_wandb:
    value:
        cli_version: 0.28.1
        e:
            a7hek814eoi9gsa23a72ksfjkbmksup7:
                args:
                    - -of
                    - /data/spatop/tttt_pilot/configs/tttt_vanilla.json
                    - -l
                    - /scratch/logs
                    - -n
                    - tttt_pilot_vanilla
                cpu_count: 64
                cpu_count_logical: 64
                cudaVersion: "13.2"
                disk:
                    /:
                        total: "1919959433216"
                        used: "1424506851328"
                email: danielprimosch@gmail.com
                executable: /opt/conda/bin/python
                gpu: NVIDIA A100 80GB PCIe
                gpu_count: 1
                gpu_nvidia:
                    - architecture: Ampere
                      cudaCores: 6912
                      memoryTotal: "85899345920"
                      name: NVIDIA A100 80GB PCIe
                      uuid: GPU-79e9234d-ec7c-e59c-b425-edaeb1d971af
                host: dp-tttt-train-vanilla-bbbs6
                memory:
                    total: "540455968768"
                os: Linux-6.8.0-136-generic-x86_64-with-glibc2.36
                program: -m spanet.train
                python: CPython 3.12.10
                root: /scratch/logs
                startedAt: "2026-08-10T18:41:36.053118Z"
                writerId: a7hek814eoi9gsa23a72ksfjkbmksup7
        m:
            - "1": trainer/global_step
              "6":
                - 3
              "7": []
            - "2": '*'
              "5": 1
              "6":
                - 1
              "7": []
        python_version: 3.12.10
        t:
            "1":
                - 1
                - 5
                - 9
                - 53
                - 103
            "2":
                - 1
                - 5
                - 9
                - 53
                - 103
            "3":
                - 7
                - 13
                - 41
                - 66
            "4": 3.12.10
            "5": 0.28.1
            "12": 0.28.1
            "13": linux-x86_64
assignment_loss_scale:
    value: 1
balance_classifications:
    value: false
balance_jets:
    value: 0
balance_losses:
    value: true
balance_particles:
    value: 0
batch_size:
    value: 256
classification_loss_scale:
    value: 0
combinatorial_scale:
    value: 0
combine_pair_loss:
    value: min
dataset_limit:
    value: 1
dataset_randomization:
    value: 0
detection_loss_scale:
    value: 1
dropout:
    value: 0.3
epochs:
    value: 60
event_info_file:
    value: /data/spatop/event_files/tttt/tttt_hadronic.yaml
focal_gamma:
    value: 0
gradient_clip:
    value: 10
hidden_dim:
    value: 64
initial_embedding_dim:
    value: 64
initial_embedding_skip_connections:
    value: 1
kl_loss_scale:
    value: 0
l2_penalty:
    value: 0.0002
learning_rate:
    value: 0.00073
learning_rate_cycles:
    value: 1
learning_rate_warmup_epochs:
    value: 2
limit_to_num_jets:
    value: 0
limit_val_batches:
    value: 1
linear_activation:
    value: gelu
linear_block_type:
    value: GRU
linear_prelu_activation:
    value: true
mask_sequence_vectors:
    value: 1
masking:
    value: Filling
mass_classes:
    value: []
moe_loss_scale:
    value: 0.01
normalization:
    value: LayerNorm
normalize_features:
    value: 1
num_attention_heads:
    value: 8
num_branch_embedding_layers:
    value: 3
num_branch_encoder_layers:
    value: 3
num_classification_layers:
    value: 3
num_dataloader_workers:
    value: 0
num_detector_layers:
    value: 2
num_embedding_layers:
    value: 10
num_encoder_layers:
    value: 4
num_experts:
    value: 8
num_experts_per_tok:
    value: 2
num_gpu:
    value: 1
num_jet_embedding_layers:
    value: 0
num_jet_encoder_layers:
    value: 2
num_pairwise_features:
    value: 4
num_regression_layers:
    value: 3
optimizer:
    value: AdamW
pairwise_embedding_dim:
    value: 8
pairwise_input_source:
    value: ""
partial_events:
    value: 1
position_embedding_dim:
    value: 64
regression_loss_scale:
    value: 0
skip_connections:
    value: 1
split_symmetric_attention:
    value: 1
testing_file:
    value: ""
train_validation_split:
    value: 0.9
training_file:
    value: /data/spatop/tttt_pilot/h5/tttt_training.h5
transformer_activation:
    value: gelu
transformer_dim:
    value: 64
transformer_dim_scale:
    value: 2
transformer_type:
    value: Gated
trial_output_dir:
    value: ./test_output
trial_time:
    value: ""
usable_gpus:
    value: ""
use_mass_classification:
    value: false
use_moe:
    value: false
use_pairwise_interactions:
    value: false
validation_file:
    value: ""
verbose_output:
    value: false