Add TF weights
#6
by
Rocketknight1
HF staff
- 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=2.766e-05; Maximum crossload hidden layer difference=3.219e-06;
Maximum conversion output difference=2.766e-05; Maximum conversion hidden layer difference=3.219e-06;
Hi
@prajjwal1
, this is an automated TF conversion of the model weights for the bert-tiny repo. All outputs have been tested to be equivalent to the PT version up to standard float32 error (~3e-6).