{ "version": "1.0-auto", "generated_at": "2026-07-22T14:01:12+02:00", "source": "huggingface", "org": "Renesas", "catalog_note": "Auto-generated from Hugging Face 'Renesas' repositories on 2026-07-22 14:01 W. Europe Daylight Time. Metrics reflect each repo's published benchmark YAML files.", "models": [ { "key": "llama-3.2-1b-instruct", "display_name": "Llama-3.2-1B-Instruct", "architecture": "Llama-3.2-1B", "model_size_m": 1000, "modality": "LLM", "variant": "Instruct", "targets": [ "NPU" ], "onnx_repo": "Renesas/Llama-3.2-1B-Instruct-ONNX", "gguf_repo": "Renesas/Llama-3.2-1B-Instruct-GGUF", "onnx_artifacts": [ "fp16", "w4a16", "w4a8", "fp32" ], "gguf_artifacts": [ "fp16", "w4a16" ], "base_model": "meta-llama/Llama-3.2-1B-Instruct", "license": "llama3.2", "tasks": [ "text-generation", "instruction-following", "chat" ], "sample_image": null, "created_at": "2026-02-20T08:59:16.000Z", "last_modified": "2026-07-02T13:35:13.000Z", "hardware_configs": [ { "id": "RCAR-X5H", "label": "R-Car X5H (NPX6-48K)", "targets": [ "NPU" ], "notes": "vendor=renesas; chip=rcar-x5h; cpu=arm-cortex-a720; npu=npx6-48k", "metrics": { "artifacts": { "fp16": { "estimate": { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 18.73, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 18.73, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "ppa-estimator", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 } }, "hil": { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 20.25, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 20.25, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "NPX_Toolkit", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "setup": "R-Car X5H (NPX6-48K)" }, "runs": [ { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 20.25, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 20.25, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "NPX_Toolkit", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "setup": "R-Car X5H (NPX6-48K)", "kind": "hil" }, { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 18.73, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 18.73, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "ppa-estimator", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "kind": "estimate" } ] }, "w4a16": { "estimate": { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 89.59, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 60323.84 }, "total": { "tok_s": 89.59, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 60323.84 }, "engine": "ppa-estimator", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "memory_bandwidth": "high", "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 } }, "hil": { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 41.38, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 41.38, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "NPX_Toolkit", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "setup": "R-Car X5H (NPX6-48K)" }, "runs": [ { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 41.38, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 41.38, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "NPX_Toolkit", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "setup": "R-Car X5H (NPX6-48K)", "kind": "hil" }, { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 89.59, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 60323.84 }, "total": { "tok_s": 89.59, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 60323.84 }, "engine": "ppa-estimator", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "memory_bandwidth": "high", "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "kind": "estimate" }, { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 63.12, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 42506.24 }, "total": { "tok_s": 63.12, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 42506.24 }, "engine": "ppa-estimator", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "memory_bandwidth": "low", "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "kind": "estimate" } ] }, "fp32": { "estimate": null, "hil": null, "runs": [] }, "w4a8": { "estimate": { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 111.01, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 73707.52 }, "total": { "tok_s": 111.01, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 73707.52 }, "engine": "ppa-estimator", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "memory_bandwidth": "high", "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 } }, "hil": null, "runs": [ { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 111.01, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 73707.52 }, "total": { "tok_s": 111.01, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 73707.52 }, "engine": "ppa-estimator", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "memory_bandwidth": "high", "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "kind": "estimate" }, { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 72.98, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 51138.56 }, "total": { "tok_s": 72.98, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 51138.56 }, "engine": "ppa-estimator", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "memory_bandwidth": "low", "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "kind": "estimate" } ] } } } } ] }, { "key": "qwen3-asr-1.7b", "display_name": "Qwen3-ASR-1.7B", "architecture": "Qwen3-ASR-1.7B", "model_size_m": 1700, "modality": "ALM", "variant": "Base", "targets": [ "NPU" ], "onnx_repo": "", "gguf_repo": "Renesas/Qwen3-ASR-1.7B-GGUF", "onnx_artifacts": [], "gguf_artifacts": [ "fp16", "w4a16" ], "base_model": "Qwen/Qwen3-ASR-1.7B", "license": "apache-2.0", "tasks": [ "automatic-speech-recognition" ], "sample_image": null, "created_at": "2026-07-14T09:34:38.000Z", "last_modified": "2026-07-22T11:55:46.000Z", "hardware_configs": [ { "id": "RCAR-X5H", "label": "R-Car X5H (NPX6-48K)", "targets": [ "NPU" ], "notes": "vendor=renesas; chip=rcar-x5h; cpu=arm-cortex-a720; npu=npx6-48k", "metrics": { "artifacts": { "fp16": { "estimate": null, "hil": { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 12.23, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 12.23, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "NPX_Toolkit", "execution_provider": "npu", "last_updated": "2026-07-22", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "setup": "R-Car X5H (NPX6-48K)" }, "runs": [ { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 12.23, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 12.23, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "NPX_Toolkit", "execution_provider": "npu", "last_updated": "2026-07-22", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "setup": "R-Car X5H (NPX6-48K)", "kind": "hil" } ] }, "w4a16": { "estimate": null, "hil": { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 20.71, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 20.71, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "NPX_Toolkit", "execution_provider": "npu", "last_updated": "2026-07-22", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "setup": "R-Car X5H (NPX6-48K)" }, "runs": [ { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 20.71, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 20.71, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "NPX_Toolkit", "execution_provider": "npu", "last_updated": "2026-07-22", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "setup": "R-Car X5H (NPX6-48K)", "kind": "hil" } ] } } } } ] }, { "key": "llama-3.2-3b-instruct", "display_name": "Llama-3.2-3B-Instruct", "architecture": "Llama-3.2-3B", "model_size_m": 3000, "modality": "LLM", "variant": "Instruct", "targets": [ "NPU" ], "onnx_repo": "Renesas/Llama-3.2-3B-Instruct-ONNX", "gguf_repo": "Renesas/Llama-3.2-3B-Instruct-GGUF", "onnx_artifacts": [ "fp16", "w4a16", "fp32" ], "gguf_artifacts": [ "fp16", "w4a16" ], "base_model": "meta-llama/Llama-3.2-3B-Instruct", "license": "llama3.2", "tasks": [ "text-generation", "instruction-following", "chat" ], "sample_image": null, "created_at": "2026-05-11T09:18:02.000Z", "last_modified": "2026-07-02T15:01:25.000Z", "hardware_configs": [ { "id": "RCAR-X5H", "label": "R-Car X5H (NPX6-48K)", "targets": [ "NPU" ], "notes": "vendor=renesas; chip=rcar-x5h; cpu=arm-cortex-a720; npu=npx6-48k", "metrics": { "artifacts": { "fp16": { "estimate": { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 11.33, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 76175.36 }, "total": { "tok_s": 11.33, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 76175.36 }, "engine": "ppa-estimator", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "memory_bandwidth": "high", "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 } }, "hil": { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 8.76, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 8.76, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "NPX_Toolkit", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "setup": "R-Car X5H (NPX6-48K)" }, "runs": [ { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 8.76, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 8.76, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "NPX_Toolkit", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "setup": "R-Car X5H (NPX6-48K)", "kind": "hil" }, { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 11.33, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 76175.36 }, "total": { "tok_s": 11.33, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 76175.36 }, "engine": "ppa-estimator", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "memory_bandwidth": "high", "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "kind": "estimate" }, { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 7.3, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 49121.28 }, "total": { "tok_s": 7.3, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 49121.28 }, "engine": "ppa-estimator", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "memory_bandwidth": "low", "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "kind": "estimate" } ] }, "w4a16": { "estimate": { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 38.1, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 67737.6 }, "total": { "tok_s": 38.1, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 67737.6 }, "engine": "ppa-estimator", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "memory_bandwidth": "high", "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 } }, "hil": { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 17.46, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 17.46, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "NPX_Toolkit", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "setup": "R-Car X5H (NPX6-48K)" }, "runs": [ { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 17.46, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 17.46, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "NPX_Toolkit", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "setup": "R-Car X5H (NPX6-48K)", "kind": "hil" }, { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 38.1, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 67737.6 }, "total": { "tok_s": 38.1, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 67737.6 }, "engine": "ppa-estimator", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "memory_bandwidth": "high", "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "kind": "estimate" }, { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 26.04, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 46223.36 }, "total": { "tok_s": 26.04, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 46223.36 }, "engine": "ppa-estimator", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "memory_bandwidth": "low", "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "kind": "estimate" } ] }, "fp32": { "estimate": null, "hil": null, "runs": [] } } } } ] }, { "key": "qwen2.5-vl-3b-instruct", "display_name": "Qwen2.5-VL-3B-Instruct", "architecture": "Qwen2.5-VL-3B", "model_size_m": 3000, "modality": "VLM", "variant": "Instruct", "targets": [ "NPU" ], "onnx_repo": "", "gguf_repo": "Renesas/Qwen2.5-VL-3B-Instruct-GGUF", "onnx_artifacts": [], "gguf_artifacts": [ "fp16", "w4a16" ], "base_model": "Qwen/Qwen2.5-VL-3B-Instruct", "license": "other", "tasks": [ "image-text-to-text", "text-generation", "instruction-following", "chat" ], "sample_image": null, "created_at": "2026-05-13T09:09:11.000Z", "last_modified": "2026-07-02T15:03:06.000Z", "hardware_configs": [ { "id": "RCAR-X5H", "label": "R-Car X5H (NPX6-48K)", "targets": [ "NPU" ], "notes": "vendor=renesas; chip=rcar-x5h; cpu=arm-cortex-a720; npu=npx6-48k", "metrics": { "artifacts": { "fp16": { "estimate": null, "hil": { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 8.21, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 8.21, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "NPX_Toolkit", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "setup": "R-Car X5H (NPX6-48K)" }, "runs": [ { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 8.21, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 8.21, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "NPX_Toolkit", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "setup": "R-Car X5H (NPX6-48K)", "kind": "hil" } ] }, "w4a16": { "estimate": null, "hil": { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 14.09, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 14.09, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "NPX_Toolkit", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "setup": "R-Car X5H (NPX6-48K)" }, "runs": [ { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 14.09, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 14.09, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "NPX_Toolkit", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "setup": "R-Car X5H (NPX6-48K)", "kind": "hil" } ] } } } } ] }, { "key": "qwen3-vl-4b-instruct", "display_name": "Qwen3-VL-4B-Instruct", "architecture": "Qwen3-VL-4B", "model_size_m": 4000, "modality": "VLM", "variant": "Instruct", "targets": [ "NPU" ], "onnx_repo": "", "gguf_repo": "Renesas/Qwen3-VL-4B-Instruct-GGUF", "onnx_artifacts": [], "gguf_artifacts": [ "fp16", "w4a16" ], "base_model": "Qwen/Qwen3-VL-4B-Instruct", "license": "apache-2.0", "tasks": [ "image-text-to-text", "text-generation", "instruction-following", "chat" ], "sample_image": null, "created_at": "2026-05-15T09:37:33.000Z", "last_modified": "2026-07-02T15:20:23.000Z", "hardware_configs": [ { "id": "RCAR-X5H", "label": "R-Car X5H (NPX6-48K)", "targets": [ "NPU" ], "notes": "vendor=renesas; chip=rcar-x5h; cpu=arm-cortex-a720; npu=npx6-48k", "metrics": { "artifacts": { "fp16": { "estimate": null, "hil": { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 6.29, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 6.29, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "NPX_Toolkit", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "setup": "R-Car X5H (NPX6-48K)" }, "runs": [ { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 6.29, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 6.29, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "NPX_Toolkit", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "setup": "R-Car X5H (NPX6-48K)", "kind": "hil" } ] }, "w4a16": { "estimate": null, "hil": { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 11.48, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 11.48, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "NPX_Toolkit", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "setup": "R-Car X5H (NPX6-48K)" }, "runs": [ { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 11.48, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 11.48, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "NPX_Toolkit", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "setup": "R-Car X5H (NPX6-48K)", "kind": "hil" } ] } } } } ] }, { "key": "llama-3.1-8b-instruct", "display_name": "Llama-3.1-8B-Instruct", "architecture": "Llama-3.1-8B", "model_size_m": 8000, "modality": "LLM", "variant": "Instruct", "targets": [ "NPU" ], "onnx_repo": "Renesas/Llama-3.1-8B-Instruct-ONNX", "gguf_repo": "Renesas/Llama-3.1-8B-Instruct-GGUF", "onnx_artifacts": [ "w4a16", "fp32" ], "gguf_artifacts": [ "w4a16", "fp16" ], "base_model": "meta-llama/Llama-3.1-8B-Instruct", "license": "llama3.1", "tasks": [ "text-generation", "instruction-following", "chat" ], "sample_image": null, "created_at": "2026-06-12T09:43:03.000Z", "last_modified": "2026-07-02T13:32:01.000Z", "hardware_configs": [ { "id": "RCAR-X5H", "label": "R-Car X5H (NPX6-48K)", "targets": [ "NPU" ], "notes": "vendor=renesas; chip=rcar-x5h; cpu=arm-cortex-a720; npu=npx6-48k", "metrics": { "artifacts": { "fp16": { "estimate": null, "hil": null, "runs": [] }, "w4a16": { "estimate": { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 17.92, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 71823.36 }, "total": { "tok_s": 17.92, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 71823.36 }, "engine": "ppa-estimator", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "memory_bandwidth": "high", "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 } }, "hil": { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 8.59, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 8.59, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "NPX_Toolkit", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "setup": "R-Car X5H (NPX6-48K)" }, "runs": [ { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 8.59, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "total": { "tok_s": 8.59, "fps": null, "latency_ms_p50": null, "peak_mem_mb": null }, "engine": "NPX_Toolkit", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "setup": "R-Car X5H (NPX6-48K)", "kind": "hil" }, { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 17.92, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 71823.36 }, "total": { "tok_s": 17.92, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 71823.36 }, "engine": "ppa-estimator", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "memory_bandwidth": "high", "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "kind": "estimate" }, { "accuracy": null, "top5_accuracy": null, "npu": { "tok_s": 11.82, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 46868.48 }, "total": { "tok_s": 11.82, "fps": null, "latency_ms_p50": null, "peak_mem_mb": 46868.48 }, "engine": "ppa-estimator", "execution_provider": "npu", "last_updated": "2026-07-02", "batch_size": 1, "memory_bandwidth": "low", "utilization": { "instances_used": 1, "instances_total": 2, "cores_used": 12, "cores_total": 24, "freq_used_mhz": 1066, "freq_total_mhz": 1066, "tops_used": 105, "tops_total": 210 }, "kind": "estimate" } ] }, "fp32": { "estimate": null, "hil": null, "runs": [] } } } } ] } ] }