nohup: ignoring input INFO 05-28 00:42:59 [__init__.py:239] Automatically detected platform cuda. INFO 05-28 00:43:01 [api_server.py:1043] vLLM API server version 0.8.5.post1 INFO 05-28 00:43:01 [api_server.py:1044] args: Namespace(host=None, port=8088, uvicorn_log_level='info', disable_uvicorn_access_log=False, allow_credentials=False, allowed_origins=['*'], allowed_methods=['*'], allowed_headers=['*'], api_key=None, lora_modules=None, prompt_adapters=None, chat_template=None, chat_template_content_format='auto', response_role='assistant', ssl_keyfile=None, ssl_certfile=None, ssl_ca_certs=None, enable_ssl_refresh=False, ssl_cert_reqs=0, root_path=None, middleware=[], return_tokens_as_token_ids=False, disable_frontend_multiprocessing=False, enable_request_id_headers=False, enable_auto_tool_choice=False, tool_call_parser=None, tool_parser_plugin='', model='myfi/V7_Parser', task='auto', tokenizer=None, hf_config_path=None, skip_tokenizer_init=False, revision=None, code_revision=None, tokenizer_revision=None, tokenizer_mode='auto', trust_remote_code=False, allowed_local_media_path=None, load_format='auto', download_dir=None, model_loader_extra_config={}, use_tqdm_on_load=True, config_format=, dtype='auto', max_model_len=5000, guided_decoding_backend='auto', reasoning_parser=None, logits_processor_pattern=None, model_impl='auto', distributed_executor_backend=None, pipeline_parallel_size=1, tensor_parallel_size=1, data_parallel_size=1, enable_expert_parallel=False, max_parallel_loading_workers=None, ray_workers_use_nsight=False, disable_custom_all_reduce=False, block_size=None, gpu_memory_utilization=0.95, swap_space=4, kv_cache_dtype='auto', num_gpu_blocks_override=600, enable_prefix_caching=None, prefix_caching_hash_algo='builtin', cpu_offload_gb=0, calculate_kv_scales=False, disable_sliding_window=False, use_v2_block_manager=True, seed=None, max_logprobs=20, disable_log_stats=False, quantization=None, rope_scaling=None, rope_theta=None, hf_token=None, hf_overrides=None, enforce_eager=False, max_seq_len_to_capture=8192, tokenizer_pool_size=0, tokenizer_pool_type='ray', tokenizer_pool_extra_config={}, limit_mm_per_prompt={}, mm_processor_kwargs=None, disable_mm_preprocessor_cache=False, enable_lora=None, enable_lora_bias=False, max_loras=1, max_lora_rank=16, lora_extra_vocab_size=256, lora_dtype='auto', long_lora_scaling_factors=None, max_cpu_loras=None, fully_sharded_loras=False, enable_prompt_adapter=None, max_prompt_adapters=1, max_prompt_adapter_token=0, device='auto', speculative_config=None, ignore_patterns=[], served_model_name=None, qlora_adapter_name_or_path=None, show_hidden_metrics_for_version=None, otlp_traces_endpoint=None, collect_detailed_traces=None, disable_async_output_proc=False, max_num_batched_tokens=3368, max_num_seqs=128, max_num_partial_prefills=1, max_long_partial_prefills=1, long_prefill_token_threshold=0, num_lookahead_slots=0, scheduler_delay_factor=0.0, preemption_mode=None, num_scheduler_steps=1, multi_step_stream_outputs=True, scheduling_policy='fcfs', enable_chunked_prefill=None, disable_chunked_mm_input=False, scheduler_cls='vllm.core.scheduler.Scheduler', override_neuron_config=None, override_pooler_config=None, compilation_config=None, kv_transfer_config=None, worker_cls='auto', worker_extension_cls='', generation_config='auto', override_generation_config=None, enable_sleep_mode=False, additional_config=None, enable_reasoning=False, disable_cascade_attn=False, disable_log_requests=False, max_log_len=None, disable_fastapi_docs=False, enable_prompt_tokens_details=False, enable_server_load_tracking=False) INFO 05-28 00:43:09 [config.py:717] This model supports multiple tasks: {'embed', 'score', 'generate', 'classify', 'reward'}. Defaulting to 'generate'. INFO 05-28 00:43:09 [config.py:2003] Chunked prefill is enabled with max_num_batched_tokens=3368. INFO 05-28 00:43:15 [__init__.py:239] Automatically detected platform cuda. INFO 05-28 00:43:17 [core.py:58] Initializing a V1 LLM engine (v0.8.5.post1) with config: model='myfi/V7_Parser', speculative_config=None, tokenizer='myfi/V7_Parser', skip_tokenizer_init=False, tokenizer_mode=auto, revision=None, override_neuron_config=None, tokenizer_revision=None, trust_remote_code=False, dtype=torch.float16, max_seq_len=5000, download_dir=None, load_format=LoadFormat.AUTO, tensor_parallel_size=1, pipeline_parallel_size=1, disable_custom_all_reduce=False, quantization=None, enforce_eager=False, kv_cache_dtype=auto, device_config=cuda, decoding_config=DecodingConfig(guided_decoding_backend='auto', reasoning_backend=None), observability_config=ObservabilityConfig(show_hidden_metrics=False, otlp_traces_endpoint=None, collect_model_forward_time=False, collect_model_execute_time=False), seed=None, served_model_name=myfi/V7_Parser, num_scheduler_steps=1, multi_step_stream_outputs=True, enable_prefix_caching=True, chunked_prefill_enabled=True, use_async_output_proc=True, disable_mm_preprocessor_cache=False, mm_processor_kwargs=None, pooler_config=None, compilation_config={"level":3,"custom_ops":["none"],"splitting_ops":["vllm.unified_attention","vllm.unified_attention_with_output"],"use_inductor":true,"compile_sizes":[],"use_cudagraph":true,"cudagraph_num_of_warmups":1,"cudagraph_capture_sizes":[512,504,496,488,480,472,464,456,448,440,432,424,416,408,400,392,384,376,368,360,352,344,336,328,320,312,304,296,288,280,272,264,256,248,240,232,224,216,208,200,192,184,176,168,160,152,144,136,128,120,112,104,96,88,80,72,64,56,48,40,32,24,16,8,4,2,1],"max_capture_size":512} WARNING 05-28 00:43:18 [utils.py:2522] Methods determine_num_available_blocks,device_config,get_cache_block_size_bytes,initialize_cache not implemented in INFO 05-28 00:43:18 [parallel_state.py:1004] rank 0 in world size 1 is assigned as DP rank 0, PP rank 0, TP rank 0 INFO 05-28 00:43:18 [cuda.py:221] Using Flash Attention backend on V1 engine. WARNING 05-28 00:43:18 [topk_topp_sampler.py:69] FlashInfer is not available. Falling back to the PyTorch-native implementation of top-p & top-k sampling. For the best performance, please install FlashInfer. INFO 05-28 00:43:18 [gpu_model_runner.py:1329] Starting to load model myfi/V7_Parser... INFO 05-28 00:43:19 [weight_utils.py:265] Using model weights format ['*.safetensors'] Loading safetensors checkpoint shards: 0% Completed | 0/2 [00:00system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nshow me some funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:44:18 [async_llm.py:252] Added request chatcmpl-dfb3939087a94c798a3d00438920efc2. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:44:23 [logger.py:39] Received request chatcmpl-890201b771094c9aaedc7dd0fcfcbc08: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nshow me low fee funds in largecap<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:44:23 [async_llm.py:252] Added request chatcmpl-890201b771094c9aaedc7dd0fcfcbc08. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:44:25 [loggers.py:111] Engine 000: Avg prompt throughput: 132.0 tokens/s, Avg generation throughput: 15.7 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 48.8% INFO 05-28 00:44:29 [logger.py:39] Received request chatcmpl-294ee5f0c56d4bffbefdac72e7307e7c: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nshow me high fee funds in midcap<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:44:29 [async_llm.py:252] Added request chatcmpl-294ee5f0c56d4bffbefdac72e7307e7c. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:44:33 [logger.py:39] Received request chatcmpl-3390c3cbd7794c45af1c162d71708ad0: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nshow me funds that have returns more than 50%<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:44:33 [async_llm.py:252] Added request chatcmpl-3390c3cbd7794c45af1c162d71708ad0. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:44:35 [loggers.py:111] Engine 000: Avg prompt throughput: 132.8 tokens/s, Avg generation throughput: 17.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 73.2% INFO 05-28 00:44:38 [logger.py:39] Received request chatcmpl-c796652eb9274d15b7e114be7f0ecb2b: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nbest midcap funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:44:38 [async_llm.py:252] Added request chatcmpl-c796652eb9274d15b7e114be7f0ecb2b. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:44:40 [logger.py:39] Received request chatcmpl-640b6199b3974af8acdba1701d0095a5: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nmidcap funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:44:40 [async_llm.py:252] Added request chatcmpl-640b6199b3974af8acdba1701d0095a5. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:44:42 [logger.py:39] Received request chatcmpl-abf96f5302994b2db08c38e21f60de76: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nfunds having AUM more than 5000 crores<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:44:42 [async_llm.py:252] Added request chatcmpl-abf96f5302994b2db08c38e21f60de76. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:44:45 [loggers.py:111] Engine 000: Avg prompt throughput: 198.3 tokens/s, Avg generation throughput: 22.6 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 83.6% INFO 05-28 00:44:46 [logger.py:39] Received request chatcmpl-4da37ae1a2674ef2a8f5ff80c7cf3901: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nfunds having AUM less than 1 crore<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:44:46 [async_llm.py:252] Added request chatcmpl-4da37ae1a2674ef2a8f5ff80c7cf3901. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:44:52 [logger.py:39] Received request chatcmpl-eded1e5908e248a0868610902364b172: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nfunds having expense ratio less than 1%<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:44:52 [async_llm.py:252] Added request chatcmpl-eded1e5908e248a0868610902364b172. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:44:55 [loggers.py:111] Engine 000: Avg prompt throughput: 132.8 tokens/s, Avg generation throughput: 13.3 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 86.7% INFO 05-28 00:44:56 [logger.py:39] Received request chatcmpl-e47b1c34f4c24eb9bc534037657bf376: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nfunds having expense ratio more than 2%<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:44:56 [async_llm.py:252] Added request chatcmpl-e47b1c34f4c24eb9bc534037657bf376. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:45:01 [logger.py:39] Received request chatcmpl-18a5bd1a45764a469e400256184f3240: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nshow me some pharma funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:45:01 [async_llm.py:252] Added request chatcmpl-18a5bd1a45764a469e400256184f3240. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:45:05 [logger.py:39] Received request chatcmpl-6f11575237f64800b80018557026f5e8: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nshow me some good midcap funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:45:05 [async_llm.py:252] Added request chatcmpl-6f11575237f64800b80018557026f5e8. INFO 05-28 00:45:05 [loggers.py:111] Engine 000: Avg prompt throughput: 198.5 tokens/s, Avg generation throughput: 14.6 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 7.2%, Prefix cache hit rate: 89.4% INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:45:10 [logger.py:39] Received request chatcmpl-1ad349df392c4d169b4fd623d3bdbd2a: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nshow me the best flexicap funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:45:10 [async_llm.py:252] Added request chatcmpl-1ad349df392c4d169b4fd623d3bdbd2a. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:45:14 [logger.py:39] Received request chatcmpl-f4f376a1aee34101b651ae0dafc996d6: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nshow me some tech funds with high returns & low risk<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:45:14 [async_llm.py:252] Added request chatcmpl-f4f376a1aee34101b651ae0dafc996d6. INFO 05-28 00:45:15 [loggers.py:111] Engine 000: Avg prompt throughput: 132.6 tokens/s, Avg generation throughput: 25.3 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 8.3%, Prefix cache hit rate: 90.6% INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:45:19 [logger.py:39] Received request chatcmpl-8555b77543bb46b698dc300be60fd578: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nwhat are the best largecap funds with financial exposures<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:45:19 [async_llm.py:252] Added request chatcmpl-8555b77543bb46b698dc300be60fd578. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:45:21 [logger.py:39] Received request chatcmpl-936f822e9e534895bf900962b5250d34: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\ndebt funds having expense ratio less than 1%<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:45:21 [async_llm.py:252] Added request chatcmpl-936f822e9e534895bf900962b5250d34. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:45:25 [loggers.py:111] Engine 000: Avg prompt throughput: 132.9 tokens/s, Avg generation throughput: 22.4 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 91.5% INFO 05-28 00:45:26 [logger.py:39] Received request chatcmpl-ea782e2f9cde46b98a7369350a72405b: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nfunds having AUM less than 663<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:45:26 [async_llm.py:252] Added request chatcmpl-ea782e2f9cde46b98a7369350a72405b. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:45:31 [logger.py:39] Received request chatcmpl-1c4d81df087f4bb692cfd59576489e97: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nhigh risk low cost small cap funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:45:31 [async_llm.py:252] Added request chatcmpl-1c4d81df087f4bb692cfd59576489e97. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:45:34 [logger.py:39] Received request chatcmpl-fc2c190476d04c79aef5cc4ddf5a4885: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nhigh volatility low cost small cap funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:45:34 [async_llm.py:252] Added request chatcmpl-fc2c190476d04c79aef5cc4ddf5a4885. INFO 05-28 00:45:35 [loggers.py:111] Engine 000: Avg prompt throughput: 198.7 tokens/s, Avg generation throughput: 32.8 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 8.3%, Prefix cache hit rate: 92.6% INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:45:37 [logger.py:39] Received request chatcmpl-ea9f9c93650a4a6c93624c2f282bf313: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nshow me some overnight funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:45:37 [async_llm.py:252] Added request chatcmpl-ea9f9c93650a4a6c93624c2f282bf313. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:45:42 [logger.py:39] Received request chatcmpl-8979bbf53d754fc1b210472683407a83: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nshow me some overnight funds with AUM <1000 crore<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:45:42 [async_llm.py:252] Added request chatcmpl-8979bbf53d754fc1b210472683407a83. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:45:45 [loggers.py:111] Engine 000: Avg prompt throughput: 132.7 tokens/s, Avg generation throughput: 19.5 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 93.0% INFO 05-28 00:45:47 [logger.py:39] Received request chatcmpl-1b22c5b66fcb4477b8331ea9c1b9db08: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nshow me some balanced hybrid funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:45:47 [async_llm.py:252] Added request chatcmpl-1b22c5b66fcb4477b8331ea9c1b9db08. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:45:52 [logger.py:39] Received request chatcmpl-06325d69ab714c2a8b5ce3643b432cb5: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nshow me some balanced hybrid funds with low cost<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:45:52 [async_llm.py:252] Added request chatcmpl-06325d69ab714c2a8b5ce3643b432cb5. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:45:55 [loggers.py:111] Engine 000: Avg prompt throughput: 132.3 tokens/s, Avg generation throughput: 15.9 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 93.4% INFO 05-28 00:45:58 [logger.py:39] Received request chatcmpl-2b220f569d2f4f79bb4325bf5710974d: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nmulti-cap funds with low expense ratios<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:45:58 [async_llm.py:252] Added request chatcmpl-2b220f569d2f4f79bb4325bf5710974d. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:46:00 [logger.py:39] Received request chatcmpl-e4a9cb3ae19d4b6182ebb65e15ac1cc6: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nlow expense ratio index funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:46:00 [async_llm.py:252] Added request chatcmpl-e4a9cb3ae19d4b6182ebb65e15ac1cc6. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:46:05 [logger.py:39] Received request chatcmpl-5173d092172d4db8906ea9511141f6fa: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nlow volatility large cap funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:46:05 [async_llm.py:252] Added request chatcmpl-5173d092172d4db8906ea9511141f6fa. INFO 05-28 00:46:05 [loggers.py:111] Engine 000: Avg prompt throughput: 197.9 tokens/s, Avg generation throughput: 19.6 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 7.3%, Prefix cache hit rate: 93.9% INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:46:08 [logger.py:39] Received request chatcmpl-33c7ebac6d3f498ab757745b3af97e30: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nbalanced funds with low risk<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:46:08 [async_llm.py:252] Added request chatcmpl-33c7ebac6d3f498ab757745b3af97e30. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:46:13 [logger.py:39] Received request chatcmpl-f28498f54cea43648a58679544fc0b57: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nshow me funds with returns more than 30% over the last 3 years<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:46:13 [async_llm.py:252] Added request chatcmpl-f28498f54cea43648a58679544fc0b57. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:46:15 [loggers.py:111] Engine 000: Avg prompt throughput: 133.0 tokens/s, Avg generation throughput: 28.3 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 94.2% INFO 05-28 00:46:18 [logger.py:39] Received request chatcmpl-c0fce03212464fe89804f94609e85d84: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nshow me funds with 3Y returns less than 30%<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:46:18 [async_llm.py:252] Added request chatcmpl-c0fce03212464fe89804f94609e85d84. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:46:23 [logger.py:39] Received request chatcmpl-7ca2efa403b846cfb92ae7e05613024b: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nshow me funds with 5Y returns more than 50%<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:46:23 [async_llm.py:252] Added request chatcmpl-7ca2efa403b846cfb92ae7e05613024b. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:46:25 [loggers.py:111] Engine 000: Avg prompt throughput: 133.6 tokens/s, Avg generation throughput: 17.7 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 94.4% INFO 05-28 00:46:29 [logger.py:39] Received request chatcmpl-b8f8c3c7c3e04b5bb0cabae1e351e033: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nshow me funds with high returns over 3 years<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:46:29 [async_llm.py:252] Added request chatcmpl-b8f8c3c7c3e04b5bb0cabae1e351e033. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:46:34 [logger.py:39] Received request chatcmpl-7f59df2de4ca447199502c052e566344: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nshow me high AUM debt funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:46:34 [async_llm.py:252] Added request chatcmpl-7f59df2de4ca447199502c052e566344. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:46:35 [loggers.py:111] Engine 000: Avg prompt throughput: 132.5 tokens/s, Avg generation throughput: 15.4 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 94.6% INFO 05-28 00:46:39 [logger.py:39] Received request chatcmpl-a2890113381148c7b68e0f6faaff8c1f: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nshow me some gold funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:46:39 [async_llm.py:252] Added request chatcmpl-a2890113381148c7b68e0f6faaff8c1f. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:46:43 [logger.py:39] Received request chatcmpl-8b20efceccab498aab049b34c3154f84: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n<|im_end|>\n<|im_start|>user\nfind some silver etfs<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:46:43 [async_llm.py:252] Added request chatcmpl-8b20efceccab498aab049b34c3154f84. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:46:45 [loggers.py:111] Engine 000: Avg prompt throughput: 131.8 tokens/s, Avg generation throughput: 15.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 94.8% INFO 05-28 00:46:55 [loggers.py:111] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 94.8% INFO 05-28 00:47:25 [logger.py:39] Received request chatcmpl-1137a706c1e84102bce2849a6f36e4e6: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nshow me some funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:47:25 [async_llm.py:252] Added request chatcmpl-1137a706c1e84102bce2849a6f36e4e6. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:47:30 [logger.py:39] Received request chatcmpl-6c49f7f0b3024320abe5a5fa813815f7: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nshow me low fee funds in largecap<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:47:30 [async_llm.py:252] Added request chatcmpl-6c49f7f0b3024320abe5a5fa813815f7. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:47:32 [logger.py:39] Received request chatcmpl-987d9cff69c74595ac061c40bed853a4: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nshow me high fee funds in midcap<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:47:32 [async_llm.py:252] Added request chatcmpl-987d9cff69c74595ac061c40bed853a4. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:47:34 [logger.py:39] Received request chatcmpl-605167efc3a244d5896b70014e55f379: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nshow me funds that have returns more than 50%<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:47:34 [async_llm.py:252] Added request chatcmpl-605167efc3a244d5896b70014e55f379. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:47:35 [loggers.py:111] Engine 000: Avg prompt throughput: 435.6 tokens/s, Avg generation throughput: 31.4 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 91.5% INFO 05-28 00:47:36 [logger.py:39] Received request chatcmpl-f622666f2ce14104b0b8d3893d6b9848: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nbest midcap funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:47:36 [async_llm.py:252] Added request chatcmpl-f622666f2ce14104b0b8d3893d6b9848. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:47:38 [logger.py:39] Received request chatcmpl-519042aba3494dd2ad11e1a6e934df6f: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nmidcap funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:47:38 [async_llm.py:252] Added request chatcmpl-519042aba3494dd2ad11e1a6e934df6f. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:47:40 [logger.py:39] Received request chatcmpl-1cc7ae1195a64afdbc5350d9b4de946e: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nfunds having AUM more than 5000 crores<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:47:40 [async_llm.py:252] Added request chatcmpl-1cc7ae1195a64afdbc5350d9b4de946e. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:47:44 [logger.py:39] Received request chatcmpl-2c44a849bfb944efa83ac7c747a9f52a: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nfunds having AUM less than 1 crore<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:47:44 [async_llm.py:252] Added request chatcmpl-2c44a849bfb944efa83ac7c747a9f52a. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:47:45 [loggers.py:111] Engine 000: Avg prompt throughput: 435.4 tokens/s, Avg generation throughput: 34.9 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 92.6% INFO 05-28 00:47:48 [logger.py:39] Received request chatcmpl-dbcd1ab6ee39423c9ca95220302a8716: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nfunds having expense ratio less than 1%<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:47:48 [async_llm.py:252] Added request chatcmpl-dbcd1ab6ee39423c9ca95220302a8716. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:47:53 [logger.py:39] Received request chatcmpl-14017537734f409687a04920dd4ff7ac: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nfunds having expense ratio more than 2%<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:47:53 [async_llm.py:252] Added request chatcmpl-14017537734f409687a04920dd4ff7ac. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:47:54 [logger.py:39] Received request chatcmpl-cde1f5b806ab4090b70d5b301b0ed0a6: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nshow me some pharma funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:47:54 [async_llm.py:252] Added request chatcmpl-cde1f5b806ab4090b70d5b301b0ed0a6. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:47:55 [loggers.py:111] Engine 000: Avg prompt throughput: 326.9 tokens/s, Avg generation throughput: 24.9 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 93.2% INFO 05-28 00:47:56 [logger.py:39] Received request chatcmpl-03539ca15b4b4541b24f92f40f1dce1c: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nshow me some good midcap funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:47:56 [async_llm.py:252] Added request chatcmpl-03539ca15b4b4541b24f92f40f1dce1c. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:47:58 [logger.py:39] Received request chatcmpl-31e2b6dfd6e0487ca4c7c8e9af62f391: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nshow me the best flexicap funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:47:58 [async_llm.py:252] Added request chatcmpl-31e2b6dfd6e0487ca4c7c8e9af62f391. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:48:03 [logger.py:39] Received request chatcmpl-82a11b5502bc471a8556411b4816df0f: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nshow me some tech funds with high returns & low risk<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:48:03 [async_llm.py:252] Added request chatcmpl-82a11b5502bc471a8556411b4816df0f. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:48:05 [loggers.py:111] Engine 000: Avg prompt throughput: 326.7 tokens/s, Avg generation throughput: 37.6 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 93.7% INFO 05-28 00:48:07 [logger.py:39] Received request chatcmpl-a611939f37db4d258032d335fe3a4fce: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nwhat are the best largecap funds with financial exposures<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:48:07 [async_llm.py:252] Added request chatcmpl-a611939f37db4d258032d335fe3a4fce. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:48:09 [logger.py:39] Received request chatcmpl-f45fec02d23b4206956dad3c7b7a966a: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\ndebt funds having expense ratio less than 1%<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:48:09 [async_llm.py:252] Added request chatcmpl-f45fec02d23b4206956dad3c7b7a966a. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:48:13 [logger.py:39] Received request chatcmpl-86ae6b06b0294e55ad291d8c39a29a62: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nfunds having AUM less than 663<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:48:13 [async_llm.py:252] Added request chatcmpl-86ae6b06b0294e55ad291d8c39a29a62. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:48:15 [loggers.py:111] Engine 000: Avg prompt throughput: 327.5 tokens/s, Avg generation throughput: 30.5 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 94.1% INFO 05-28 00:48:17 [logger.py:39] Received request chatcmpl-2f2bb88132cb4a8fbb945f4f175b7e32: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nhigh risk low cost small cap funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:48:17 [async_llm.py:252] Added request chatcmpl-2f2bb88132cb4a8fbb945f4f175b7e32. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:48:23 [logger.py:39] Received request chatcmpl-ab22b74d55bd4f25a64958f588865a56: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nhigh volatility low cost small cap funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:48:23 [async_llm.py:252] Added request chatcmpl-ab22b74d55bd4f25a64958f588865a56. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:48:25 [loggers.py:111] Engine 000: Avg prompt throughput: 217.6 tokens/s, Avg generation throughput: 33.1 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 94.4% INFO 05-28 00:48:28 [logger.py:39] Received request chatcmpl-112431e256eb4c21b2284daa7c04b57e: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nshow me some overnight funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:48:28 [async_llm.py:252] Added request chatcmpl-112431e256eb4c21b2284daa7c04b57e. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:48:30 [logger.py:39] Received request chatcmpl-8bad543669c6475cae6ea0dd258d4b3e: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nshow me some overnight funds with AUM <1000 crore<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:48:30 [async_llm.py:252] Added request chatcmpl-8bad543669c6475cae6ea0dd258d4b3e. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:48:31 [logger.py:39] Received request chatcmpl-8dab79cbfd9b4988a1aebcb2a1c75ad4: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nshow me some balanced hybrid funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:48:31 [async_llm.py:252] Added request chatcmpl-8dab79cbfd9b4988a1aebcb2a1c75ad4. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:48:33 [logger.py:39] Received request chatcmpl-a175eb9ea28f48ca97d5048345b520be: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nshow me some balanced hybrid funds with low cost<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:48:33 [async_llm.py:252] Added request chatcmpl-a175eb9ea28f48ca97d5048345b520be. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:48:35 [logger.py:39] Received request chatcmpl-b5dcf20a212b4ab090ae6ae8e5f21024: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nmulti-cap funds with low expense ratios<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:48:35 [async_llm.py:252] Added request chatcmpl-b5dcf20a212b4ab090ae6ae8e5f21024. INFO 05-28 00:48:35 [loggers.py:111] Engine 000: Avg prompt throughput: 544.6 tokens/s, Avg generation throughput: 35.6 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 12.2%, Prefix cache hit rate: 95.0% INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:48:36 [logger.py:39] Received request chatcmpl-2f2d98a40ae04980867945c73632a674: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nlow expense ratio index funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:48:36 [async_llm.py:252] Added request chatcmpl-2f2d98a40ae04980867945c73632a674. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:48:41 [logger.py:39] Received request chatcmpl-3c05b7c59f0d4e59b7b698e8a4fc76b4: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nlow volatility large cap funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:48:41 [async_llm.py:252] Added request chatcmpl-3c05b7c59f0d4e59b7b698e8a4fc76b4. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:48:44 [logger.py:39] Received request chatcmpl-0d70ea95e0c1433b97ed5c45a8532413: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nbalanced funds with low risk<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:48:44 [async_llm.py:252] Added request chatcmpl-0d70ea95e0c1433b97ed5c45a8532413. INFO 05-28 00:48:45 [loggers.py:111] Engine 000: Avg prompt throughput: 325.7 tokens/s, Avg generation throughput: 39.5 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 12.7%, Prefix cache hit rate: 95.3% INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:48:49 [logger.py:39] Received request chatcmpl-d42d47c24e7c45c7ad40b78901643a83: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nshow me funds with returns more than 30% over the last 3 years<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:48:49 [async_llm.py:252] Added request chatcmpl-d42d47c24e7c45c7ad40b78901643a83. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:48:54 [logger.py:39] Received request chatcmpl-e9b65070cce84b0ab8245f1ddeea4a1d: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nshow me funds with 3Y returns less than 30%<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:48:54 [async_llm.py:252] Added request chatcmpl-e9b65070cce84b0ab8245f1ddeea4a1d. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:48:55 [loggers.py:111] Engine 000: Avg prompt throughput: 219.3 tokens/s, Avg generation throughput: 21.5 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.4% INFO 05-28 00:48:58 [logger.py:39] Received request chatcmpl-3b9d32621f6f464ca6b197cc17cc1039: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nshow me funds with 5Y returns more than 50%<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:48:58 [async_llm.py:252] Added request chatcmpl-3b9d32621f6f464ca6b197cc17cc1039. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:49:00 [logger.py:39] Received request chatcmpl-5610bafdbd2e4786912f29b73cfc61bc: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nshow me funds with high returns over 3 years<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:49:00 [async_llm.py:252] Added request chatcmpl-5610bafdbd2e4786912f29b73cfc61bc. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:49:05 [logger.py:39] Received request chatcmpl-3b038c440a7e45f3a20702707eeedf17: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nshow me high AUM debt funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:49:05 [async_llm.py:252] Added request chatcmpl-3b038c440a7e45f3a20702707eeedf17. INFO 05-28 00:49:05 [loggers.py:111] Engine 000: Avg prompt throughput: 327.4 tokens/s, Avg generation throughput: 23.8 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 12.2%, Prefix cache hit rate: 95.6% INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:49:12 [logger.py:39] Received request chatcmpl-c0d026bf2b154aefac3975c4fc77c1cd: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nshow me some gold funds<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:49:12 [async_llm.py:252] Added request chatcmpl-c0d026bf2b154aefac3975c4fc77c1cd. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:49:14 [logger.py:39] Received request chatcmpl-06a1be3752574977a3f592916a309c2b: prompt: '<|im_start|>system\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n <|im_end|>\n<|im_start|>user\nfind some silver etfs<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 00:49:14 [async_llm.py:252] Added request chatcmpl-06a1be3752574977a3f592916a309c2b. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 00:49:15 [loggers.py:111] Engine 000: Avg prompt throughput: 217.2 tokens/s, Avg generation throughput: 14.5 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.8% INFO 05-28 00:49:25 [loggers.py:111] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.8% INFO 05-28 09:47:06 [logger.py:39] Received request chatcmpl-485e83773935437f855d3774fbb16cd3: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nshow me some funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:47:06 [async_llm.py:252] Added request chatcmpl-485e83773935437f855d3774fbb16cd3. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:47:07 [loggers.py:111] Engine 000: Avg prompt throughput: 113.8 tokens/s, Avg generation throughput: 4.5 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 94.0% INFO 05-28 09:47:15 [logger.py:39] Received request chatcmpl-8f98bc47858648a0925842b874673862: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nshow me low fee funds in largecap\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:47:15 [async_llm.py:252] Added request chatcmpl-8f98bc47858648a0925842b874673862. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:47:17 [loggers.py:111] Engine 000: Avg prompt throughput: 114.2 tokens/s, Avg generation throughput: 7.2 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 94.0% INFO 05-28 09:47:20 [logger.py:39] Received request chatcmpl-ae416b22d76b497ea17f91be13fe6e49: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nshow me high fee funds in midcap\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:47:20 [async_llm.py:252] Added request chatcmpl-ae416b22d76b497ea17f91be13fe6e49. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:47:24 [logger.py:39] Received request chatcmpl-d873ae26ad144b31a5ddf3a52c72af5b: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nshow me funds that have returns more than 50%\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:47:24 [async_llm.py:252] Added request chatcmpl-d873ae26ad144b31a5ddf3a52c72af5b. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:47:27 [loggers.py:111] Engine 000: Avg prompt throughput: 228.7 tokens/s, Avg generation throughput: 16.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 94.2% INFO 05-28 09:47:29 [logger.py:39] Received request chatcmpl-7ea99c3c700f4118b233dc2a259fe158: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nbest midcap funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:47:29 [async_llm.py:252] Added request chatcmpl-7ea99c3c700f4118b233dc2a259fe158. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:47:33 [logger.py:39] Received request chatcmpl-47bd713a1f894857a09bce740992caf1: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nmidcap funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:47:33 [async_llm.py:252] Added request chatcmpl-47bd713a1f894857a09bce740992caf1. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:47:37 [loggers.py:111] Engine 000: Avg prompt throughput: 227.5 tokens/s, Avg generation throughput: 20.7 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 94.4% INFO 05-28 09:47:38 [logger.py:39] Received request chatcmpl-a475cbf1737e4ccd8ee742f623d6f400: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nfunds having AUM more than 5000 crores\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:47:38 [async_llm.py:252] Added request chatcmpl-a475cbf1737e4ccd8ee742f623d6f400. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:47:45 [logger.py:39] Received request chatcmpl-d7ba2d26a317424ab9c7c4c717e73c5a: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nfunds having AUM less than 1 crore\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:47:45 [async_llm.py:252] Added request chatcmpl-d7ba2d26a317424ab9c7c4c717e73c5a. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:47:47 [loggers.py:111] Engine 000: Avg prompt throughput: 229.2 tokens/s, Avg generation throughput: 17.8 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 94.5% INFO 05-28 09:47:51 [logger.py:39] Received request chatcmpl-6d7d5d0e1d734f249295c27eb3c97a46: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nfunds having expense ratio less than 1%\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:47:51 [async_llm.py:252] Added request chatcmpl-6d7d5d0e1d734f249295c27eb3c97a46. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:47:56 [logger.py:39] Received request chatcmpl-51e64dcfa3314d83aeffadc966b47802: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nfunds having expense ratio more than 2%\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:47:56 [async_llm.py:252] Added request chatcmpl-51e64dcfa3314d83aeffadc966b47802. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:47:57 [loggers.py:111] Engine 000: Avg prompt throughput: 228.6 tokens/s, Avg generation throughput: 14.8 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 94.6% INFO 05-28 09:48:00 [logger.py:39] Received request chatcmpl-6375131a98d44b50b3d44fba9b74cd74: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nshow me some pharma funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:48:00 [async_llm.py:252] Added request chatcmpl-6375131a98d44b50b3d44fba9b74cd74. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:48:05 [logger.py:39] Received request chatcmpl-2ea23a9a7ddb405598e7c1bba5039389: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nshow me some good midcap funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:48:05 [async_llm.py:252] Added request chatcmpl-2ea23a9a7ddb405598e7c1bba5039389. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:48:07 [loggers.py:111] Engine 000: Avg prompt throughput: 228.1 tokens/s, Avg generation throughput: 20.4 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 94.8% INFO 05-28 09:48:10 [logger.py:39] Received request chatcmpl-6eb3065d1684468e85a8ae94ddba2376: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nshow me the best flexicap funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:48:10 [async_llm.py:252] Added request chatcmpl-6eb3065d1684468e85a8ae94ddba2376. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:48:17 [logger.py:39] Received request chatcmpl-e68e1b287665486fb32f14badf156e4f: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nshow me some tech funds with high returns & low risk\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:48:17 [async_llm.py:252] Added request chatcmpl-e68e1b287665486fb32f14badf156e4f. INFO 05-28 09:48:17 [loggers.py:111] Engine 000: Avg prompt throughput: 228.6 tokens/s, Avg generation throughput: 13.4 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 12.5%, Prefix cache hit rate: 94.9% INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:48:22 [logger.py:39] Received request chatcmpl-1029860ecf404280b38c7b5ddc33a697: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nwhat are the best largecap funds with financial exposures\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:48:22 [async_llm.py:252] Added request chatcmpl-1029860ecf404280b38c7b5ddc33a697. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:48:27 [logger.py:39] Received request chatcmpl-87c5e08678fd4b4aaeeded457d434a32: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\ndebt funds having expense ratio less than 1%\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:48:27 [async_llm.py:252] Added request chatcmpl-87c5e08678fd4b4aaeeded457d434a32. INFO 05-28 09:48:27 [loggers.py:111] Engine 000: Avg prompt throughput: 228.8 tokens/s, Avg generation throughput: 26.9 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 12.5%, Prefix cache hit rate: 95.0% INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:48:31 [logger.py:39] Received request chatcmpl-84156e15c058486ab72b91ceae1ef719: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nfunds having AUM less than 663\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:48:31 [async_llm.py:252] Added request chatcmpl-84156e15c058486ab72b91ceae1ef719. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:48:36 [logger.py:39] Received request chatcmpl-daa2e582a1694d1ba9eccdd2f92dd4a6: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nhigh risk low cost small cap funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:48:36 [async_llm.py:252] Added request chatcmpl-daa2e582a1694d1ba9eccdd2f92dd4a6. INFO 05-28 09:48:37 [loggers.py:111] Engine 000: Avg prompt throughput: 228.6 tokens/s, Avg generation throughput: 19.8 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 13.0%, Prefix cache hit rate: 95.1% INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:48:42 [logger.py:39] Received request chatcmpl-7fc8aab0c16748a9b74fad6c04c82256: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nhigh volatility low cost small cap funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:48:42 [async_llm.py:252] Added request chatcmpl-7fc8aab0c16748a9b74fad6c04c82256. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:48:47 [logger.py:39] Received request chatcmpl-8845c14bfd554361a96c327ce4afbfe4: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nshow me some overnight funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:48:47 [async_llm.py:252] Added request chatcmpl-8845c14bfd554361a96c327ce4afbfe4. INFO 05-28 09:48:47 [loggers.py:111] Engine 000: Avg prompt throughput: 228.0 tokens/s, Avg generation throughput: 25.4 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 12.5%, Prefix cache hit rate: 95.2% INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:48:51 [logger.py:39] Received request chatcmpl-4381948b87f748668aa189c2d90c8df6: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nshow me some overnight funds with AUM <1000 crore\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:48:51 [async_llm.py:252] Added request chatcmpl-4381948b87f748668aa189c2d90c8df6. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:48:56 [logger.py:39] Received request chatcmpl-2f5e352c2c6e4a08947b53690918429c: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nshow me some balanced hybrid funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:48:56 [async_llm.py:252] Added request chatcmpl-2f5e352c2c6e4a08947b53690918429c. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:48:57 [loggers.py:111] Engine 000: Avg prompt throughput: 228.8 tokens/s, Avg generation throughput: 19.5 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.3% INFO 05-28 09:49:02 [logger.py:39] Received request chatcmpl-5e63e315d48f42b9a18e2752a687bf98: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nshow me some balanced hybrid funds with low cost\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:49:02 [async_llm.py:252] Added request chatcmpl-5e63e315d48f42b9a18e2752a687bf98. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:49:07 [logger.py:39] Received request chatcmpl-eaad13ddbb964d93bbbe97c6f7a8eb96: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nmulti-cap funds with low expense ratios\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:49:07 [async_llm.py:252] Added request chatcmpl-eaad13ddbb964d93bbbe97c6f7a8eb96. INFO 05-28 09:49:07 [loggers.py:111] Engine 000: Avg prompt throughput: 228.4 tokens/s, Avg generation throughput: 8.4 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 12.2%, Prefix cache hit rate: 95.4% INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:49:12 [logger.py:39] Received request chatcmpl-148cdf899aa4436fb8c602136f6708c9: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nlow expense ratio index funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:49:12 [async_llm.py:252] Added request chatcmpl-148cdf899aa4436fb8c602136f6708c9. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:49:16 [logger.py:39] Received request chatcmpl-2b61f323fb79475f99f47c9056d8051e: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nlow volatility large cap funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:49:16 [async_llm.py:252] Added request chatcmpl-2b61f323fb79475f99f47c9056d8051e. INFO 05-28 09:49:17 [loggers.py:111] Engine 000: Avg prompt throughput: 227.8 tokens/s, Avg generation throughput: 23.6 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 13.0%, Prefix cache hit rate: 95.4% INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:49:21 [logger.py:39] Received request chatcmpl-4a2a13e8fc564221ab295c38b590bd5b: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nbalanced funds with low risk\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:49:21 [async_llm.py:252] Added request chatcmpl-4a2a13e8fc564221ab295c38b590bd5b. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:49:26 [logger.py:39] Received request chatcmpl-81d8e1ac5851453a9374aa719dfe7b73: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nshow me funds with returns more than 30% over the last 3 years\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:49:26 [async_llm.py:252] Added request chatcmpl-81d8e1ac5851453a9374aa719dfe7b73. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:49:27 [loggers.py:111] Engine 000: Avg prompt throughput: 229.0 tokens/s, Avg generation throughput: 20.3 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.5% INFO 05-28 09:49:31 [logger.py:39] Received request chatcmpl-251c82da34a44acbb40fa0f54c28297d: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nshow me funds with 3Y returns less than 30%\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:49:31 [async_llm.py:252] Added request chatcmpl-251c82da34a44acbb40fa0f54c28297d. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:49:36 [logger.py:39] Received request chatcmpl-ed8729e5eea243668a4ad1dede4700db: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nshow me funds with 5Y returns more than 50%\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:49:36 [async_llm.py:252] Added request chatcmpl-ed8729e5eea243668a4ad1dede4700db. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:49:37 [loggers.py:111] Engine 000: Avg prompt throughput: 229.4 tokens/s, Avg generation throughput: 17.9 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.6% INFO 05-28 09:49:40 [logger.py:39] Received request chatcmpl-033eeaa833004841bb63ffe61323104a: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nshow me funds with high returns over 3 years\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:49:40 [async_llm.py:252] Added request chatcmpl-033eeaa833004841bb63ffe61323104a. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:49:45 [logger.py:39] Received request chatcmpl-c13c9032f31b432bbb1744e51043c47e: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nshow me high AUM debt funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:49:45 [async_llm.py:252] Added request chatcmpl-c13c9032f31b432bbb1744e51043c47e. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:49:47 [loggers.py:111] Engine 000: Avg prompt throughput: 228.5 tokens/s, Avg generation throughput: 17.9 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.7% INFO 05-28 09:49:49 [logger.py:39] Received request chatcmpl-05540f8b30c34356b9837090825185b1: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nshow me some gold funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:49:49 [async_llm.py:252] Added request chatcmpl-05540f8b30c34356b9837090825185b1. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:49:54 [logger.py:39] Received request chatcmpl-476de96be62b4abea7b4359d75281f9b: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nfind some silver etfs\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:49:54 [async_llm.py:252] Added request chatcmpl-476de96be62b4abea7b4359d75281f9b. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:49:57 [loggers.py:111] Engine 000: Avg prompt throughput: 227.8 tokens/s, Avg generation throughput: 13.4 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.7% INFO 05-28 09:50:07 [loggers.py:111] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.7% INFO 05-28 09:50:52 [logger.py:39] Received request chatcmpl-d061dbe8c0ea4cb289caffbcee90bef5: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nshow me some funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:50:52 [async_llm.py:252] Added request chatcmpl-d061dbe8c0ea4cb289caffbcee90bef5. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:50:56 [logger.py:39] Received request chatcmpl-81bd3f88d9764928ac29a986f86276ec: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nshow me low fee funds in largecap\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:50:56 [async_llm.py:252] Added request chatcmpl-81bd3f88d9764928ac29a986f86276ec. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:50:57 [loggers.py:111] Engine 000: Avg prompt throughput: 142.6 tokens/s, Avg generation throughput: 13.9 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.1% INFO 05-28 09:51:02 [logger.py:39] Received request chatcmpl-ecf203c956ce46208944adbe6e741982: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nshow me high fee funds in midcap\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:51:02 [async_llm.py:252] Added request chatcmpl-ecf203c956ce46208944adbe6e741982. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:51:07 [logger.py:39] Received request chatcmpl-ea76a4472b724b2ca04742d57d1e4183: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nshow me funds that have returns more than 50%\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:51:07 [async_llm.py:252] Added request chatcmpl-ea76a4472b724b2ca04742d57d1e4183. INFO 05-28 09:51:07 [loggers.py:111] Engine 000: Avg prompt throughput: 143.3 tokens/s, Avg generation throughput: 9.3 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 8.0%, Prefix cache hit rate: 95.2% INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:51:12 [logger.py:39] Received request chatcmpl-69b4d354752a4b9eb17befdd1c6c7f33: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nbest midcap funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:51:12 [async_llm.py:252] Added request chatcmpl-69b4d354752a4b9eb17befdd1c6c7f33. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:51:17 [logger.py:39] Received request chatcmpl-b1c05263ece64477a511b62cb4478430: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nmidcap funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:51:17 [async_llm.py:252] Added request chatcmpl-b1c05263ece64477a511b62cb4478430. INFO 05-28 09:51:17 [loggers.py:111] Engine 000: Avg prompt throughput: 142.1 tokens/s, Avg generation throughput: 20.4 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 8.0%, Prefix cache hit rate: 95.2% INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:51:22 [logger.py:39] Received request chatcmpl-addf3ac30dd8439ba4252e4ca31837ae: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nfunds having AUM more than 5000 crores\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:51:22 [async_llm.py:252] Added request chatcmpl-addf3ac30dd8439ba4252e4ca31837ae. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:51:26 [logger.py:39] Received request chatcmpl-f9eeb0890be4482da19fd7aa0ca08bd6: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nfunds having AUM less than 1 crore\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:51:26 [async_llm.py:252] Added request chatcmpl-f9eeb0890be4482da19fd7aa0ca08bd6. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:51:27 [loggers.py:111] Engine 000: Avg prompt throughput: 143.8 tokens/s, Avg generation throughput: 19.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.2% INFO 05-28 09:51:30 [logger.py:39] Received request chatcmpl-7692dd17d50440ef917277a9149de76f: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nfunds having expense ratio less than 1%\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:51:30 [async_llm.py:252] Added request chatcmpl-7692dd17d50440ef917277a9149de76f. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:51:35 [logger.py:39] Received request chatcmpl-b763951498804644bc38d57d61bd74bb: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nfunds having expense ratio more than 2%\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:51:35 [async_llm.py:252] Added request chatcmpl-b763951498804644bc38d57d61bd74bb. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:51:37 [loggers.py:111] Engine 000: Avg prompt throughput: 143.2 tokens/s, Avg generation throughput: 17.8 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.3% INFO 05-28 09:51:40 [logger.py:39] Received request chatcmpl-51e8839afe9c43a8bab2fa3d0dd41a25: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nshow me some pharma funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:51:40 [async_llm.py:252] Added request chatcmpl-51e8839afe9c43a8bab2fa3d0dd41a25. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:51:44 [logger.py:39] Received request chatcmpl-6e6529dcef36404482dad9ff7c490f60: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nshow me some good midcap funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:51:44 [async_llm.py:252] Added request chatcmpl-6e6529dcef36404482dad9ff7c490f60. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:51:47 [loggers.py:111] Engine 000: Avg prompt throughput: 142.7 tokens/s, Avg generation throughput: 14.3 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.3% INFO 05-28 09:51:49 [logger.py:39] Received request chatcmpl-b7424df7202740a6a522599c65faf8f2: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nshow me the best flexicap funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:51:49 [async_llm.py:252] Added request chatcmpl-b7424df7202740a6a522599c65faf8f2. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:51:54 [logger.py:39] Received request chatcmpl-2894c236d6f04f968079739d9455465a: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nshow me some tech funds with high returns & low risk\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:51:54 [async_llm.py:252] Added request chatcmpl-2894c236d6f04f968079739d9455465a. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:51:57 [loggers.py:111] Engine 000: Avg prompt throughput: 143.2 tokens/s, Avg generation throughput: 20.8 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.3% INFO 05-28 09:52:01 [logger.py:39] Received request chatcmpl-1f2da7f769774f63a433409bfbd0670f: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nwhat are the best largecap funds with financial exposures\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:52:01 [async_llm.py:252] Added request chatcmpl-1f2da7f769774f63a433409bfbd0670f. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:52:06 [logger.py:39] Received request chatcmpl-904cc9e6460c4af2b4b8d035d96aaac3: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\ndebt funds having expense ratio less than 1%\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:52:06 [async_llm.py:252] Added request chatcmpl-904cc9e6460c4af2b4b8d035d96aaac3. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:52:07 [loggers.py:111] Engine 000: Avg prompt throughput: 143.4 tokens/s, Avg generation throughput: 20.3 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.4% INFO 05-28 09:52:10 [logger.py:39] Received request chatcmpl-df8e34550b24434da248c3be207d432e: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nfunds having AUM less than 663\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:52:10 [async_llm.py:252] Added request chatcmpl-df8e34550b24434da248c3be207d432e. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:52:14 [logger.py:39] Received request chatcmpl-556d9b7182c14b13901f0c891a1ffa30: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nhigh risk low cost small cap funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:52:14 [async_llm.py:252] Added request chatcmpl-556d9b7182c14b13901f0c891a1ffa30. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:52:17 [loggers.py:111] Engine 000: Avg prompt throughput: 143.2 tokens/s, Avg generation throughput: 18.7 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.4% INFO 05-28 09:52:19 [logger.py:39] Received request chatcmpl-f6a282abc2fc45ffb9bcdeb615646a5b: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nhigh volatility low cost small cap funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:52:19 [async_llm.py:252] Added request chatcmpl-f6a282abc2fc45ffb9bcdeb615646a5b. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:52:24 [logger.py:39] Received request chatcmpl-9f5f5fd7100448deb89d7bfae02b8c21: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nshow me some overnight funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:52:24 [async_llm.py:252] Added request chatcmpl-9f5f5fd7100448deb89d7bfae02b8c21. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:52:27 [loggers.py:111] Engine 000: Avg prompt throughput: 142.6 tokens/s, Avg generation throughput: 21.3 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.4% INFO 05-28 09:52:29 [logger.py:39] Received request chatcmpl-8c3f8065205e4d278db4e1eaba90962a: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nshow me some overnight funds with AUM <1000 crore\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:52:29 [async_llm.py:252] Added request chatcmpl-8c3f8065205e4d278db4e1eaba90962a. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:52:33 [logger.py:39] Received request chatcmpl-538da2be059f4da78940508d7a298b7d: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nshow me some balanced hybrid funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:52:33 [async_llm.py:252] Added request chatcmpl-538da2be059f4da78940508d7a298b7d. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:52:37 [logger.py:39] Received request chatcmpl-aa0edba5c136442ab53920849442d062: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nshow me some balanced hybrid funds with low cost\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:52:37 [async_llm.py:252] Added request chatcmpl-aa0edba5c136442ab53920849442d062. INFO 05-28 09:52:37 [loggers.py:111] Engine 000: Avg prompt throughput: 215.0 tokens/s, Avg generation throughput: 16.3 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 7.8%, Prefix cache hit rate: 95.5% INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:52:41 [logger.py:39] Received request chatcmpl-913dfe9bc60044658dcf2de3c70d412a: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nmulti-cap funds with low expense ratios\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:52:41 [async_llm.py:252] Added request chatcmpl-913dfe9bc60044658dcf2de3c70d412a. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:52:47 [loggers.py:111] Engine 000: Avg prompt throughput: 71.4 tokens/s, Avg generation throughput: 12.4 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.5% INFO 05-28 09:52:49 [logger.py:39] Received request chatcmpl-3e8183056a6a42ffb4ef2ff2cf7a6d84: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nlow expense ratio index funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:52:49 [async_llm.py:252] Added request chatcmpl-3e8183056a6a42ffb4ef2ff2cf7a6d84. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:52:53 [logger.py:39] Received request chatcmpl-6948696b569343c3984a3e5db4a16901: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nlow volatility large cap funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:52:53 [async_llm.py:252] Added request chatcmpl-6948696b569343c3984a3e5db4a16901. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:52:57 [loggers.py:111] Engine 000: Avg prompt throughput: 142.4 tokens/s, Avg generation throughput: 19.2 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.5% INFO 05-28 09:53:07 [loggers.py:111] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.5% INFO 05-28 09:53:13 [logger.py:39] Received request chatcmpl-ab28c82422a14a6b900476957fa4b93a: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nbalanced funds with low risk\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:53:13 [async_llm.py:252] Added request chatcmpl-ab28c82422a14a6b900476957fa4b93a. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:53:17 [loggers.py:111] Engine 000: Avg prompt throughput: 71.2 tokens/s, Avg generation throughput: 8.5 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.5% INFO 05-28 09:53:19 [logger.py:39] Received request chatcmpl-e87237785ef141c58e710eababde3bb9: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nshow me funds with returns more than 30% over the last 3 years\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:53:19 [async_llm.py:252] Added request chatcmpl-e87237785ef141c58e710eababde3bb9. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:53:27 [loggers.py:111] Engine 000: Avg prompt throughput: 72.4 tokens/s, Avg generation throughput: 9.1 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.5% INFO 05-28 09:53:29 [logger.py:39] Received request chatcmpl-e306b123d18f4c1e968c18ece4abdbfb: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nshow me funds with 3Y returns less than 30%\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:53:29 [async_llm.py:252] Added request chatcmpl-e306b123d18f4c1e968c18ece4abdbfb. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:53:34 [logger.py:39] Received request chatcmpl-775ddd783d5c420f8bc3c51bf4a7bb39: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nshow me funds with 5Y returns more than 50%\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:53:34 [async_llm.py:252] Added request chatcmpl-775ddd783d5c420f8bc3c51bf4a7bb39. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:53:37 [loggers.py:111] Engine 000: Avg prompt throughput: 144.0 tokens/s, Avg generation throughput: 17.7 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.6% INFO 05-28 09:53:42 [logger.py:39] Received request chatcmpl-e757ed65459d498198de554b0c0b0b4a: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nshow me funds with high returns over 3 years\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:53:42 [async_llm.py:252] Added request chatcmpl-e757ed65459d498198de554b0c0b0b4a. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:53:47 [logger.py:39] Received request chatcmpl-3c4a315784a144d28cd9d92106b1e1ad: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nshow me high AUM debt funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:53:47 [async_llm.py:252] Added request chatcmpl-3c4a315784a144d28cd9d92106b1e1ad. INFO 05-28 09:53:47 [loggers.py:111] Engine 000: Avg prompt throughput: 143.1 tokens/s, Avg generation throughput: 11.8 tokens/s, Running: 1 reqs, Waiting: 0 reqs, GPU KV cache usage: 8.0%, Prefix cache hit rate: 95.6% INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:53:54 [logger.py:39] Received request chatcmpl-087b4a4506244c5aa2f007a5027f33ee: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nshow me some gold funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:53:54 [async_llm.py:252] Added request chatcmpl-087b4a4506244c5aa2f007a5027f33ee. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:53:57 [loggers.py:111] Engine 000: Avg prompt throughput: 71.2 tokens/s, Avg generation throughput: 11.3 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.6% INFO 05-28 09:54:02 [logger.py:39] Received request chatcmpl-c3880ef58d1741ab88afadfb4dcc4a04: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}]\n}\n\n\n# Guidelines\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\')\n - For queries related to large-cap, mid-cap, small-cap, use \'attribute/style/size/large_cap\', \'attribute/style/size/mid_cap\', \'attribute/style/size/small_cap\'\n - Qualifier: "high", "low", or null\n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: \n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\')\n - Use \'exposure/ticker/security_name\' when there is a ticker which is being queried as holdings/compositions of the main security.\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n - Tickers - Exposure tickers refers to the list of tickers that are being queried as holdings/compositions of the main security.\n\n3. Tickers:\n - Include any mentioned securities names (e.g., \'Nippon India Nifty smallcap\', \'Tata\', \'Edelweiss\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - If stock or equity fund is mentioned, use \'asset_type/equity\'\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi/equity_schemes/large_cap_fund\')\n - For sector or theme-specific requests, use \'sebi/equity_schemes/sectoral_/_thematic_fund\'\n - For tax-saving funds, use \'sebi/equity_schemes/elss\'\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/funds/etf\')\n - Always include if a specific investment vehicle is mentioned or implied\n\n# Important Notes:\n- Today\'s date: 2025-05-28\n- For qualitative performance: use \'attribute/technical/performance\'\n- Be precise with quantifiers\n- Provide only the structured output, no additional text\n\n\n### Input:\nfind some silver etfs\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:54:02 [async_llm.py:252] Added request chatcmpl-c3880ef58d1741ab88afadfb4dcc4a04. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:54:07 [loggers.py:111] Engine 000: Avg prompt throughput: 71.2 tokens/s, Avg generation throughput: 7.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.6% INFO 05-28 09:54:17 [loggers.py:111] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.6% INFO 05-28 09:55:55 [logger.py:39] Received request chatcmpl-a487bca521324f22bef2a81ccb828e50: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nshow me some funds\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:55:55 [async_llm.py:252] Added request chatcmpl-a487bca521324f22bef2a81ccb828e50. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:55:57 [loggers.py:111] Engine 000: Avg prompt throughput: 113.8 tokens/s, Avg generation throughput: 4.5 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.6% INFO 05-28 09:56:01 [logger.py:39] Received request chatcmpl-68b29c9b42914005b54a629c99d85a4b: prompt: '<|im_start|>system\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\n<|im_start|>user\nBelow is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.\n\n### Instruction:\n\nYou are a semantic parser specializing in financial queries. Your task is to analyze financial queries and extract relevant information into a structured format.\n\n# Output Schema\n\n{\n "attributes": [{"node": string, "qualifier": string|null, "time": string|null, "quantifier": string|null}],\n "exposures": [{"node": string|null, "qualifier": string|null, "quantifier": string|null}],\n "tickers": [{"name": string}],\n "amc": [{"name": string}],\n "asset_types": [{"node": string}],\n "sebi": [{"node": string}],\n "vehicles": [{"node": string}],\n "objectives": [{"node": string}]\n}\n\n# Guidelines\n\nFollow the guidelines respectively for each section\n\n1. Attributes:\n - Use full hierarchical paths (e.g., \'attribute/technical/returns/annualized_returns\', \'attribute/technical/aum\')\n - IMPORTANT:- [For performance-related queries or queries that include "best" or "good" or "better" or "top" or "worse" or "bad" options, always explicitly include \'attribute/technical/returns\' or \'attribute/technical/performance\']\n - If volatility is mentioned in queries , use \'attribute/technical/risk/volatility\' else use \'attribute/technical/risk\'\n - IMPORTANT:- [Qualifier: Only use "high" or "low" when explicitly mentioned in the query, Else Default to null] \n - Time: Use \'d\', \'w\', \'m\', \'y\', \'ytd\' units. For recent/current, use \'1w\'\n - Quantifier format: Use standardized format (e.g., ">5%", "<20%", "75%")\n\n2. Exposures:\n - Use full sector paths (e.g., \'exposure/sector/financials\', \'exposure/sector/consumer_staples\')\n - For tech/ technology fund use \'exposure/sector/information_technology\' \n - IMPORTANT : [ FOR QUERIES THAT HAS "HOLDS"/ "HOLD"/ "WITH" A TICKER, USE \'EXPOSURE/TICKER/\' ]\n - Qualifier: "high" (default), "low", or null\n - Quantifier format: % (e.g., ">5%")\n\n3. Tickers:\n - IMPORTANT : Include any mentioned securities names (e.g., \'Edelweiss\', \'Axis Large Cap\', \'SBI Mid Cap\', \'Kotak Large cap\')\n - Use exact names as mentioned in the query\n\n4. Amc:\n - IMPORTANT : Include any mentioned fund house names (e.g., \'dsp\', \'HDFC\', \'Axis\', \'Kotak\')\n - Use exact names as mentioned in the query\n\n4. Asset Types:\n - Use full paths (e.g., \'asset_type/equity\', \'asset_type/bonds\')\n - IMPORTANT :- [If \'equity\' keyword is mentioned, use \'asset_type/equity\']\n\n5. SEBI:\n - Use full classification paths (e.g., \'sebi_classification/equity_schemes/sectoral_/_thematic_fund\', \'sebi_classification/equity_schemes/mid_cap_fund\', \'sebi_classification/hybrid_schemes/balanced_hybrid_fund)\n - IMPORTANT : [ For QUERIES HAVING "NIFTY"/"nifty" KEYWORD use \'sebi_classification/other/index_funds_etfs\']\n - IMPORTANT : [ For all "multi cap"/"multicap" keywords in the query, always include "sebi_classification/equity_schemes/multi_cap_fund" ]\n - FOR TAX-SAVING FUNDS, USE \'sebi_classification/equity_schemes/elss\'\n - Include SEBI classifications if specific sectors, themes, or solutions are mentioned\n\n6. Vehicles:\n - Use full paths (e.g., \'vehicle/funds/mutual_fund\', \'vehicle/stock\', \'vehicle/portfolio\')\n - Use ETF when explicitly mentioned, else always default to \'vehicle/funds/mutual_fund\'\n - For NFO-related queries (e.g., "Show me SBI\'s NFOs", "NFOs to invest in"), use \'vehicle/method/nfo\' ONLY\n - IMPORTANT: When NFO is mentioned, do NOT include \'vehicle/funds/mutual_fund\' - use only \'vehicle/method/nfo\'\n - IMPORTANT: Do not leave the vehicles empty, if the vehicle is unclear or not mentioned, default to \'vechicle/mutual_fund\'\n - Always include if a specific investment vehicle is mentioned or implied\n\n7. Objectives:\n - Use full paths (e.g., \'objectives/growth\', \'objectives/momentum\')\n\n# Important Notes:\n - For tickers, if the query mentions two AMCs followed by the fund name, include both AMCs along with the fund name in the ticker list.\n - For sector-based thematic funds (e.g., tech, FMCG, banking), always include "sebi_classification/equity_schemes/sectoral_/_thematic_fund"\n - For debt funds, always include "sebi_classification/debt_schemes"\n \n\n### Input:\nshow me low fee funds in largecap\n\n### Response:<|im_end|>\n<|im_start|>assistant\n', params: SamplingParams(n=1, presence_penalty=0.0, frequency_penalty=0.0, repetition_penalty=1.05, temperature=0.01, top_p=0.95, top_k=20, min_p=0.0, seed=1, stop=[], stop_token_ids=[], bad_words=[], include_stop_str_in_output=False, ignore_eos=False, max_tokens=1024, min_tokens=0, logprobs=None, prompt_logprobs=None, skip_special_tokens=True, spaces_between_special_tokens=True, truncate_prompt_tokens=None, guided_decoding=None, extra_args=None), prompt_token_ids: None, lora_request: None, prompt_adapter_request: None. INFO 05-28 09:56:01 [async_llm.py:252] Added request chatcmpl-68b29c9b42914005b54a629c99d85a4b. INFO: 2401:4900:1f27:690a:f1f0:b19c:8a51:b27b:0 - "POST /v1/chat/completions HTTP/1.1" 200 OK INFO 05-28 09:56:07 [loggers.py:111] Engine 000: Avg prompt throughput: 114.2 tokens/s, Avg generation throughput: 7.2 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.7% INFO 05-28 09:56:17 [loggers.py:111] Engine 000: Avg prompt throughput: 0.0 tokens/s, Avg generation throughput: 0.0 tokens/s, Running: 0 reqs, Waiting: 0 reqs, GPU KV cache usage: 0.2%, Prefix cache hit rate: 95.7%