diff --git "a/healed/stableopd_cold_keep50/vllm_server.log" "b/healed/stableopd_cold_keep50/vllm_server.log" new file mode 100644--- /dev/null +++ "b/healed/stableopd_cold_keep50/vllm_server.log" @@ -0,0 +1,8665 @@ +Skipping import of cpp extensions due to incompatible torch version 2.10.0+cu128 for torchao version 0.15.0 Please see https://github.com/pytorch/ao/issues/2919 for more info +WARNING 08-01 14:11:24 [registry.py:915] Model architecture OlmoeForCausalLM is already registered, and will be overwritten by the new model class glean_vllm.pruned_olmoe:PrunedOlmoeForCausalLM. +(APIServer pid=41276) INFO 08-01 14:11:24 [utils.py:299] +(APIServer pid=41276) INFO 08-01 14:11:24 [utils.py:299] █ █ █▄ ▄█ +(APIServer pid=41276) INFO 08-01 14:11:24 [utils.py:299] ▄▄ ▄█ █ █ █ ▀▄▀ █ version 0.19.0 +(APIServer pid=41276) INFO 08-01 14:11:24 [utils.py:299] █▄█▀ █ █ █ █ model outputs/pruned/glean-0125inst-math-keep50 +(APIServer pid=41276) INFO 08-01 14:11:24 [utils.py:299] ▀▀ ▀▀▀▀▀ ▀▀▀▀▀ ▀ ▀ +(APIServer pid=41276) INFO 08-01 14:11:24 [utils.py:299] +(APIServer pid=41276) INFO 08-01 14:11:24 [utils.py:233] non-default args: {'model_tag': 'outputs/pruned/glean-0125inst-math-keep50', 'host': '127.0.0.1', 'port': 8377, 'model': 'outputs/pruned/glean-0125inst-math-keep50', 'max_model_len': 3584, 'enforce_eager': True, 'served_model_name': ['student'], 'gpu_memory_utilization': 0.85} +(APIServer pid=41276) INFO 08-01 14:11:32 [model.py:549] Resolved architecture: OlmoeForCausalLM +(APIServer pid=41276) INFO 08-01 14:11:32 [model.py:1678] Using max model len 3584 +(APIServer pid=41276) INFO 08-01 14:11:33 [vllm.py:790] Asynchronous scheduling is enabled. +(APIServer pid=41276) WARNING 08-01 14:11:33 [vllm.py:848] Enforce eager set, disabling torch.compile and CUDAGraphs. This is equivalent to setting -cc.mode=none -cc.cudagraph_mode=none +(APIServer pid=41276) WARNING 08-01 14:11:33 [vllm.py:859] Inductor compilation was disabled by user settings, optimizations settings that are only active during inductor compilation will be ignored. +(APIServer pid=41276) INFO 08-01 14:11:33 [vllm.py:1025] Cudagraph is disabled under eager mode +(APIServer pid=41276) INFO 08-01 14:11:33 [compilation.py:290] Enabled custom fusions: norm_quant, act_quant +Skipping import of cpp extensions due to incompatible torch version 2.10.0+cu128 for torchao version 0.15.0 Please see https://github.com/pytorch/ao/issues/2919 for more info +(EngineCore pid=41613) WARNING 08-01 14:11:41 [registry.py:915] Model architecture OlmoeForCausalLM is already registered, and will be overwritten by the new model class glean_vllm.pruned_olmoe:PrunedOlmoeForCausalLM. +(EngineCore pid=41613) INFO 08-01 14:11:41 [core.py:105] Initializing a V1 LLM engine (v0.19.0) with config: model='outputs/pruned/glean-0125inst-math-keep50', speculative_config=None, tokenizer='outputs/pruned/glean-0125inst-math-keep50', skip_tokenizer_init=False, tokenizer_mode=auto, revision=None, tokenizer_revision=None, trust_remote_code=False, dtype=torch.bfloat16, max_seq_len=3584, download_dir=None, load_format=auto, tensor_parallel_size=1, pipeline_parallel_size=1, data_parallel_size=1, decode_context_parallel_size=1, dcp_comm_backend=ag_rs, disable_custom_all_reduce=False, quantization=None, enforce_eager=True, enable_return_routed_experts=False, kv_cache_dtype=auto, device_config=cuda, structured_outputs_config=StructuredOutputsConfig(backend='auto', disable_any_whitespace=False, disable_additional_properties=False, reasoning_parser='', reasoning_parser_plugin='', enable_in_reasoning=False), observability_config=ObservabilityConfig(show_hidden_metrics_for_version=None, otlp_traces_endpoint=None, collect_detailed_traces=None, kv_cache_metrics=False, kv_cache_metrics_sample=0.01, cudagraph_metrics=False, enable_layerwise_nvtx_tracing=False, enable_mfu_metrics=False, enable_mm_processor_stats=False, enable_logging_iteration_details=False), seed=0, served_model_name=student, enable_prefix_caching=True, enable_chunked_prefill=True, pooler_config=None, compilation_config={'mode': , 'debug_dump_path': None, 'cache_dir': '', 'compile_cache_save_format': 'binary', 'backend': 'inductor', 'custom_ops': ['all'], 'splitting_ops': [], 'compile_mm_encoder': False, 'cudagraph_mm_encoder': False, 'encoder_cudagraph_token_budgets': [], 'encoder_cudagraph_max_images_per_batch': 0, 'compile_sizes': [], 'compile_ranges_endpoints': [2048], 'inductor_compile_config': {'enable_auto_functionalized_v2': False, 'size_asserts': False, 'alignment_asserts': False, 'scalar_asserts': False, 'combo_kernels': True, 'benchmark_combo_kernel': True}, 'inductor_passes': {}, 'cudagraph_mode': , 'cudagraph_num_of_warmups': 0, 'cudagraph_capture_sizes': [], 'cudagraph_copy_inputs': False, 'cudagraph_specialize_lora': True, 'use_inductor_graph_partition': False, 'pass_config': {'fuse_norm_quant': True, 'fuse_act_quant': True, 'fuse_attn_quant': False, 'enable_sp': False, 'fuse_gemm_comms': False, 'fuse_allreduce_rms': False}, 'max_cudagraph_capture_size': 0, 'dynamic_shapes_config': {'type': , 'evaluate_guards': False, 'assume_32_bit_indexing': False}, 'local_cache_dir': None, 'fast_moe_cold_start': True, 'static_all_moe_layers': []} +(EngineCore pid=41613) INFO 08-01 14:11:41 [parallel_state.py:1400] world_size=1 rank=0 local_rank=0 distributed_init_method=tcp://192.168.0.27:48291 backend=nccl +(EngineCore pid=41613) INFO 08-01 14:11:41 [parallel_state.py:1716] rank 0 in world size 1 is assigned as DP rank 0, PP rank 0, PCP rank 0, TP rank 0, EP rank 0, EPLB rank N/A +(EngineCore pid=41613) INFO 08-01 14:11:42 [gpu_model_runner.py:4735] Starting to load model outputs/pruned/glean-0125inst-math-keep50... +(EngineCore pid=41613) INFO 08-01 14:11:42 [cuda.py:334] Using FLASH_ATTN attention backend out of potential backends: ['FLASH_ATTN', 'FLASHINFER', 'TRITON_ATTN', 'FLEX_ATTENTION']. +(EngineCore pid=41613) INFO 08-01 14:11:42 [flash_attn.py:596] Using FlashAttention version 2 +(EngineCore pid=41613) Loading safetensors checkpoint shards: 0% Completed | 0/2 [00:00