Device agnostic testing (#25870)

* adds agnostic decorators and availability fns

* renaming decorators and fixing imports

* updating some representative example tests
bloom, opt, and reformer for now

* wip device agnostic functions

* lru cache to device checking functions

* adds `TRANSFORMERS_TEST_DEVICE_SPEC`
if present, imports the target file and updates device to function
mappings

* comments `TRANSFORMERS_TEST_DEVICE_SPEC` code

* extra checks on device name

* `make style; make quality`

* updates default functions for agnostic calls

* applies suggestions from review

* adds `is_torch_available` guard

* Add spec file to docs, rename function dispatch names to backend_*

* add backend import to docs example for spec file

* change instances of  to

* Move register backend to before device check as per @statelesshz changes

* make style

* make opt test require fp16 to run

---------

Co-authored-by: arsalanu <arsalanu@graphcore.ai>
Co-authored-by: arsalanu <hzji210@gmail.com>

tests/models/opt/test_modeling_opt.py

@@ -22,7 +22,7 @@ import unittest
 import timeout_decorator  # noqa
 
 from transformers import OPTConfig, is_torch_available
-from transformers.testing_utils import require_torch, require_torch_gpu, slow, torch_device
+from transformers.testing_utils import require_torch, require_torch_fp16, require_torch_gpu, slow, torch_device
 
 from ...generation.test_utils import GenerationTesterMixin
 from ...test_configuration_common import ConfigTester
@@ -286,13 +286,13 @@ class OPTModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMixin,
             with torch.no_grad():
                 model(**inputs)[0]
 
+    @require_torch_fp16
     def test_generate_fp16(self):
         config, input_dict = self.model_tester.prepare_config_and_inputs()
         input_ids = input_dict["input_ids"]
         attention_mask = input_ids.ne(1).to(torch_device)
         model = OPTForCausalLM(config).eval().to(torch_device)
-        if torch_device == "cuda":
-            model.half()
+        model.half()
         model.generate(input_ids, attention_mask=attention_mask)
         model.generate(num_beams=4, do_sample=True, early_stopping=False, num_return_sequences=3)