.models.jax.utils.quantization.quantization_utils.ptq.create_quantized_param'
+        'tpu_inference.models.jax.utils.qwix.qwix_utils.ptq.create_quantized_param'
     )
     def test_manually_quantize_qwix_weight(self, mock_create_param):
         """Test that manually_quantize_qwix_weight calls ptq.create_quantized_param correctly."""
@@ -672,9 +655,7 @@ class TestManualQwixQuantization(unittest.TestCase):
         self.assertEqual(passed_how_to_quantize.calibration_method,
                          self.calibration_method)
 
-    @patch(
-        'tpu_inference.models.jax.utils.quantization.quantization_utils.ptq.quantize_act'
-    )
+    @patch('tpu_inference.models.jax.utils.qwix.qwix_utils.ptq.quantize_act')
     @patch('qwix.pallas.get_current_rule')
     def test_manually_quantize_qwix_activation(self, mock_get_rule,
                                                mock_quantize_act):
@@ -832,5 +813,157 @@ class TestGetQuantDtypeFromQwixConfig(unittest.TestCase):
         self.assertIsNone(quant_dtype)
 
 
+class TestGetDefaultQwixQuantizationConfig(unittest.TestCase):
+    """Tests for the get_default_qwix_quantization_config function."""
+
+    def setUp(self):
+        # Mocking the default configs that the function expects to find in the module
+        self.mock_deepseek_config = {
+            "qwix": {
+                "rules": [{
+                    "module_path": ".*",
+                    "tile_size": 0
+                }]
+            }
+        }
+        self.mock_llama_config = {"qwix": {"rules": [{"name": "llama_rule"}]}}
+        self.mock_gpt_oss_config = {"qwix": {"rules": [{"name": "gpt_rule"}]}}
+
+        # Patch the constants in the module where the function resides
+        self.patchers = [
+            patch(
+                "tpu_inference.models.jax.utils.qwix.qwix_utils.DEFAULT_DEEPSEEK_FP4_MLP_MOE_FP8_ATTN_CONFIG",
+                self.mock_deepseek_config),
+            patch(
+                "tpu_inference.models.jax.utils.qwix.qwix_utils.DEFAULT_LLAMA4_FP8_CONFIG",
+                self.mock_llama_config),
+            patch(
+                "tpu_inference.models.jax.utils.qwix.qwix_utils.DEFAULT_GPT_OSS_FP4_CONFIG",
+                self.mock_gpt_oss_config),
+            patch("tpu_inference.models.jax.utils.qwix.qwix_utils.logger",
+                  MagicMock())
+        ]
+        for p in self.patchers:
+            p.start()
+
+    def tearDown(self):
+        for p in self.patchers:
+            p.stop()
+
+    def test_skip_quantization_returns_none(self):
+        """Test that skip_quantization=True returns None immediately."""
+        result = quantize_qwix.get_default_qwix_quantization_config(
+            MagicMock(), True)
+        self.assertIsNone(result)
+
+    def test_unsupported_model_returns_none(self):
+        """Test that an unknown model type returns None."""
+        hf_config = MagicMock()
+        hf_config.model_type = "unknown_model"
+        result = quantize_qwix.get_default_qwix_quantization_config(
+            hf_config, False)
+        self.assertIsNone(result)
+
+    def test_deepseek_v3_success(self):
+        """Test DeepSeek V3 config with valid weight_block_size."""
+        hf_config = MagicMock()
+        hf_config.model_type = "DeepSeek_V3"
+        hf_config.quantization_config = {
+            "quant_method": "fp8",
+            "weight_block_size": [1, 128]
+        }
+
+        result = quantize_qwix.get_default_qwix_quantization_config(
+            hf_config, False)
+
+        # Check if tile_size was updated from 0 to 128
+        self.assertEqual(result["qwix"]["rules"][0]["tile_size"], 128)
+        # Ensure it's a deep copy (original mock shouldn't change)
+        self.assertEqual(
+            self.mock_deepseek_config["qwix"]["rules"][0]["tile_size"], 0)
+
+    def test_deepseek_v3_invalid_block_size(self):
+        """Test DeepSeek V3 raises ValueError on invalid block size format."""
+        hf_config = MagicMock()
+        hf_config.model_type = "deepseek_v3"
+        hf_config.quantization_config = {
+            "quant_method": "fp8",
+            "weight_block_size": [128]
+        }
+
+        with self.assertRaisesRegex(ValueError, "Invalid weight_block_size"):
+            quantize_qwix.get_default_qwix_quantization_config(
+                hf_config, False)
+
+    def test_deepseek_v3_invalid_block_size_2d_subchannel(self):
+        """Test DeepSeek V3 raises ValueError on invalid block size format."""
+        hf_config = MagicMock()
+        hf_config.model_type = "deepseek_v3"
+        hf_config.quantization_config = {
+            "quant_method": "fp8",
+            "weight_block_size": [512, 512]
+        }
+
+        with self.assertRaisesRegex(AssertionError,
+                                    "Expected first dimension to be 1"):
+            quantize_qwix.get_default_qwix_quantization_config(
+                hf_config, False)
+
+    def test_deepseek_v3_no_weight_block_size(self):
+        """Test DeepSeek V3 config with valid weight_block_size."""
+        hf_config = MagicMock()
+        hf_config.model_type = "DeepSeek_V3"
+        hf_config.quantization_config = {
+            "quant_method": "fp8",
+        }
+
+        with self.assertRaisesRegex(
+                AssertionError,
+                "Expected weight_block_size in quantization_config"):
+
+            quantize_qwix.get_default_qwix_quantization_config(
+                hf_config, False)
+
+    def test_deepseek_v3_tile_size_assertion(self):
+        """Test DeepSeek V3 raises AssertionError if tile_size is <= 1."""
+        hf_config = MagicMock()
+        hf_config.model_type = "deepseek_v3"
+        hf_config.quantization_config = {
+            "quant_method": "fp8",
+            "weight_block_size": [1, 1]
+        }
+
+        with self.assertRaises(AssertionError):
+            quantize_qwix.get_default_qwix_quantization_config(
+                hf_config, False)
+
+    def test_llama4_success(self):
+        """Test Llama 4 default config path."""
+        hf_config = MagicMock()
+        hf_config.model_type = "llama4"
+        hf_config.quantization_config = {"quant_method": "compressed-tensors"}
+
+        result = quantize_qwix.get_default_qwix_quantization_config(
+            hf_config, False)
+        self.assertEqual(result, self.mock_llama_config)
+
+    def test_gpt_oss_success(self):
+        """Test GPT-OSS default config path."""
+        hf_config = MagicMock()
+        hf_config.model_type = "gpt_oss"
+        hf_config.quantization_config = {"quant_method": "mxfp4"}
+
+        result = quantize_qwix.get_default_qwix_quantization_config(
+            hf_config, False)
+        self.assertEqual(result, self.mock_gpt_oss_config)
+
+    def test_missing_attributes_handled(self):
+        """Test that function handles hf_config objects missing model_type safely."""
+        hf_config = object()  # No attributes
+        result = quantize_qwix.get_default_qwix_quantization_config(
+            hf_config, False)
+        self.assertIsNone(result)
+
+
 if __name__ == '__main__':
     unittest.main()

tests/layers/jax/test_rope.py

@@ -0,0 +1,93 @@
+# Copyright 2025 Google LLC
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import jax
+from jax import numpy as jnp
+from jax._src import test_util as jtu
+from jax.sharding import Mesh
+
+from tpu_inference.layers.jax.rope import (DeepseekScalingRotaryEmbedding,
+                                           RotaryEmbedding)
+
+
+class RotaryEmbeddingTest(jtu.JaxTestCase):
+
+    def test_apply_rope(self):
+        head_dim = 2
+        rope_theta = 10000
+        original_max_position_embeddings = 2
+        rope = RotaryEmbedding(
+            rotary_dim=head_dim,
+            rope_theta=rope_theta,
+            original_max_position_embeddings=original_max_position_embeddings,
+            dtype=jnp.float32)
+        rope.initialize_cache()
+        self.assertTrue(
+            rope.sin_cos_cache.shape == (original_max_position_embeddings,
+                                         head_dim))
+        expected_sin_cos = jnp.array([[1, 0], [0.5403023, 0.841471]],
+                                     dtype=jnp.float32)
+        self.assertArraysAllClose(rope.sin_cos_cache, expected_sin_cos)
+
+        num_tokens = 2
+        num_heads = 1
+        positions = jnp.arange(num_tokens)
+        x = jnp.ones((num_tokens, num_heads, head_dim))
+        x_rope = rope.apply_rope(positions, x)
+        expected_x_rope = jnp.array([[[1, 1]], [[-0.30116874, 1.3817732]]],
+                                    dtype=jnp.float32)
+        self.assertTrue(x_rope.shape == x.shape)
+        self.assertArraysAllClose(x_rope, expected_x_rope)
+
+
+class DeepseekScalingRotaryEmbeddingTest(jtu.JaxTestCase):
+
+    def test_apply_rope(self):
+        head_dim = 2
+        rope_theta = 10000
+        original_max_position_embeddings = 1
+        scaling_factor = 2
+        devices = jax.devices()
+        mesh = Mesh(devices, ('data', ))
+
+        rope = DeepseekScalingRotaryEmbedding(
+            rotary_dim=head_dim,
+            rope_theta=rope_theta,
+            original_max_position_embeddings=original_max_position_embeddings,
+            scaling_factor=scaling_factor,
+            dtype=jnp.float32)
+        rope.initialize_cache(mesh)
+        expected_padded_dim = 128
+        self.assertTrue(
+            rope.sin_cos_cache.shape == (scaling_factor *
+                                         original_max_position_embeddings,
+                                         expected_padded_dim))
+
+        valid_cache_slice = rope.sin_cos_cache[:, :head_dim]
+
+        expected_sin_cos = jnp.array([[1.0693147, 0], [0.5777532, 0.8997973]],
+                                     dtype=jnp.float32)
+
+        self.assertArraysAllClose(valid_cache_slice, expected_sin_cos)
+
+        num_tokens = 2
+        num_heads = 1
+        positions = jnp.arange(num_tokens)
+        x = jnp.ones((num_tokens, num_heads, head_dim))
+        x_rope = rope.apply_rope(positions, x)
+        expected_x_rope = jnp.array(
+            [[[1.0693147, 1.0693147]], [[-0.32204413, 1.4775505]]],
+            dtype=jnp.float32)
+        self.assertTrue(x_rope.shape == x.shape)
+        self.assertArraysAllClose(x_rope, expected_x_rope)

tests/layers/jax/test_sharding.py

@@ -0,0 +1,159 @@
+# Copyright 2025 Google LLC
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import unittest
+from unittest.mock import MagicMock
+
+import jax
+
+from tpu_inference.layers.common.sharding import (Sharding, ShardingConfig,
+                                                  ShardingRulesConfig,
+                                                  ShardingStrategy)
+
+
+class TestSharding(unittest.TestCase):
+    """Unit test suite for the sharding configuration logic."""
+
+    def setUp(self):
+        """Sets up the testing environment before each test."""
+
+        self.mock_devices = [MagicMock(coords=i) for i in range(8)]
+        self.original_jax_devices = jax.devices
+        jax.devices = lambda: self.mock_devices
+
+    def tearDown(self):
+        """Restores the original jax.devices function after tests."""
+        jax.devices = self.original_jax_devices
+
+    def test_sharding_strategy_init(self):
+        """Tests the initialization of the ShardingStrategy."""
+        strategy = ShardingStrategy(
+            tensor_parallelism=2,
+            expert_parallelism=4,
+            data_parallelism=1,
+            sequence_parallelism=1,
+        )
+        self.assertEqual(strategy.tensor_parallelism, 2)
+        self.assertEqual(strategy.expert_parallelism, 4)
+
+    def test_sharding_config_init(self):
+        """Tests the initialization of ShardingConfig."""
+        config = ShardingConfig()
+        self.assertIsInstance(config.prefill_rules, ShardingRulesConfig)
+        self.assertIsInstance(config.generate_rules, ShardingRulesConfig)
+
+        custom_rules = ShardingRulesConfig(activation_ffw_td=("model", None))
+        config_with_rules = ShardingConfig(prefill_rules=custom_rules)
+        self.assertEqual(config_with_rules.prefill_rules.activation_ffw_td,
+                         ("model", None))
+
+    def test_apply_overrides(self):
+        """Tests the _apply_overrides method for valid and invalid keys."""
+        sharding = Sharding(
+            prefill_rules={},
+            generate_rules={},
+        )
+        config_obj = ShardingRulesConfig()
+
+        valid_overrides = {"activation_ffw_td": ("model", None)}
+        sharding._apply_overrides(config_obj, valid_overrides)
+        self.assertEqual(config_obj.activation_ffw_td, ("model", None))
+
+        invalid_overrides = {"non_existent_attribute": (None, "model")}
+        with self.assertRaises(AttributeError):
+            sharding._apply_overrides(config_obj, invalid_overrides)
+
+    def test_default_sharding_config(self):
+        """Tests that default sharding rules are created correctly."""
+        sharding = Sharding(
+            prefill_rules={},
+            generate_rules={},
+        )
+
+        sharding_cfg = sharding.get_sharding_cfg()
+        generate_rules = sharding_cfg.generate_rules
+
+        self.assertEqual(generate_rules.ffw_weight_df, (None, "model"))
+        self.assertEqual(generate_rules.moe_router_de, (None, "model"))
+        self.assertEqual(generate_rules.attn_q_weight_dnh,
+                         (None, "model", None))
+
+    def test_sharding_init_with_overrides(self):
+        """Tests Sharding initialization with programmatic overrides."""
+        generate_overrides = {"logits_tv": ("data", "model")}
+
+        sharding = Sharding(
+            generate_rules=generate_overrides,
+            prefill_rules={},
+        )
+
+        sharding_cfg = sharding.get_sharding_cfg()
+        self.assertNotEqual(sharding_cfg.generate_rules.logits_tv,
+                            (None, "model"))
+        self.assertEqual(sharding_cfg.generate_rules.logits_tv,
+                         ("data", "model"))
+
+    def test_get_overrides_from_vllm_config(self):
+        """Tests fetching sharding overrides from a mock VllmConfig."""
+
+        mock_vllm_config_prefill = MagicMock()
+        mock_vllm_config_prefill.additional_config = {
+            "sharding": {
+                "logical_rules": {
+                    "all": {
+                        "norm_scale": ("model", )
+                    },
+                    "prefill": {
+                        "activation_ffw_td": ("data", "model")
+                    },
+                }
+            }
+        }
+        sharding_prefill = Sharding(
+            vllm_config=mock_vllm_config_prefill,
+            prefill_rules={},
+            generate_rules={},
+        )
+        prefill_overrides = sharding_prefill._get_overrides("prefill")
+
+        self.assertEqual(prefill_overrides["norm_scale"], ("model", ))
+        self.assertEqual(prefill_overrides["activation_ffw_td"],
+                         ("data", "model"))
+
+        mock_vllm_config_generate = MagicMock()
+        mock_vllm_config_generate.additional_config = {
+            "sharding": {
+                "logical_rules": {
+                    "all": {
+                        "norm_scale": ("model", )
+                    },
+                    "prefill": {
+                        "activation_ffw_td": ("data", "model")
+                    },
+                }
+            }
+        }
+        sharding_generate = Sharding(
+            vllm_config=mock_vllm_config_generate,
+            prefill_rules={},
+            generate_rules={},
+        )
+        generate_overrides = sharding_generate._get_overrides("generate")
+
+        self.assertEqual(generate_overrides["norm_scale"], ("model", ))
+        self.assertNotIn("activation_ffw_td", generate_overrides)
+
+
+if __name__ == "__main__":
+    unittest.main()