TeleopWM / benchmark.json
awskhalil's picture
Upload folder using huggingface_hub
607693f verified
{
"action_dim": 3,
"aux_dynamics_hidden_dim": null,
"batch_size": 1,
"checkpoint": "outputs/simvp_baseline/2026-05-25_08-44-43_full5ep_motionctxV2_grid2x4_fmScale3_dualWM10_delta05_MSE_dLong05_dSteer15_corr01_noDetach_conv1x1_ctrlSteerS030/checkpoints/best.pt",
"conditioning_dim": 32,
"conditioning_fusion": null,
"conditioning_injection": "single",
"conditioning_representation": "longitudinal_steer_speed",
"control_steer_input_scale": 0.3,
"device": "cuda",
"dual_conv1x1_fusion_stats": {
"bias_norm": 0.050001209756022555,
"enabled": true,
"param_count": 2080,
"simvp_half_weight_norm": 34.854761816360536,
"weight_norm": 35.060819933559266,
"wm_half_weight_norm": 3.7956123535264505,
"wm_to_simvp_weight_norm_ratio": 0.10889795700009121
},
"dual_fusion": "conv1x1",
"dual_wm_conditioning": "film",
"dual_wm_gated": true,
"dual_wm_hidden_dim": 512,
"dual_wm_num_layers": 3,
"dual_wm_scale": 1.0,
"future_action_detach_latents": false,
"future_action_future_motion_scale": 3.0,
"future_action_head_variant": "motion_context_v2",
"future_action_source": "wm",
"future_action_spatial_grid": [
2,
4
],
"future_action_spatial_pooling": "grid",
"future_action_token_dim": 256,
"future_frames_per_second": 205.4795994525576,
"has_future_action_cls_heads": true,
"has_future_action_head": true,
"height": 320,
"iters": 200,
"latency_ms_per_batch": 38.93330540508032,
"model_variant": "av_wm_dual_bigwm",
"normalize_controls": true,
"parameters": 11452811,
"peak_vram_mb": 1243.41455078125,
"pred_future_actions_shape": [
1,
8,
2
],
"samples_per_second": 25.6849499315697,
"simvp_conditioning": "none",
"simvp_conditioning_stage": "input",
"speed_dim": 1,
"speed_scale": 20.0,
"warmup": 20,
"width": 512,
"wm_latent_residual": false,
"wm_residual_gated": true,
"wm_residual_hidden_dim": 128,
"wm_residual_scale": 0.1
}