Add TF weights
#1
by CCMat - opened
Model converted by the transformers' pt_to_tf CLI. All converted model outputs and hidden layers were validated against its Pytorch counterpart.
Maximum crossload output difference=1.135e-04; Maximum crossload hidden layer difference=2.536e-02;
Maximum conversion output difference=1.135e-04; Maximum conversion hidden layer difference=2.536e-02;
CAUTION: The maximum admissible error was manually increased to 0.03!
@joaogante , @nielsr , @sgugger
The max error was increased due to batch normalization creating differences that get amplified through the forward pass.
This is the corresponding github PR : https://github.com/huggingface/transformers/pull/18597
joaogante changed pull request status to merged