Instructions to use molkab/dashboard with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Keras
How to use molkab/dashboard with Keras:
# Available backend options are: "jax", "torch", "tensorflow". import os os.environ["KERAS_BACKEND"] = "jax" import keras model = keras.saving.load_model("hf://molkab/dashboard") - Notebooks
- Google Colab
- Kaggle
| { | |
| "economies": { | |
| "1 — Sleep mode secteur": { | |
| "economie_kwh": 157190.58, | |
| "economie_pct": 9.24, | |
| "co2_evite_t": 83.31, | |
| "economie_dt": 62876.23, | |
| "n_eligible": 81755, | |
| "couverture_pct": 15.6, | |
| "objectif_10pct": false | |
| }, | |
| "2 — Réduction puissance": { | |
| "economie_kwh": 50573.1, | |
| "economie_pct": 2.97, | |
| "co2_evite_t": 26.8, | |
| "economie_dt": 20229.24, | |
| "n_eligible": 139256, | |
| "couverture_pct": 26.5, | |
| "objectif_10pct": false | |
| }, | |
| "3 — Free cooling": { | |
| "economie_kwh": 23500.67, | |
| "economie_pct": 1.38, | |
| "co2_evite_t": 12.46, | |
| "economie_dt": 9400.27, | |
| "n_eligible": 47626, | |
| "couverture_pct": 9.1, | |
| "objectif_10pct": false | |
| }, | |
| "4 — Mode éco calendaire": { | |
| "economie_kwh": 1214.72, | |
| "economie_pct": 0.07, | |
| "co2_evite_t": 0.64, | |
| "economie_dt": 485.89, | |
| "n_eligible": 5259, | |
| "couverture_pct": 1.0, | |
| "objectif_10pct": false | |
| }, | |
| "5 — Combinée (1+2+3+4)": { | |
| "economie_kwh": 232479.07, | |
| "economie_pct": 13.67, | |
| "co2_evite_t": 123.21, | |
| "economie_dt": 92991.63, | |
| "objectif_10pct": true | |
| }, | |
| "5 — Alerte saturation 2G (surveillance)": { | |
| "n_observations": 11035, | |
| "couverture_pct": 2.1, | |
| "economie_kwh": 0, | |
| "action": "migration_3G_4G_ou_extension_capacite" | |
| }, | |
| "baseline_kwh": 1700808.91, | |
| "baseline_co2_t": 901.43, | |
| "baseline_dt": 680323.56 | |
| }, | |
| "kpi_reseau": { | |
| "nb_stations": 66, | |
| "nb_mesures": 525706, | |
| "conso_totale_kwh": 1700808.91, | |
| "conso_moyenne_kwh": 3.2353, | |
| "eei_moyen": 103.36, | |
| "score_qos_moyen": 0.7614, | |
| "pct_mode_eco": 26.7, | |
| "pct_mode_critique": 3.01, | |
| "pct_anomalies": 6.47, | |
| "n_alertes_2g": 11035, | |
| "economie_combinee_pct": 13.67, | |
| "economie_combinee_kwh": 232479.0, | |
| "co2_evite_t": 123.2, | |
| "economie_dt": 92992.0, | |
| "meilleur_agent_rl": "Q-Learning UCB", | |
| "economie_rl_pct": 17.36, | |
| "economie_rl_kwh": 73490.0, | |
| "rl_violations_qos": 0, | |
| "rl_agents_comparaison": { | |
| "Q-Learning": { | |
| "eco_pct": 9.31, | |
| "n_viols": 0 | |
| }, | |
| "SARSA": { | |
| "eco_pct": 14.5, | |
| "n_viols": 0 | |
| }, | |
| "Double Q-Learning": { | |
| "eco_pct": 15.3, | |
| "n_viols": 0 | |
| }, | |
| "Expected SARSA": { | |
| "eco_pct": 15.25, | |
| "n_viols": 0 | |
| }, | |
| "Q-Learning UCB": { | |
| "eco_pct": 17.36, | |
| "n_viols": 0 | |
| }, | |
| "SARSA(λ)": { | |
| "eco_pct": 15.2, | |
| "n_viols": 0 | |
| }, | |
| "Monte Carlo": { | |
| "eco_pct": 8.15, | |
| "n_viols": 0 | |
| } | |
| }, | |
| "generated_at": "2026-05-21T13:56:10.701085", | |
| "economie_rl_episodique_pct": 0.68, | |
| "ecart_statique_vs_episodique": -16.68, | |
| "note_baseline": "Baseline = consommation mesurée 2024 sans optimisation active. Ne tient pas compte des pratiques manuelles déjà en place." | |
| }, | |
| "couvertures": { | |
| "strat_1_sleep_pct": 15.6, | |
| "strat_2_reduction_pct": 26.5, | |
| "strat_3_free_cooling_pct": 16.1, | |
| "strat_4_eco_calendaire_pct": 1.0, | |
| "strat_5_alerte_2g_pct": 2.1 | |
| }, | |
| "rl_resultats_tous_agents": { | |
| "Q-Learning": { | |
| "economie_pct": 9.3074, | |
| "economie_kwh": 39410.43, | |
| "n_violations": 0, | |
| "pct_violations": 0.0, | |
| "class_name": "AgentQLearning", | |
| "reference": "Watkins & Dayan (1992)", | |
| "is_best": false | |
| }, | |
| "SARSA": { | |
| "economie_pct": 14.4981, | |
| "economie_kwh": 61389.35, | |
| "n_violations": 0, | |
| "pct_violations": 0.0, | |
| "class_name": "AgentSARSA", | |
| "reference": "Rummery & Niranjan (1994)", | |
| "is_best": false | |
| }, | |
| "Double Q-Learning": { | |
| "economie_pct": 15.3014, | |
| "economie_kwh": 64790.82, | |
| "n_violations": 0, | |
| "pct_violations": 0.0, | |
| "class_name": "AgentDoubleQLearning", | |
| "reference": "van Hasselt (2010)", | |
| "is_best": false | |
| }, | |
| "Expected SARSA": { | |
| "economie_pct": 15.2459, | |
| "economie_kwh": 64555.67, | |
| "n_violations": 0, | |
| "pct_violations": 0.0, | |
| "class_name": "AgentExpectedSARSA", | |
| "reference": "van Seijen et al. (2009)", | |
| "is_best": false | |
| }, | |
| "Q-Learning UCB": { | |
| "economie_pct": 17.3558, | |
| "economie_kwh": 73489.6, | |
| "n_violations": 0, | |
| "pct_violations": 0.0, | |
| "class_name": "AgentQLearningUCB", | |
| "reference": "Auer et al. (2002)", | |
| "is_best": true | |
| }, | |
| "SARSA(λ)": { | |
| "economie_pct": 15.1972, | |
| "economie_kwh": 64349.72, | |
| "n_violations": 0, | |
| "pct_violations": 0.0, | |
| "class_name": "AgentSARSALambda", | |
| "reference": "Sutton (1988)", | |
| "is_best": false | |
| }, | |
| "Monte Carlo": { | |
| "economie_pct": 8.151, | |
| "economie_kwh": 34513.93, | |
| "n_violations": 0, | |
| "pct_violations": 0.0, | |
| "class_name": "AgentMonteCarlo", | |
| "reference": "Sutton & Barto (1998)", | |
| "is_best": false | |
| } | |
| }, | |
| "meilleur_agent": "Q-Learning UCB", | |
| "top_eco_agent": "Q-Learning UCB", | |
| "n_episodes": 1000, | |
| "generated_at": "2026-05-21T13:56:19.813317" | |
| } |