diff --git a/tensorrt_llm/_torch/autotuner.py b/tensorrt_llm/_torch/autotuner.py index f68d9c7f6c..89fb89a054 100644 --- a/tensorrt_llm/_torch/autotuner.py +++ b/tensorrt_llm/_torch/autotuner.py @@ -505,9 +505,8 @@ class AutoTunerProfilingCache: fcntl.flock(f, fcntl.LOCK_SH) current_cache_contents = json.load(f) self._deserialize_metadata(current_cache_contents["metadata"]) - assert f"rank_{rank}" in current_cache_contents, f"Rank {rank} cache not found in {file_path}" self.cache = self._deserialize_cache_data( - current_cache_contents[f'rank_{rank}']) + current_cache_contents.get(f'rank_{rank}', {})) logger.info( f"[AutoTuner] Successfully loaded cache from {file_path} using JSON format" ) diff --git a/tests/integration/test_lists/waives.txt b/tests/integration/test_lists/waives.txt index b98d108a53..8fb581a1ed 100644 --- a/tests/integration/test_lists/waives.txt +++ b/tests/integration/test_lists/waives.txt @@ -356,7 +356,6 @@ accuracy/test_llm_api_pytorch.py::TestNemotronV3Super::test_auto_dtype_4gpus[4-4 accuracy/test_llm_api_pytorch.py::TestDeepSeekR1::test_nvfp4_multi_gpus[latency] SKIP (https://nvbugs/5814309) accuracy/test_llm_api_pytorch.py::TestDeepSeekV3Lite::test_bfloat16_4gpus[tp4-mtp_nextn=2-attention_dp=False-cuda_graph=False-overlap_scheduler=False-torch_compile=True] SKIP (https://nvbugs/5800646) unittest/_torch/thop/parallel/test_fp4_swizzle.py::test_swizzle_sf SKIP (https://nvbugs/5811159) -unittest/_torch/misc/test_autotuner.py::test_autotuner_distributed_strategy[2-DistributedTuningStrategy.BROADCAST] SKIP (https://nvbugs/5814253) stress_test/stress_test.py::test_run_stress_test[llama-v3-8b-instruct-hf_tp1-stress_time_300s_timeout_450s-MAX_UTILIZATION-pytorch-stress-test] SKIP (https://nvbugs/5814203) unittest/_torch/attention/test_trtllm_flashinfer_symbol_collision.py::test_flashinfer_fused_moe_matches_torch_moe SKIP (https://nvbugs/5814215) full:sm89/accuracy/test_llm_api_pytorch.py::TestNemotronV3Nano::test_auto_dtype SKIP (https://nvbugs/5814504)