TestBufferAliasingSurvivesSaveLoad.test_buffer_kv_save_load RuntimeError: execute_engine (with implementation in <module 'torch._library.custom_ops' from 'T:\\conda_environment_28905825685\\lib\\site-packages\\torch\\_library\\custom_ops.py'>): The output of this custom operator (1) must not also be an input to this custom operator and (2) may not alias any inputs to this custom operator or other returns. The most common way to trigger this error is if we have y = custom_op(x) and y and x are the same Tensor. Please instead return a clone of the offending output tensor(s) (e.g. return x.clone()) or refactor the custom operator to not return y. To execute this test, run the following from the base repo dir: python test_aliased_io_serialization.py TestBufferAliasingSurvivesSaveLoad.test_buffer_kv_save_load This message can be suppressed by setting PYTORCH_PRINT_REPRO_ON_FAILURE=0

TestUserInputAliasingSurvivesSaveLoad.test_kv_cache_user_input_save_load RuntimeError: execute_engine (with implementation in <module 'torch._library.custom_ops' from 'T:\\conda_environment_28905825685\\lib\\site-packages\\torch\\_library\\custom_ops.py'>): The output of this custom operator (1) must not also be an input to this custom operator and (2) may not alias any inputs to this custom operator or other returns. The most common way to trigger this error is if we have y = custom_op(x) and y and x are the same Tensor. Please instead return a clone of the offending output tensor(s) (e.g. return x.clone()) or refactor the custom operator to not return y. To execute this test, run the following from the base repo dir: python test_aliased_io_serialization.py TestUserInputAliasingSurvivesSaveLoad.test_kv_cache_user_input_save_load This message can be suppressed by setting PYTORCH_PRINT_REPRO_ON_FAILURE=0

TestLiftMutableBuffersAPI.test_caller_threads_buffer_for_in_place_mutation RuntimeError: execute_engine (with implementation in <module 'torch._library.custom_ops' from 'T:\\conda_environment_28905825685\\lib\\site-packages\\torch\\_library\\custom_ops.py'>): The output of this custom operator (1) must not also be an input to this custom operator and (2) may not alias any inputs to this custom operator or other returns. The most common way to trigger this error is if we have y = custom_op(x) and y and x are the same Tensor. Please instead return a clone of the offending output tensor(s) (e.g. return x.clone()) or refactor the custom operator to not return y. To execute this test, run the following from the base repo dir: python test_lift_mutable_buffers_api.py TestLiftMutableBuffersAPI.test_caller_threads_buffer_for_in_place_mutation This message can be suppressed by setting PYTORCH_PRINT_REPRO_ON_FAILURE=0

TestIndexCopyKVFastPath.test_single_position_write_aliased RuntimeError: execute_engine (with implementation in <module 'torch._library.custom_ops' from 'T:\\conda_environment_28905825685\\lib\\site-packages\\torch\\_library\\custom_ops.py'>): The output of this custom operator (1) must not also be an input to this custom operator and (2) may not alias any inputs to this custom operator or other returns. The most common way to trigger this error is if we have y = custom_op(x) and y and x are the same Tensor. Please instead return a clone of the offending output tensor(s) (e.g. return x.clone()) or refactor the custom operator to not return y. To execute this test, run the following from the base repo dir: python test_index_copy_kv.py TestIndexCopyKVFastPath.test_single_position_write_aliased This message can be suppressed by setting PYTORCH_PRINT_REPRO_ON_FAILURE=0

TestBufferBackedKVCache.test_paired_buffer_caches RuntimeError: execute_engine (with implementation in <module 'torch._library.custom_ops' from 'T:\\conda_environment_28905825685\\lib\\site-packages\\torch\\_library\\custom_ops.py'>): The output of this custom operator (1) must not also be an input to this custom operator and (2) may not alias any inputs to this custom operator or other returns. The most common way to trigger this error is if we have y = custom_op(x) and y and x are the same Tensor. Please instead return a clone of the offending output tensor(s) (e.g. return x.clone()) or refactor the custom operator to not return y. To execute this test, run the following from the base repo dir: python test_aliased_io.py TestBufferBackedKVCache.test_paired_buffer_caches This message can be suppressed by setting PYTORCH_PRINT_REPRO_ON_FAILURE=0

TestBufferBackedKVCache.test_buffer_streaming_persists_across_calls RuntimeError: execute_engine (with implementation in <module 'torch._library.custom_ops' from 'T:\\conda_environment_28905825685\\lib\\site-packages\\torch\\_library\\custom_ops.py'>): The output of this custom operator (1) must not also be an input to this custom operator and (2) may not alias any inputs to this custom operator or other returns. The most common way to trigger this error is if we have y = custom_op(x) and y and x are the same Tensor. Please instead return a clone of the offending output tensor(s) (e.g. return x.clone()) or refactor the custom operator to not return y. To execute this test, run the following from the base repo dir: python test_aliased_io.py TestBufferBackedKVCache.test_buffer_streaming_persists_across_calls This message can be suppressed by setting PYTORCH_PRINT_REPRO_ON_FAILURE=0

TestBufferBackedKVCache.test_buffer_mutation_in_place RuntimeError: execute_engine (with implementation in <module 'torch._library.custom_ops' from 'T:\\conda_environment_28905825685\\lib\\site-packages\\torch\\_library\\custom_ops.py'>): The output of this custom operator (1) must not also be an input to this custom operator and (2) may not alias any inputs to this custom operator or other returns. The most common way to trigger this error is if we have y = custom_op(x) and y and x are the same Tensor. Please instead return a clone of the offending output tensor(s) (e.g. return x.clone()) or refactor the custom operator to not return y. To execute this test, run the following from the base repo dir: python test_aliased_io.py TestBufferBackedKVCache.test_buffer_mutation_in_place This message can be suppressed by setting PYTORCH_PRINT_REPRO_ON_FAILURE=0

TestUserInputKVCache.test_streaming_state_accumulates RuntimeError: execute_engine (with implementation in <module 'torch._library.custom_ops' from 'T:\\conda_environment_28905825685\\lib\\site-packages\\torch\\_library\\custom_ops.py'>): The output of this custom operator (1) must not also be an input to this custom operator and (2) may not alias any inputs to this custom operator or other returns. The most common way to trigger this error is if we have y = custom_op(x) and y and x are the same Tensor. Please instead return a clone of the offending output tensor(s) (e.g. return x.clone()) or refactor the custom operator to not return y. To execute this test, run the following from the base repo dir: python test_aliased_io.py TestUserInputKVCache.test_streaming_state_accumulates This message can be suppressed by setting PYTORCH_PRINT_REPRO_ON_FAILURE=0

TestUserInputKVCache.test_single_slice_write_in_place RuntimeError: execute_engine (with implementation in <module 'torch._library.custom_ops' from 'T:\\conda_environment_28905825685\\lib\\site-packages\\torch\\_library\\custom_ops.py'>): The output of this custom operator (1) must not also be an input to this custom operator and (2) may not alias any inputs to this custom operator or other returns. The most common way to trigger this error is if we have y = custom_op(x) and y and x are the same Tensor. Please instead return a clone of the offending output tensor(s) (e.g. return x.clone()) or refactor the custom operator to not return y. To execute this test, run the following from the base repo dir: python test_aliased_io.py TestUserInputKVCache.test_single_slice_write_in_place This message can be suppressed by setting PYTORCH_PRINT_REPRO_ON_FAILURE=0

TestUserInputKVCache.test_paired_kv_caches RuntimeError: execute_engine (with implementation in <module 'torch._library.custom_ops' from 'T:\\conda_environment_28905825685\\lib\\site-packages\\torch\\_library\\custom_ops.py'>): The output of this custom operator (1) must not also be an input to this custom operator and (2) may not alias any inputs to this custom operator or other returns. The most common way to trigger this error is if we have y = custom_op(x) and y and x are the same Tensor. Please instead return a clone of the offending output tensor(s) (e.g. return x.clone()) or refactor the custom operator to not return y. To execute this test, run the following from the base repo dir: python test_aliased_io.py TestUserInputKVCache.test_paired_kv_caches This message can be suppressed by setting PYTORCH_PRINT_REPRO_ON_FAILURE=0

Node.js 20 is deprecated. The following actions target Node.js 20 but are being forced to run on Node.js 24: ./test-infra/.github/actions/setup-ssh, actions/checkout@v4, pmeier/pytest-results-action@v0.5.0. For more information see: https://github.blog/changelog/2025-09-19-deprecation-of-node-20-on-github-actions-runners/

`torch.jit.script_method` is deprecated. Please switch to `torch.compile` or `torch.export`.

`torch.jit.script_method` is deprecated. Please switch to `torch.compile` or `torch.export`.

`torch.jit.script_method` is deprecated. Please switch to `torch.compile` or `torch.export`.

`torch.jit.script_method` is deprecated. Please switch to `torch.compile` or `torch.export`.

`torch.jit.script_method` is deprecated. Please switch to `torch.compile` or `torch.export`.

`torch.jit.script_method` is deprecated. Please switch to `torch.compile` or `torch.export`.

`torch.jit.script` is deprecated. Please switch to `torch.compile` or `torch.export`.

Your device is potentially capable to use triton kernel to speed up quantization simulations. Try to install triton with `pip install triton`.

builtin type SwigPyObject has no __module__ attribute