onnx_diagnostic.helpers.torch_test_helper

onnx_diagnostic.helpers.torch_test_helper.dummy_llm(cls_name: str | None = None, dynamic_shapes: bool = False) → Tuple[Module, Tuple[Tensor, ...]] | Tuple[Module, Tuple[Tensor, ...], Any]

Creates a dummy LLM for test purposes.

Parameters:

• cls_name – None for whole model or a piece of it

• dynamic_shapes – returns dynamic shapes as well

<<<

from onnx_diagnostic.helpers.torch_test_helper import dummy_llm

print(dummy_llm())

>>>

    (LLM(
      (embedding): Embedding(
        (embedding): Embedding(1024, 16)
        (pe): Embedding(1024, 16)
      )
      (decoder): DecoderLayer(
        (attention): MultiAttentionBlock(
          (attention): ModuleList(
            (0-1): 2 x AttentionBlock(
              (query): Linear(in_features=16, out_features=16, bias=False)
              (key): Linear(in_features=16, out_features=16, bias=False)
              (value): Linear(in_features=16, out_features=16, bias=False)
            )
          )
          (linear): Linear(in_features=32, out_features=16, bias=True)
        )
        (feed_forward): FeedForward(
          (linear_1): Linear(in_features=16, out_features=128, bias=True)
          (relu): ReLU()
          (linear_2): Linear(in_features=128, out_features=16, bias=True)
        )
        (norm_1): LayerNorm((16,), eps=1e-05, elementwise_affine=True)
        (norm_2): LayerNorm((16,), eps=1e-05, elementwise_affine=True)
      )
    ), (tensor([[ 988,  119,  833,  922,  526,  410,  311,  323,  623,  292,  745,  907,
              952,  358,  999,  332,  785,  219,  672,  233,  828,  255,   93,  377,
              480,  961,  488, 1000,   42,   76]]),))

onnx_diagnostic.helpers.torch_test_helper.is_torchdynamo_exporting() → bool

Tells if torch is exporting a model.

onnx_diagnostic.helpers.torch_test_helper.replace_string_by_dynamic(dynamic_shapes: Any) → Any

Replaces strings by torch.export.Dim.DYNAMIC.

onnx_diagnostic.helpers.torch_test_helper.steal_forward(model: Module, with_shape: bool = True, with_min_max: bool = False)

The necessary modification to steem forward method and prints out inputs and outputs. See example Steel method forward to guess the dynamic shapes (with Tiny-LLM).

onnx_diagnostic.helpers.torch_test_helper.to_any(value: Any, to_value: dtype | device) → Any

Applies torch.to is applicable. Goes recursively.

onnx_diagnostic.helpers.torch_test_helper.to_numpy(tensor: Tensor)

Converts a torch tensor to numy.

onnx_diagnostic.helpers.torch_test_helper.torch_deepcopy(value: Any) → Any

Makes a deepcopy.