A pivot-model abstraction method enables automatic migration of neural network implementations between frameworks such as PyTorch and TensorFlow while preserving functional equivalence.
Title resolution pending
1 Pith paper cite this work, alongside 1 external citations. Polarity classification is still indexing.
1
Pith paper citing it
1
external citations · OpenAlex
fields
cs.LG 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Towards Migrating Neural Network Implementations
A pivot-model abstraction method enables automatic migration of neural network implementations between frameworks such as PyTorch and TensorFlow while preserving functional equivalence.