b54e99fb4c6b3d513c4eecc11b27fadafef91a4a,agents/dagger/agent.py,Agent,get_next_action,#Agent#Any#Any#,112

Before Change


        desired_throttle = min(max(desired_throttle, 0.), 1.)

        log.info("desired_steering %f", desired_steering)
        log.debug("desired_throttle %f", desired_throttle)
        if self.previous_action:
            smoothed_steering = 0.2 * self.previous_action.steering + 0.5 * desired_steering
        else:
            smoothed_steering = desired_steering * 0.7

After Change


            log.debug("net out is None")
            return self.previous_action or Action()

        net_out = net_out[0]  // We currently only have one environment

        desired_spin, desired_direction, desired_speed, desired_speed_change, desired_steering, desired_throttle = \
            net_out

        desired_spin = desired_spin * c.SPIN_NORMALIZATION_FACTOR
        desired_speed = desired_speed * c.SPEED_NORMALIZATION_FACTOR
        desired_speed_change = desired_speed_change * c.SPEED_NORMALIZATION_FACTOR

        log.debug("desired_steering %f", desired_steering)
        log.debug("desired_throttle %f", desired_throttle)

        log.debug("desired_direction %f", desired_direction)
        log.debug("desired_speed %f", desired_speed)
        log.debug("desired_speed_change %f", desired_speed_change)
        log.debug("desired_throttle %f", desired_throttle)
        log.debug("desired_spin %f", desired_spin)

        actual_speed = obz["speed"]
        log.debug("actual_speed %f", actual_speed)
        log.debug("desired_speed %f", desired_speed)

        if isinstance(self.net, MobileNetV2):
            // target_speed = 8 * 100
            target_speed = desired_speed
            // desired_throttle = abs(target_speed / max(actual_speed, 1e-3))
            // desired_throttle = min(max(desired_throttle, 0.), 1.)
            target_speed = 8 * 100
            desired_throttle = abs(target_speed / max(actual_speed, 1e-3))
            desired_throttle = min(max(desired_throttle, 0.), 1.)

            // if self.previous_net_out:
            //     desired_throttle = 0.2 * self.previous_action.throttle + 0.7 * desired_throttle
            // else:
            // desired_throttle = desired_throttle * 0.95

        else:
            // AlexNet
            target_speed = 9 * 100
            // Network overfit on speed, plus it"s nice to be able to change it,
            // so we just ignore output speed of net
            desired_throttle = abs(target_speed / max(actual_speed, 1e-3))
            desired_throttle = min(max(desired_throttle, 0.), 1.)
        log.debug("actual_speed %r" % actual_speed)

        // log.info("desired_steering %f", desired_steering)
        // log.info("desired_throttle %f", desired_throttle)
Italian Trulli
In pattern: SUPERPATTERN

Frequency: 3

Non-data size: 6

Instances


Project Name: deepdrive/deepdrive
Commit Name: b54e99fb4c6b3d513c4eecc11b27fadafef91a4a
Time: 2018-05-29
Author: cquiter@gmail.com
File Name: agents/dagger/agent.py
Class Name: Agent
Method Name: get_next_action


Project Name: vatlab/SoS
Commit Name: c8788d2eedcdb2671289d7d47a41b8fdcb0294f1
Time: 2017-09-11
Author: ben.bog@gmail.com
File Name: src/sos/sos_executor.py
Class Name: Base_Executor
Method Name: resolve_dangling_targets


Project Name: deepfakes/faceswap
Commit Name: bcf38b02cc7209d1baccd1302b5224f5faf2f00a
Time: 2021-01-31
Author: 36920800+torzdf@users.noreply.github.com
File Name: plugins/train/model/_base.py
Class Name: _Inference
Method Name: _make_inference_model