diff options
Diffstat (limited to 'scripts/test_local_reward_15667317.out')
| -rw-r--r-- | scripts/test_local_reward_15667317.out | 87 |
1 files changed, 87 insertions, 0 deletions
diff --git a/scripts/test_local_reward_15667317.out b/scripts/test_local_reward_15667317.out new file mode 100644 index 0000000..6e45a21 --- /dev/null +++ b/scripts/test_local_reward_15667317.out @@ -0,0 +1,87 @@ +=== Local LLM Reward Model Batch Test === +Model: models/llama-3.1-8b-instruct +GPU: NVIDIA A100-SXM4-40GB + +Starting vLLM server on port 8005... +Waiting for vLLM server to start... +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:03:02 [api_server.py:1351] vLLM API server version 0.13.0 +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:03:02 [utils.py:253] non-default args: {'port': 8005, 'model': 'models/llama-3.1-8b-instruct', 'dtype': 'bfloat16', 'max_model_len': 4096, 'gpu_memory_utilization': 0.85} +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:03:02 [model.py:514] Resolved architecture: LlamaForCausalLM +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:03:02 [model.py:1661] Using max model len 4096 +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:03:03 [scheduler.py:230] Chunked prefill is enabled with max_num_batched_tokens=2048. +[0;36m(EngineCore_DP0 pid=2400286)[0;0m INFO 01-27 12:03:15 [core.py:93] Initializing a V1 LLM engine (v0.13.0) with config: model='models/llama-3.1-8b-instruct', speculative_config=None, tokenizer='models/llama-3.1-8b-instruct', skip_tokenizer_init=False, tokenizer_mode=auto, revision=None, tokenizer_revision=None, trust_remote_code=False, dtype=torch.bfloat16, max_seq_len=4096, download_dir=None, load_format=auto, tensor_parallel_size=1, pipeline_parallel_size=1, data_parallel_size=1, disable_custom_all_reduce=False, quantization=None, enforce_eager=False, kv_cache_dtype=auto, device_config=cuda, structured_outputs_config=StructuredOutputsConfig(backend='auto', disable_fallback=False, 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), seed=0, served_model_name=models/llama-3.1-8b-instruct, enable_prefix_caching=True, enable_chunked_prefill=True, pooler_config=None, compilation_config={'level': None, 'mode': <CompilationMode.VLLM_COMPILE: 3>, 'debug_dump_path': None, 'cache_dir': '', 'compile_cache_save_format': 'binary', 'backend': 'inductor', 'custom_ops': ['none'], 'splitting_ops': ['vllm::unified_attention', 'vllm::unified_attention_with_output', 'vllm::unified_mla_attention', 'vllm::unified_mla_attention_with_output', 'vllm::mamba_mixer2', 'vllm::mamba_mixer', 'vllm::short_conv', 'vllm::linear_attention', 'vllm::plamo2_mamba_mixer', 'vllm::gdn_attention_core', 'vllm::kda_attention', 'vllm::sparse_attn_indexer'], 'compile_mm_encoder': False, 'compile_sizes': [], 'compile_ranges_split_points': [2048], 'inductor_compile_config': {'enable_auto_functionalized_v2': False, 'combo_kernels': True, 'benchmark_combo_kernel': True}, 'inductor_passes': {}, 'cudagraph_mode': <CUDAGraphMode.FULL_AND_PIECEWISE: (2, 1)>, 'cudagraph_num_of_warmups': 1, 'cudagraph_capture_sizes': [1, 2, 4, 8, 16, 24, 32, 40, 48, 56, 64, 72, 80, 88, 96, 104, 112, 120, 128, 136, 144, 152, 160, 168, 176, 184, 192, 200, 208, 216, 224, 232, 240, 248, 256, 272, 288, 304, 320, 336, 352, 368, 384, 400, 416, 432, 448, 464, 480, 496, 512], 'cudagraph_copy_inputs': False, 'cudagraph_specialize_lora': True, 'use_inductor_graph_partition': False, 'pass_config': {'fuse_norm_quant': False, 'fuse_act_quant': False, 'fuse_attn_quant': False, 'eliminate_noops': True, 'enable_sp': False, 'fuse_gemm_comms': False, 'fuse_allreduce_rms': False}, 'max_cudagraph_capture_size': 512, 'dynamic_shapes_config': {'type': <DynamicShapesType.BACKED: 'backed'>, 'evaluate_guards': False}, 'local_cache_dir': None} +[0;36m(EngineCore_DP0 pid=2400286)[0;0m INFO 01-27 12:03:17 [parallel_state.py:1203] world_size=1 rank=0 local_rank=0 distributed_init_method=tcp://141.142.254.16:34265 backend=nccl +[0;36m(EngineCore_DP0 pid=2400286)[0;0m INFO 01-27 12:03:17 [parallel_state.py:1411] rank 0 in world size 1 is assigned as DP rank 0, PP rank 0, PCP rank 0, TP rank 0, EP rank 0 +[0;36m(EngineCore_DP0 pid=2400286)[0;0m INFO 01-27 12:03:19 [gpu_model_runner.py:3562] Starting to load model models/llama-3.1-8b-instruct... +[0;36m(EngineCore_DP0 pid=2400286)[0;0m INFO 01-27 12:03:20 [cuda.py:351] Using FLASH_ATTN attention backend out of potential backends: ('FLASH_ATTN', 'FLASHINFER', 'TRITON_ATTN', 'FLEX_ATTENTION') +[0;36m(EngineCore_DP0 pid=2400286)[0;0m INFO 01-27 12:03:37 [default_loader.py:308] Loading weights took 16.45 seconds +[0;36m(EngineCore_DP0 pid=2400286)[0;0m INFO 01-27 12:03:37 [gpu_model_runner.py:3659] Model loading took 14.9889 GiB memory and 17.567555 seconds +[0;36m(EngineCore_DP0 pid=2400286)[0;0m INFO 01-27 12:03:47 [backends.py:643] Using cache directory: /u/yurenh2/.cache/vllm/torch_compile_cache/1c763cd906/rank_0_0/backbone for vLLM's torch.compile +[0;36m(EngineCore_DP0 pid=2400286)[0;0m INFO 01-27 12:03:47 [backends.py:703] Dynamo bytecode transform time: 9.88 s +[0;36m(EngineCore_DP0 pid=2400286)[0;0m INFO 01-27 12:03:54 [backends.py:261] Cache the graph of compile range (1, 2048) for later use +[0;36m(EngineCore_DP0 pid=2400286)[0;0m INFO 01-27 12:03:59 [backends.py:278] Compiling a graph for compile range (1, 2048) takes 8.04 s +[0;36m(EngineCore_DP0 pid=2400286)[0;0m INFO 01-27 12:03:59 [monitor.py:34] torch.compile takes 17.92 s in total +[0;36m(EngineCore_DP0 pid=2400286)[0;0m INFO 01-27 12:04:00 [gpu_worker.py:375] Available KV cache memory: 17.35 GiB +[0;36m(EngineCore_DP0 pid=2400286)[0;0m INFO 01-27 12:04:00 [kv_cache_utils.py:1291] GPU KV cache size: 142,160 tokens +[0;36m(EngineCore_DP0 pid=2400286)[0;0m INFO 01-27 12:04:00 [kv_cache_utils.py:1296] Maximum concurrency for 4,096 tokens per request: 34.71x +[0;36m(EngineCore_DP0 pid=2400286)[0;0m INFO 01-27 12:04:05 [gpu_model_runner.py:4587] Graph capturing finished in 5 secs, took 0.56 GiB +[0;36m(EngineCore_DP0 pid=2400286)[0;0m INFO 01-27 12:04:05 [core.py:259] init engine (profile, create kv cache, warmup model) took 27.83 seconds +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [api_server.py:1099] Supported tasks: ['generate'] +[0;36m(APIServer pid=2399803)[0;0m WARNING 01-27 12:04:06 [model.py:1487] Default sampling parameters have been overridden by the model's Hugging Face generation config recommended from the model creator. If this is not intended, please relaunch vLLM instance with `--generation-config vllm`. +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [serving_responses.py:201] Using default chat sampling params from model: {'temperature': 0.6, 'top_p': 0.9} +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [serving_chat.py:137] Using default chat sampling params from model: {'temperature': 0.6, 'top_p': 0.9} +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [serving_completion.py:77] Using default completion sampling params from model: {'temperature': 0.6, 'top_p': 0.9} +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [serving_chat.py:137] Using default chat sampling params from model: {'temperature': 0.6, 'top_p': 0.9} +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [api_server.py:1425] Starting vLLM API server 0 on http://0.0.0.0:8005 +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:38] Available routes are: +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /openapi.json, Methods: GET, HEAD +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /docs, Methods: GET, HEAD +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /docs/oauth2-redirect, Methods: GET, HEAD +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /redoc, Methods: GET, HEAD +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /scale_elastic_ep, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /is_scaling_elastic_ep, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /tokenize, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /detokenize, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /inference/v1/generate, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /pause, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /resume, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /is_paused, Methods: GET +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /metrics, Methods: GET +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /health, Methods: GET +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /load, Methods: GET +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /v1/models, Methods: GET +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /version, Methods: GET +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /v1/responses, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /v1/responses/{response_id}, Methods: GET +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /v1/responses/{response_id}/cancel, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /v1/messages, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /v1/chat/completions, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /v1/completions, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /v1/audio/transcriptions, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /v1/audio/translations, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /ping, Methods: GET +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /ping, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /invocations, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /classify, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /v1/embeddings, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /score, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /v1/score, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /rerank, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /v1/rerank, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /v2/rerank, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO 01-27 12:04:06 [launcher.py:46] Route: /pooling, Methods: POST +[0;36m(APIServer pid=2399803)[0;0m INFO: 127.0.0.1:32858 - "GET /health HTTP/1.1" 200 OK +vLLM server ready after 100s +[0;36m(APIServer pid=2399803)[0;0m INFO: 127.0.0.1:32866 - "GET /health HTTP/1.1" 200 OK + +Running batch test... +[0;36m(APIServer pid=2399803)[0;0m INFO: 127.0.0.1:32870 - "GET /v1/models HTTP/1.1" 200 OK +====================================================================== +Local LLM Reward Model Batch Test +====================================================================== +vLLM URL: http://localhost:8005/v1 + +Model: models/llama-3.1-8b-instruct + + +=== Test Complete === |