Improves documentation

_doc/cmds/validate.rst

@@ -100,6 +100,7 @@ Let's export with ONNX this time and checks for discrepancies.
     python -m onnx_diagnostic validate -m arnir0/Tiny-LLM --run -v 1 --export onnx-dynamo -o dump_models --patch --opt ir
 
 .. runpython::
+    :process:
 
     from onnx_diagnostic._command_lines_parser import main
 
@@ -117,6 +118,7 @@ of function :func:`onnx_diagnostic.torch_models.validate.run_ort_fusion`.
     python -m onnx_diagnostic validate -m arnir0/Tiny-LLM --run -v 1 --export onnx-dynamo -o dump_models --patch --opt ir --ortfusiontype ALL
 
 .. runpython::
+    :process:
 
     from onnx_diagnostic._command_lines_parser import main
 

_doc/examples/plot_dump_intermediate_results.py

@@ -128,7 +128,8 @@
 # results to intermediate results in ONNX.
 # Let's create the ONNX model.
 
-epo = torch.onnx.export(model, inputs, dynamic_shapes=ds, dynamo=True)
+ep = torch.export.export(model, inputs, dynamic_shapes=ds)
+epo = torch.onnx.export(ep, dynamo=True)
 epo.optimize()
 epo.save("plot_dump_intermediate_results.onnx")
 

_doc/recipes/plot_export_dim1.py

@@ -54,9 +54,12 @@ def forward(self, x, y, z):
 # Same model, a dynamic dimension = 1 and backed_size_oblivious=True
 # ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 
-with torch.fx.experimental._config.patch(backed_size_oblivious=True):
-    ep = torch.export.export(model, (x, y, z), dynamic_shapes=(ds, ds, ds))
-    print(ep.graph)
+try:
+    with torch.fx.experimental._config.patch(backed_size_oblivious=True):
+        ep = torch.export.export(model, (x, y, z), dynamic_shapes=(ds, ds, ds))
+        print(ep.graph)
+except RuntimeError as e:
+    print("ERROR", e)
 
 # %%
 # It worked.

onnx_diagnostic/helpers/cache_helper.py

@@ -96,13 +96,16 @@ def flatten_unflatten_for_dynamic_shapes(
             return tuple(subtrees)
         if spec.type is list:
             return list(subtrees)
+        if spec.type is None and not subtrees:
+            return None
         if spec.context:
             # This is a custom class with attributes.
             # It is returned as a list.
             return list(subtrees)
         raise ValueError(
             f"Unable to interpret spec type {spec.type} "
-            f"(type is {type(spec.type)}, context is {spec.context})."
+            f"(type is {type(spec.type)}, context is {spec.context}), "
+            f"spec={spec}, subtrees={subtrees}"
         )
     # This is a list.
     return subtrees