Fix SDPA tests (#28552)

* skip bf16 test if not supported by device

* fix

* fix bis

* use is_torch_bf16_available_on_device

* use is_torch_fp16_available_on_device

* fix & use public llama

* use 1b model

* fix flacky test

---------
Co-authored-by: Your Name <you@example.com>

tests/models/llama/test_modeling_llama.py

@@ -457,10 +457,10 @@ class LlamaModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMixi
         """
         max_new_tokens = 30
 
-        tokenizer = LlamaTokenizer.from_pretrained("meta-llama/Llama-2-7b-hf")
+        tokenizer = LlamaTokenizer.from_pretrained("saibo/llama-1B")
 
         model_sdpa = LlamaForCausalLM.from_pretrained(
-            "meta-llama/Llama-2-7b-hf",
+            "saibo/llama-1B",
             torch_dtype=torch.float16,
             low_cpu_mem_usage=True,
         ).to(torch_device)
@@ -468,7 +468,7 @@ class LlamaModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMixi
         self.assertTrue(model_sdpa.config._attn_implementation == "sdpa")
 
         model_eager = LlamaForCausalLM.from_pretrained(
-            "meta-llama/Llama-2-7b-hf",
+            "saibo/llama-1B",
             torch_dtype=torch.float16,
             low_cpu_mem_usage=True,
             attn_implementation="eager",
@@ -488,7 +488,11 @@ class LlamaModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMixi
         if not has_sdpa:
             raise ValueError("The SDPA model should have SDPA attention layers")
 
-        texts = ["hi", "Hello this is a very long sentence my friend", "Today I am in Paris and"]
+        texts = [
+            "hi here's a longer context, getting longer and",
+            "Hello this is a very long sentence my friend, very long for real",
+            "Today I am in Paris and",
+        ]
 
         for padding_side in ["left", "right"]:
             tokenizer.padding_side = padding_side

tests/test_modeling_common.py

@@ -84,6 +84,8 @@ from transformers.utils import (
     is_accelerate_available,
     is_flax_available,
     is_tf_available,
+    is_torch_bf16_available_on_device,
+    is_torch_fp16_available_on_device,
     is_torch_fx_available,
     is_torch_sdpa_available,
 )
@@ -3382,8 +3384,13 @@ class ModelTesterMixin:
         if not self.all_model_classes[0]._supports_sdpa:
             self.skipTest(f"{self.all_model_classes[0].__name__} does not support SDPA")
 
-        if torch_device == "cpu" and torch_dtype == "float16":
-            self.skipTest("float16 not supported on cpu")
+        if torch_dtype == "float16" and not is_torch_fp16_available_on_device(torch_device):
+            self.skipTest(f"float16 not supported on {torch_device} (on the specific device currently used)")
+
+        if torch_dtype == "bfloat16" and not is_torch_bf16_available_on_device(torch_device):
+            self.skipTest(
+                f"bfloat16 not supported on {torch_device} (on the specific device currently used, e.g. Nvidia T4 GPU)"
+            )
 
         # Not sure whether it's fine to put torch.XXX in a decorator if torch is not available so hacking it here instead.
         if torch_dtype == "float16":
@@ -3400,7 +3407,7 @@ class ModelTesterMixin:
             ("cpu", True, torch.bfloat16): 1e-2,
             ("cuda", False, torch.float32): 1e-6,
             ("cuda", False, torch.bfloat16): 1e-2,
-            ("cuda", False, torch.float16): 1e-3,
+            ("cuda", False, torch.float16): 5e-3,
             ("cuda", True, torch.float32): 1e-6,
             ("cuda", True, torch.bfloat16): 1e-2,
             ("cuda", True, torch.float16): 5e-3,
@@ -3412,7 +3419,7 @@ class ModelTesterMixin:
             ("cpu", True, torch.bfloat16): 1e-2,
             ("cuda", False, torch.float32): 1e-4,
             ("cuda", False, torch.bfloat16): 1e-2,
-            ("cuda", False, torch.float16): 1e-3,
+            ("cuda", False, torch.float16): 5e-3,
             ("cuda", True, torch.float32): 1e-4,
             ("cuda", True, torch.bfloat16): 3e-2,
             ("cuda", True, torch.float16): 5e-3,