Ammar N. Abbas

Referring to "https://web.ece.ucsb.edu/Faculty/Rabiner/ece259/Reprints/tutorial%20on%20hmm%20and%20applications.pdf" and library "https://hmmlearn.readthedocs.io/en/latest/" I have found this solution: 1- Through log_gamma (posterior distribution): ``` state_sequences = [] for i in range(100): for j in range(lengths[i]): state_sequences.append(np.argmax(np.exp(SHMM.log_gammas[i])[j])) pred_state_seq...

The only problem is that with the synchronous mode, carla-ros bridge is too slow

@cambel thank you for your reply. The issue is resolved with your branch However, now it shows this error: ''' GYM Environment:simulation/task_space.yaml Grasping pluging can't be loaded ft_sensor: False, ee_link:...

@cambel thanks a lot this error is resolved now. Getting this attribute error `AttributeError: module 'ur_control.transformations' has no attribute 'pose_from_angular_veloticy'`

@cambel you are right, it was due to tensorflow and tensorflow_probability compatibility issues it's working fine now. Thanks Thank you for sharing another work example, really interesting...

The problem is related to updating the critic and then the actor_loss which is based on critic. One solution might be to move the actor_loss after the critic update, but...

Similar issue faced by my team too