Add TF weights
#12
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=5.970e-03; Maximum crossload hidden layer difference=1.053e-01;
Maximum conversion output difference=5.970e-03; Maximum conversion hidden layer difference=1.053e-01;
CAUTION: The maximum admissible error was manually increased to 0.2!
Thanks!
ybelkada
changed pull request status to
merged
Hi, sorry to bump this thread.
Could you please give me some pointers on how to use the .h5 file with TF? It only contains weights so I understand the model structure needs to be created in code beforehand. What is the transformers module that the PyTorch example imports?