[UnitTest] adding more properties testing (#1306)

Co-authored-by: Somasundaram <sindhuso@5c52309d1b09.ant.amazon.com>

engines/python/setup/djl_python/tests/test_properties_manager.py

@@ -101,6 +101,38 @@ def test_tnx_configs(self):
         self.assertEqual(tnx_configs.compiled_graph_path,
                          str(properties['compiled_graph_path']))
 
+    def test_tnx_all_configs(self):
+        # TODO: Replace with actual example of context_length_estimate
+
+        properties = {
+            "n_positions":
+            "2048",
+            "load_split_model":
+            "true",
+            "load_in_8bit":
+            "true",
+            "compiled_graph_path":
+            "s3://test/bucket/folder",
+            "low_cpu_mem_usage":
+            "true",
+            'context_length_estimate':
+            '{"context_length": "128", "variable_size": "12"}'
+        }
+        tnx_configs = TransformerNeuronXProperties(**common_properties,
+                                                   **properties)
+        self.assertEqual(tnx_configs.n_positions, 2048)
+        self.assertEqual(tnx_configs.compiled_graph_path,
+                         properties['compiled_graph_path'])
+
+        self.assertTrue(tnx_configs.load_split_model)
+        self.assertTrue(tnx_configs.load_in_8bit)
+        self.assertTrue(tnx_configs.low_cpu_mem_usage)
+
+        self.assertDictEqual(tnx_configs.context_length_estimate, {
+            'context_length': '128',
+            'variable_size': '12'
+        })
+
     def test_tnx_configs_error_case(self):
         properties = {
             "n_positions": "256",
@@ -273,15 +305,43 @@ def test_ds_invalid_dtype():
         test_ds_invalid_dtype()
 
     def test_hf_configs(self):
+        properties = {
+            "model_id": "model_id",
+            "model_dir": "model_dir",
+            "low_cpu_mem_usage": "true",
+            "disable_flash_attn": "false",
+            "engine": "MPI",
+        }
+
+        hf_configs = HuggingFaceProperties(**properties)
+        self.assertIsNone(hf_configs.load_in_8bit)
+        self.assertIsNone(hf_configs.device)
+        self.assertTrue(hf_configs.low_cpu_mem_usage)
+        self.assertFalse(hf_configs.disable_flash_attn)
+        self.assertIsNone(hf_configs.device_map)
+        self.assertTrue(hf_configs.is_mpi)
+        self.assertDictEqual(hf_configs.kwargs, {
+            'trust_remote_code': False,
+            "low_cpu_mem_usage": True,
+        })
+
+    def test_hf_all_configs(self):
         properties = {
             "model_id": "model_id",
             "model_dir": "model_dir",
             "tensor_parallel_degree": "4",
+            "load_in_4bit": "false",
             "load_in_8bit": "true",
-            "low_cpu_mem_usage": "True",
+            "low_cpu_mem_usage": "true",
             "disable_flash_attn": "false",
             "engine": "MPI",
-            "device_map": "auto"
+            "device_map": "cpu",
+            "quantize": "bitsandbytes8",
+            "output_formatter": "jsonlines",
+            "waiting_steps": '12',
+            "trust_remote_code": "true",
+            "rolling_batch": "auto",
+            "dtype": "bf16"
         }
 
         hf_configs = HuggingFaceProperties(**properties)
@@ -292,10 +352,13 @@ def test_hf_configs(self):
         self.assertTrue(hf_configs.is_mpi)
         self.assertDictEqual(
             hf_configs.kwargs, {
-                'trust_remote_code': False,
+                'trust_remote_code': True,
                 "low_cpu_mem_usage": True,
-                "device_map": 'auto',
-                "load_in_8bit": True
+                "device_map": 'cpu',
+                "load_in_8bit": True,
+                "waiting_steps": 12,
+                "output_formatter": "jsonlines",
+                "torch_dtype": torch.bfloat16
             })
 
     def test_hf_quantize(self):