Upload clic/hits/v3.1.0/pyg-clic-hits-v1_clic_20260328_144021_479374/hyperparameters.json with huggingface_hub

clic/hits/v3.1.0/pyg-clic-hits-v1_clic_20260328_144021_479374/hyperparameters.json

@@ -0,0 +1 @@
+{"num_mlpf_params": 3247890, "dataset": "clic_hits", "data_dir": "/local/joosep/mlpf/clic/v1.2.5_key4hep_2025-05-29/tfds/", "model": {"type": "attention", "input_encoding": "split", "learned_representation_mode": "last", "pt_mode": "direct-elemtype-split", "eta_mode": "linear", "sin_phi_mode": "linear", "cos_phi_mode": "linear", "energy_mode": "direct-elemtype-split", "trainable": "all", "gnn_lsh": null, "attention": {"conv_type": "attention", "embedding_dim": 128, "width": 128, "num_convs": 3, "dropout_ff": 0.0, "activation": "elu", "layernorm": true, "num_heads": 16, "head_dim": 16, "attention_type": "flash", "dropout_conv_reg_mha": 0.0, "dropout_conv_reg_ff": 0.0, "dropout_conv_id_mha": 0.0, "dropout_conv_id_ff": 0.0, "use_pre_layernorm": true, "use_simplified_attention": false, "export_onnx_fused": false, "save_attention": false}}, "conv_type": "attention", "batch_size": 1, "hyperparameters": {"batch_size": 1, "lr": 0.0001}, "backend": "pytorch", "threads": 16, "gpu_type": "l40", "mem_per_gpu_mb": 80000, "slurm_partition": "gpu", "slurm_runtime": "168h", "input_dim": 12, "num_classes": 6, "elemtypes_nonzero": [1, 2], "num_steps": 100000, "patience": 10000, "checkpoint_freq": 10000, "val_freq": 10000, "num_workers": 4, "prefetch_factor": 2, "gpu_batch_multiplier": 16, "dtype": "bfloat16", "lr": 0.0001, "weight_decay": 0.01, "optimizer": "adamw", "lr_schedule": "cosinedecay", "lr_schedule_config": {}, "pad_to_multiple_elements": null, "train": true, "test": true, "compile": false, "make_plots": false, "sort_data": false, "load": null, "relaxed_load": true, "comet": false, "comet_offline": false, "comet_name": "particleflow", "comet_step_freq": 10000, "raytune": {"local_dir": null, "sched": "asha", "search_alg": null, "default_metric": "val_loss", "default_mode": "min", "asha": {"max_t": 200, "reduction_factor": 4, "brackets": 1, "grace_period": 10}, "hyperband": {"max_t": 200, "reduction_factor": 4}, "hyperopt": {"n_random_steps": 10}, "nevergrad": {"n_random_steps": 10}}, "train_dataset": {"clic_hits": {"physical": {"batch_size": 1, "samples": {"clic_edm_ttbar_hits": {"version": "3.1.0", "splits": ["1", "2", "3", "4", "5", "6", "7", "8", "9", "10"], "batch_size": null}, "clic_edm_ww_fullhad_hits": {"version": "3.1.0", "splits": ["1", "2", "3", "4", "5", "6", "7", "8", "9", "10"], "batch_size": null}, "clic_edm_qq_hits": {"version": "3.1.0", "splits": ["1", "2", "3", "4", "5", "6", "7", "8", "9", "10"], "batch_size": null}}}}}, "valid_dataset": {"clic_hits": {"physical": {"batch_size": 1, "samples": {"clic_edm_ttbar_hits": {"version": "3.1.0", "splits": ["1", "2", "3", "4", "5", "6", "7", "8", "9", "10"], "batch_size": null}, "clic_edm_ww_fullhad_hits": {"version": "3.1.0", "splits": ["1", "2", "3", "4", "5", "6", "7", "8", "9", "10"], "batch_size": null}, "clic_edm_qq_hits": {"version": "3.1.0", "splits": ["1", "2", "3", "4", "5", "6", "7", "8", "9", "10"], "batch_size": null}}}}}, "test_dataset": {"clic_edm_ttbar_hits": {"version": "3.1.0", "splits": ["1", "2", "3", "4", "5", "6", "7", "8", "9", "10"], "batch_size": 1}, "clic_edm_ww_fullhad_hits": {"version": "3.1.0", "splits": ["1", "2", "3", "4", "5", "6", "7", "8", "9", "10"], "batch_size": 1}, "clic_edm_qq_hits": {"version": "3.1.0", "splits": ["1", "2", "3", "4", "5", "6", "7", "8", "9", "10"], "batch_size": 1}}, "enabled_test_datasets": ["clic_edm_ttbar_hits", "clic_edm_ww_fullhad_hits", "clic_edm_qq_hits"], "ntrain": null, "nvalid": null, "ntest": 100, "data_config": null, "gpus": 1}