Merge pull request #24 from zhaoyi11/rp1m

kevinzakka · web-flow · commit 0d9736c64eba · 2024-11-02T15:36:07.000-07:00
Lift the requirement of human fingering with RP1M
diff --git a/robopianist/suite/tasks/piano_with_shadow_hands.py b/robopianist/suite/tasks/piano_with_shadow_hands.py
@@ -17,6 +17,7 @@
 from typing import List, Optional, Sequence, Tuple
 
 import numpy as np
+from scipy.optimize import linear_sum_assignment
 from dm_control import mjcf
 from dm_control.composer import variation as base_variation
 from dm_control.composer.observation import observable
@@ -134,6 +135,11 @@ def _set_rewards(self) -> None:
         )
         if not self._disable_fingering_reward:
             self._reward_fn.add("fingering_reward", self._compute_fingering_reward)
+        else:
+            # use OT based fingering
+            print('Fingering is unavailable. OT fingering reward is used.')
+            self._reward_fn.add("ot_fingering_reward", self._compute_ot_fingering_reward)
+
         if not self._disable_forearm_reward:
             self._reward_fn.add("forearm_reward", self._compute_forearm_reward)
 
@@ -324,6 +330,44 @@ def _distance_finger_to_key(
         )
         return float(np.mean(rews))
 
+    def _compute_ot_fingering_reward(self, physics: mjcf.Physics) -> float:
+        """ OT reward calculation from RP1M https://arxiv.org/abs/2408.11048 """
+        # calcuate fingertip positions
+        fingertip_pos = [physics.bind(finger).xpos.copy() for finger in self.left_hand.fingertip_sites]
+        fingertip_pos += [physics.bind(finger).xpos.copy() for finger in self.right_hand.fingertip_sites]
+        
+        # calcuate the positions of piano keys to press.
+        keys_to_press = np.flatnonzero(self._goal_current[:-1]) # keys to press
+        # if no key is pressed
+        if keys_to_press.shape[0] == 0:
+            return 1.
+
+        # calculate key pos
+        key_pos = []
+        for key in keys_to_press:
+            key_geom = self.piano.keys[key].geom[0]
+            key_geom_pos = physics.bind(key_geom).xpos.copy()
+            key_geom_pos[-1] += 0.5 * physics.bind(key_geom).size[2]
+            key_geom_pos[0] += 0.35 * physics.bind(key_geom).size[0]
+            key_pos.append(key_geom_pos.copy())
+
+        # calcualte the distance between keys and fingers
+        dist = np.full((len(fingertip_pos), len(key_pos)), 100.)
+        for i, finger in enumerate(fingertip_pos):
+            for j, key in enumerate(key_pos):
+                dist[i, j] = np.linalg.norm(key - finger)
+        
+        # calculate the shortest distance
+        row_ind, col_ind = linear_sum_assignment(dist)
+        dist = dist[row_ind, col_ind]
+        rews = tolerance(
+            dist,
+            bounds=(0, _FINGER_CLOSE_ENOUGH_TO_KEY),
+            margin=(_FINGER_CLOSE_ENOUGH_TO_KEY * 10),
+            sigmoid="gaussian",
+        )
+        return float(np.mean(rews))        
+
     def _update_goal_state(self) -> None:
         # Observable callables get called after `after_step` but before
         # `should_terminate_episode`. Since we increment `self._t_idx` in `after_step`,
