Changing Backend Tensor initialization (meta-pytorch#3484)

Summary:
X-link: pytorch/FBGEMM#5056

Pull Request resolved: meta-pytorch#3484

X-link: https://github.com/facebookresearch/FBGEMM/pull/2066

**Context:**
Currently, we are enabling SSD optimizer offloading for the ssd tbe kernel

**In this diff:**
We retrieve the newly added parameters from the tbe config and pass it down to the tbe

Differential Revision: D85353134

Author: Raahul Kalyaan Jakka

torchrec/distributed/batched_embedding_kernel.py

@@ -246,6 +246,12 @@ def _populate_ssd_tbe_params(config: GroupedEmbeddingConfig) -> Dict[str, Any]:
         ssd_tbe_params["kvzch_eviction_tbe_config"] = fused_params.get(
             "kvzch_eviction_tbe_config"
         )
+    if "enable_optimizer_offloading" in fused_params:
+        ssd_tbe_params["enable_optimizer_offloading"] = fused_params.get(
+            "enable_optimizer_offloading"
+        )
+    else:
+        ssd_tbe_params["enable_optimizer_offloading"] = False
 
     ssd_tbe_params["table_names"] = [table.name for table in config.embedding_tables]