# file: /home/runner/work/flameox/flameox/src/flameox/adapters/inference.py
# hypothesis_version: 6.161.4

[0.0, 0.95, 1.0, 1000.0, 1000000.0, 1000000000.0, 100, 256, 999, 1024, 10000, 65536, 100000, 1000000, 1000000000, '*', '-', '.', '2', '95', '_', '_ms', 'actual_duration', 'after', 'aggregate', 'aggregation', 'aiperf', 'aiperf.input_tokens', 'aiperf.output_tokens', 'aiperf.requests', 'aiperf_0.12', 'artifact_id', 'authenticat', 'authentication', 'bad_request', 'before', 'block_id', 'cache_hit', 'cancel', 'cancelled', 'code', 'completed', 'connect', 'connection', 'conversation_id', 'count', 'credit_issued_ns', 'data', 'data.item', 'data.item.payloads', 'data.item.session_id', 'deadline', 'decode_ns', 'derived', 'dimensions', 'duration', 'e2el', 'end_array', 'end_map', 'end_to_end_latency', 'error', 'error_code', 'error_type', 'errors', 'evidence_level', 'evidence_run_id', 'failed_requests', 'flameox.inference', 'forbidden', 'generated_texts', 'hash_ids', 'ignore', 'inference_requests', 'input_length', 'input_lens', 'input_tokens', 'inter_token_latency', 'internal', 'invalid', 'invalid_request', 'is_warmup', 'itl', 'itls', 'json', 'kind', 'latency_ns', 'line_index', 'loop_count', 'map_key', 'mean', 'mean_', 'mean_e2el_ms', 'mean_itl_ms', 'mean_itl_ns', 'mean_tpot_ms', 'mean_ttft_ms', 'measurement_id', 'measurements', 'median', 'median_', 'median_e2el_ms', 'median_itl_ms', 'median_tpot_ms', 'median_ttft_ms', 'metadata', 'metrics', 'mooncake', 'ms', 'name', 'network', 'not_found', 'ns', 'num_prompts', 'observed', 'observed_started_ns', 'order_in_block', 'output_length', 'output_lens', 'output_throughput', 'output_tokens', 'p', 'parser_version', 'percentile', 'percentiles_e2el_ms', 'percentiles_itl_ms', 'percentiles_tpot_ms', 'percentiles_ttft_ms', 'permission', 'permission_denied', 'phase', 'prefill_ns', 'prefix_hash_count', 'producer', 'profile_export.jsonl', 'provider_error', 'provider_request_id', 'queue_ns', 'rate', 'rate_limit', 'rate_limited', 'rb', 'request_goodput', 'request_id', 'request_latency', 'request_start_ns', 'request_throughput', 'requests', 'requests/s', 'requests/sec', 'row', 'run_id', 's', 'scheduled_ns', 'scope', 'server', 'server_error', 'session_num', 'sglang.bench_serving', 'sglang_bench_serving', 'source_request_id', 'start_array', 'start_map', 'stat', 'std', 'std_', 'std_e2el_ms', 'std_itl_ms', 'std_tpot_ms', 'std_ttft_ms', 'steady_state', 'string', 'success', 'successful_requests', 'sum', 'throttl', 'throughput', 'time_scale', 'time_to_first_token', 'timed_out', 'timeout', 'timestamp', 'timestamp_ms', 'tokens', 'tokens/sec', 'total_input', 'total_input_tokens', 'total_output', 'total_output_tokens', 'total_requests', 'tpot', 'tpot_ns', 'trial_id', 'ttft', 'ttft_ns', 'ttfts', 'turn_index', 'type', 'unauthor', 'unavailable', 'unit', 'us', 'utf-8', 'validation', 'value', 'value_float', 'value_index', 'value_int', 'variant_id', 'vllm.failed_requests', 'vllm.request_goodput', 'vllm.total_requests', 'vllm_bench', 'was_cancelled', 'worker_id', 'worker_run_index', 'workload', 'x_request_id', 'µs']