dif

diff --git a/torch/_inductor/config.py b/torch/_inductor/config.py
index 7900ecb46dc..fb5f6bb8595 100644
--- a/torch/_inductor/config.py
+++ b/torch/_inductor/config.py
@@ -1291,7 +1291,7 @@ class auto_chunker:
     output_size_threshold = 1024 * 1024
 
     # Don't chunk from a node if it does not 'amplify' the inputs a lot
-    amplify_ratio_threshold = 8
+    amplify_ratio_threshold = 6
 
     num_chunk = (
         int(os.environ.get("TORCHINDUCTOR_CHUNKER_NUM_CHUNKS"))  # type: ignore[arg-type]
diff --git a/torch/_inductor/fx_passes/auto_chunker/core.py b/torch/_inductor/fx_passes/auto_chunker/core.py
index 647638578a7..739c8416b5a 100644
--- a/torch/_inductor/fx_passes/auto_chunker/core.py
+++ b/torch/_inductor/fx_passes/auto_chunker/core.py
@@ -125,19 +125,24 @@ def find_amplifier_node(graph: Graph) -> Optional[Node]:
 
     amplifier_nodes_ratio = []
 
+    import torch.distributed as dist
+    # dist.breakpoint() # TODO
     for node in graph.nodes:
         if use_tangent(node):
             # enter backward part of the graph
+            log.debug("Node %s uses tangent", node.format_node())
             break
 
         # Only trigger chunking for a small set of nodes like matmul for now
         if node.op != "call_function" or node.target not in eligible_amplifier_node:
+            log.debug("Node %s not calling eligible function", node.format_node())
             continue
 
         input_size = compute_tensor_size(node.args, node.kwargs)
         output_size = compute_tensor_size(node)
 
         if input_size == 0:
+            log.debug("Node %s has 0 input", node.format_node())
             continue
 
         ratio = output_size / input_size
@@ -148,6 +153,7 @@ def find_amplifier_node(graph: Graph) -> Optional[Node]:
             amplifier_nodes_ratio.append((node, ratio))
         elif ratio >= 4 and output_size >= 64_000:
             log.debug("Node '%s' get skipped as amplifier_node due to small amplification ratio or size. ratio %s, size %s", node.format_node(), ratio, output_size)
+        log.debug("Node '%s' get skipped as amplifier_node due to small amplification ratio or size. ratio %s, size %s", node.format_node(), ratio, output_size)
 
     amplifier_nodes_ratio = sorted(
         amplifier_nodes_ratio, key=lambda x: x[1], reverse=True
diff --git a/torch/_inductor/fx_passes/auto_chunker/propagate_scale_by.py b/torch/_inductor/fx_passes/auto_chunker/propagate_scale_by.py
index 0937fd58f77..65f3cf2febe 100644
--- a/torch/_inductor/fx_passes/auto_chunker/propagate_scale_by.py
+++ b/torch/_inductor/fx_passes/auto_chunker/propagate_scale_by.py
@@ -89,9 +89,9 @@ def propagate_scale_by(nodes_with_chunking_meta: Sequence[Node]) -> None:
 def propagate_div(div_node: Node) -> bool:
     lhs_node, rhs_node = div_node.args[:2]
     assert isinstance(lhs_node, Node)
-    assert isinstance(rhs_node, Node)
+    assert isinstance(rhs_node, (Node, int, float))
     lhs_scale_by = get_scale_by_from_node(lhs_node)
-    rhs_scale_by = get_scale_by_from_node(rhs_node)
+    rhs_scale_by = get_scale_by_from_node(rhs_node) if isinstance(rhs_node, Node) else None
     if lhs_scale_by and rhs_scale_by is None:
         update_chunking_meta(div_node, scale_by=lhs_scale_by)
         return True