RecursionError: maximum recursion depth exceeded while calling a Python object. When compiling a transformer model with JIT about adapters

[TimoImhof]

Hi @Sardhendu, thanks for opening this issue.

The adapter-transformer package is deprecated and the new and actively maintained package we provide is the adapters package (see here or #584 for more information).

However, I tried to reproduce the error with the new package and it is still possible; here is my slimmed-down version to reproduce the error with adapters:

from transformers import  CLIPModel
import torch
import torch.nn as nn
from adapters import init

testmodel = CLIPModel.from_pretrained("openai/clip-vit-base-patch32")
testmodel.eval()
# enable adapter support
init(testmodel)

shape = [1, 3, 224, 224]
_random_example = torch.rand(shape, requires_grad=False, dtype=torch.float32)
x = torch.index_select(
    _random_example,
    dim=0,
    index=torch.Tensor([0]).to(dtype=torch.int32),
)
print(x)
_inputs = {"X": x}

traced_model = torch.jit.trace(
    testmodel, 
    example_inputs=(
        _inputs["X"], 
    )
)
torch.jit.save(traced_model, "jit_model.pth")

Since this problem still exists in the new package I will have a look into this and get back to you.
If you have gained any new insights since you posted this issue, please let me know.

Cheers,
Timo