{ "key": "opt08_mint_quant/bits_sweep_fine", "tag": "bits_sweep_fine", "script": "opt08_mint_quant.py", "script_sha256_16": "dd139bbe634ea732", "hf_folder": "applied/opt08_mint_quant", "priority": 2, "note": "finer membrane-precision ladder in mint_sweep.json", "phase": 9, "n_phases": 10, "method": "harmonic_matched", "command": [ "/usr/bin/python3", "-m", "torch.distributed.run", "--standalone", "--nproc_per_node=2", "/kaggle/working/scripts/opt08_mint_quant.py", "--method", "harmonic_matched", "--data-root", "/kaggle/working/data", "--epochs", "60", "--seed", "1234", "--batch-size", "64", "--event-eval", "--no-resume", "--decision-margin", "2.0", "--hub-dir", "/kaggle/working/hub/opt08_mint_quant/bits_sweep_fine", "--keep-epoch-weights", "1", "--resplit", "block", "--drop-rest", "--bench", "--bits-sweep", "0,2,3,4,6,8" ], "flags": [ "--bits-sweep", "0,2,3,4,6,8" ], "epochs_requested": 60, "gpus": 2, "relative_epoch_cost": 1.0, "exit_code": 0, "status": "ok", "duration_s": 323.5, "started_utc": "2026-08-23T04:51:25.594890+00:00", "finished_utc": "2026-08-23T04:56:49.132449+00:00", "optimization_summary": { "optimization": "mint_quant", "research_section": "4. Membrane-Potential Memory-Wall Optimizations (multiplier-less integer quantization)", "reference": "Yin et al., MINT: Multiplier-less INTeger Quantization for Energy Efficient Spiking Neural Networks, arXiv:2305.09850 (ASP-DAC 2024)", "script": "opt08_mint_quant.py", "baseline_mode": false, "method": "harmonic_matched", "accuracy": { "test_acc": 0.22428571428571428, "test_target_acc": 0.22428571428571428, "best_val_acc": 0.49357142857142855 }, "efficiency": { "spikes_per_inference": null, "syn_ops_per_inference": 737711.1282467532, "dense_macs_per_inference": 122675200.0, "event_fraction_of_dense": 0.006013531082458013, "syn_ops_per_inference_dense_fanout": 63661272.290584415, "weight_sparsity": 0.9789982001252087, "effective_ticks": 27.373571428571427, "num_ticks": 100, "membrane_bytes_per_inference": 0, "weight_bytes": 613376, "input_raster_bytes": null, "train_ms_per_step": 88.68454749999728, "infer_ms_per_sample": 0.6045492706249433, "peak_vram_mb": 2835.119104 }, "optimization_specific": { "w_bits": 4, "u_bits": 2, "leak_shift": 0, "beta_eff": 1.0, "qat": true, "batchnorm": true, "membrane_memory_reduction_x": 16.0, "weight_memory_reduction_x": 8.0, "membrane_bytes_per_inference": 28200.0, "working_set_bytes": 613658.0, "baseline_working_set_bytes": 4911520.0, "l2_cache_fits": true, "baseline_l2_cache_fits": false, "l2_budget_bytes": 2097152, "membrane_saturation_frac": 0.018447992852393617, "acc_vs_ubits": { "0": 0.18428571428571427, "2": 0.22428571428571428, "3": 0.21571428571428572, "4": 0.21428571428571427, "6": 0.21571428571428572, "8": 0.21214285714285713 }, "bits_sweep_rows": [ { "u_bits": 0, "acc": 0.18428571428571427, "target_acc": 0.18428571428571427, "membrane_bytes_per_neuron_per_tick": 4.0, "membrane_memory_reduction_x": 1.0, "membrane_saturation_frac": null, "acc_drop_vs_fp32": 0.0 }, { "u_bits": 2, "acc": 0.22428571428571428, "target_acc": 0.22428571428571428, "membrane_bytes_per_neuron_per_tick": 0.25, "membrane_memory_reduction_x": 16.0, "membrane_saturation_frac": 0.016339829898049645, "acc_drop_vs_fp32": -0.04000000000000001 }, { "u_bits": 3, "acc": 0.21571428571428572, "target_acc": 0.21571428571428572, "membrane_bytes_per_neuron_per_tick": 0.375, "membrane_memory_reduction_x": 10.666666666666666, "membrane_saturation_frac": 0.07418640569592198, "acc_drop_vs_fp32": -0.031428571428571445 }, { "u_bits": 4, "acc": 0.21428571428571427, "target_acc": 0.21428571428571427, "membrane_bytes_per_neuron_per_tick": 0.5, "membrane_memory_reduction_x": 8.0, "membrane_saturation_frac": 0.1724607297207447, "acc_drop_vs_fp32": -0.03 }, { "u_bits": 6, "acc": 0.21571428571428572, "target_acc": 0.21571428571428572, "membrane_bytes_per_neuron_per_tick": 0.75, "membrane_memory_reduction_x": 5.333333333333333, "membrane_saturation_frac": 0.3710663923980496, "acc_drop_vs_fp32": -0.031428571428571445 }, { "u_bits": 8, "acc": 0.21214285714285713, "target_acc": 0.21214285714285713, "membrane_bytes_per_neuron_per_tick": 1.0, "membrane_memory_reduction_x": 4.0, "membrane_saturation_frac": 0.37882293051861704, "acc_drop_vs_fp32": -0.027857142857142858 } ], "fp32_membrane_acc": 0.18428571428571427, "per_layer_scales": [ { "layer": 0, "weight_scale": 0.03549237549304962, "membrane_scale": 1.0, "shared_scale_gauge_factor": 28.17506537976945 }, { "layer": 1, "weight_scale": 0.03008062924657549, "membrane_scale": 1.0, "shared_scale_gauge_factor": 33.24398541675601 }, { "layer": 2, "weight_scale": 0.027376641120229448, "membrane_scale": 1.0, "shared_scale_gauge_factor": 36.52749055694305 }, { "layer": 3, "weight_scale": 0.02388627827167511, "membrane_scale": 1.0, "shared_scale_gauge_factor": 41.865040196982996 } ], "shared_scale_note": "weight grid and membrane grid are the same grid up to a per-layer positive scalar that BatchNorm absorbs exactly; exact only with --batchnorm on", "multiplier_less_compare": true, "wall_clock_caveat": "this reference kernel computes in fp32 registers, so the traffic win does not appear in train_ms_per_step; a fixed-point core is where it becomes energy" }, "config": { "method": "harmonic_matched", "data_root": "/kaggle/working/data", "subjects": "", "blocks": "", "limit": 0, "drop_rest": true, "resplit": "block", "test_blocks": "6", "class_weight": false, "epochs": 60, "batch_size": 64, "lr": 0.002, "weight_decay": 0.0005, "grad_clip": 5.0, "num_bins": 100, "window_sec": 5.0, "hidden": "512,256", "freq_groups": 8, "dropout": 0.3, "spike_drop": 0.1, "time_jitter": 2, "label_smoothing": 0.05, "no_batchnorm": false, "binary_input": false, "beta": 1.0, "thresh": 1.0, "window": 0.5, "gain": 1.0, "alpha": 0.9, "readout": "spikecount", "fake_quant": 0, "amp": false, "spike_reg": 0.0, "event_eval": true, "decision_margin": 2.0, "export": false, "seed": 1234, "hub_dir": "/kaggle/working/hub/opt08_mint_quant/bits_sweep_fine", "repo_id": "UWU-R-13/SSVEP-SNN", "push_hf": false, "keep_epoch_weights": 1, "no_resume": true, "selftest": false, "baseline": false, "bench": true, "bench_warmup": 10, "bench_iters": 50, "bench_train_steps": 20, "w_bits": 4, "u_bits": 2, "leak_shift": 0, "qat": true, "bits_sweep": "0,2,3,4,6,8", "l2_bytes": 2097152, "sampling_rate": 250.0 } }, "final_summary": { "final_test_acc": 0.22428571428571428, "final_test_target_acc": 0.22428571428571428, "final_test_rest_recall": NaN, "best_acc": 0.49357142857142855, "best_epoch": 7, "resplit": "block", "class_weight": false, "drop_rest": true, "epochs": 60, "method": "harmonic_matched", "n_in": 2880, "n_classes": 40, "num_bins": 100 }, "event_driven_summary": { "acc": 0.22428571428571428, "anytime_acc": { "0.25": 0.11142857142857143, "0.5": 0.16214285714285714, "0.75": 0.19142857142857142, "1.0": 0.22428571428571428 }, "syn_per_inference": 737711.1282467532, "dense_per_inference": 122675200.0, "event_fraction_of_dense": 0.006013531082458013, "syn_per_inference_dense_fanout": 63661272.290584415, "weight_sparsity": 0.9789982001252087, "num_ticks": 100, "dt_ms": 50.0, "margin_decision_acc": 0.10714285714285714, "mean_decision_tick": 27.373571428571427, "mean_decision_ms": 1368.6785714285713, "dense_forward_test_acc": 0.22428571428571428 }, "extra_sweep_file_present": true, "selftest": { "mode": "gate", "passed": true }, "driver_version": 1, "zip": { "name": "bits_sweep_fine.zip", "files": 81, "bytes": 12373116, "MB": 12.4, "sha256": "2fe3f4aad883a9ccc3ebe2dd4259aaa6e4053f00e53278a241e2f51f70d48e6c" } }