Revert "pin nightly to 2.5.0.dev20240709+cu121 (#505)"

This reverts commit cc871c5.

Author: jerryzh168

.github/workflows/regression_test.yml

@@ -33,7 +33,7 @@ jobs:
             gpu-arch-version: "12.1"
           - name: CUDA Nightly
             runs-on: linux.g5.12xlarge.nvidia.gpu
-            torch-spec: '--pre torch==2.5.0.dev20240709+cu121 --index-url https://download.pytorch.org/whl/nightly/cu121'
+            torch-spec: '--pre torch --index-url https://download.pytorch.org/whl/nightly/cu121'
             gpu-arch-type: "cuda"
             gpu-arch-version: "12.1"
           - name: CPU 2.2.2
@@ -48,7 +48,7 @@ jobs:
             gpu-arch-version: ""
           - name: CPU Nightly
             runs-on: linux.4xlarge
-            torch-spec: '--pre torch==2.5.0.dev20240709+cpu --index-url https://download.pytorch.org/whl/nightly/cpu'
+            torch-spec: '--pre torch --index-url https://download.pytorch.org/whl/nightly/cpu'
             gpu-arch-type: "cpu"
             gpu-arch-version: ""
 

test/integration/test_integration.py

@@ -631,6 +631,7 @@ def test_dequantize_int8_weight_only_quant_subclass(self, device, dtype):
 
     @parameterized.expand(COMMON_DEVICE_DTYPE)
     @unittest.skipIf(not TORCH_VERSION_AFTER_2_3, "int4 requires torch nightly.")
+    @unittest.skipIf(TORCH_VERSION_AFTER_2_5, "int4 skipping 2.5+ for now")
     def test_dequantize_int4_weight_only_quant_subclass(self, device, dtype):
         if dtype != torch.bfloat16:
             self.skipTest("Currently only supports bfloat16.")
@@ -641,6 +642,7 @@ def test_dequantize_int4_weight_only_quant_subclass(self, device, dtype):
 
     @parameterized.expand(COMMON_DEVICE_DTYPE)
     @unittest.skipIf(not TORCH_VERSION_AFTER_2_3, "int4 requires torch nightly.")
+    @unittest.skipIf(TORCH_VERSION_AFTER_2_5, "int4 skipping 2.5+ for now")
     def test_dequantize_int4_weight_only_quant_subclass_grouped(self, device, dtype):
         if dtype != torch.bfloat16:
             self.skipTest("Currently only supports bfloat16.")
@@ -821,6 +823,7 @@ def test_int8_weight_only_quant_with_freeze(self, device, dtype):
 
     @parameterized.expand(COMMON_DEVICE_DTYPE)
     @unittest.skipIf(not TORCH_VERSION_AFTER_2_3, "int4 requires torch nightly.")
+    @unittest.skipIf(TORCH_VERSION_AFTER_2_5, "int4 skipping 2.5+ for now")
     def test_int4_weight_only_quant_subclass_api(self, device, dtype):
         if dtype != torch.bfloat16:
             self.skipTest(f"Fails for {dtype}")
@@ -835,6 +838,7 @@ def test_int4_weight_only_quant_subclass_api(self, device, dtype):
 
     @parameterized.expand(COMMON_DEVICE_DTYPE)
     @unittest.skipIf(not TORCH_VERSION_AFTER_2_3, "int4 requires torch nightly.")
+    @unittest.skipIf(TORCH_VERSION_AFTER_2_5, "int4 skipping 2.5+ for now")
     def test_int4_weight_only_quant_subclass_api_grouped(self, device, dtype):
         if dtype != torch.bfloat16:
             self.skipTest(f"Fails for {dtype}")