obs, action_set, reward, done, info = env.reset(instance) returns 'None' in obs, action_set and 'True' in done variables in set cover problem.

[azraaziz00]

import ecole env = ecole.environment.Branching( observation_function=ecole.observation.NodeBipartite(),) reward_function=-ecole.reward.NNodes())

set up an instance generator

instances = ecole.instance.SetCoverGenerator(n_rows=100, n_cols=1000, density=0.05, max_coef=100)

generate ten MDP episodes

for i in range(10): instance = next(instances) obs, action_set, reward, done, info = env.reset(instance)

Why does the env.reset returns None in obs and action_set and True in done variables?

Although, it works well with the instances generated with CombinatorialAuctionGenerator function i.e.,

import ecole env = ecole.environment.Branching( observation_function=ecole.observation.NodeBipartite(),) reward_function=-ecole.reward.NNodes())

set up an instance generator

instances = ecole.instance.CombinatorialAuctionGenerator (n_items=100, n_bids=100)

generate ten MDP episodes

for i in range(10): instance = next(instances) obs, action_set, reward, done, info = env.reset(instance)

Where am I going wrong?