fix(config): accept pretrained_model dir for --config_path on resume (#4023)

Co-authored-by: Steven Palma <imstevenpmwork@ieee.org>

docs/source/hil_data_collection.mdx

@@ -59,6 +59,7 @@ The `lerobot-rollout --strategy.type=dagger` mode requires **teleoperators with
 
 - `bi_openarm_mini` - Bimanual OpenArm Mini
 - `so_leader` - SO100 / SO101 leader arm
+- `bi_so_leader` - Bimanual SO100 / SO101 leader arms
 
 > [!IMPORTANT]
 > The provided commands default to `bi_openarm_follower` + `bi_openarm_mini`.

docs/source/so101.mdx

@@ -338,7 +338,7 @@ It is advisable to install one 3-pin cable in the motor after placing them befor
 <hfoption id="Leader">
 
 - Mount the leader holder onto the wrist and secure it with 4 M3x6mm screws.
-- Attach the handle to motor 5 using 1 M2x6mm screw.
+- Attach the handle to the leader holder using 1 M2x6mm screw.
 - Insert the gripper motor, secure it with 2 M2x6mm screws on each side, attach a motor horn using a M3x6mm horn screw.
 - Attach the follower trigger with 4 M3x6mm screws.
 

pyproject.toml

@@ -67,7 +67,7 @@ dependencies = [
     "einops>=0.8.0,<0.9.0",
 
     # Config & Hub
-    "draccus==0.10.0", # TODO: Relax version constraint
+    "draccus>=0.11.6,<0.12.0",
     "huggingface-hub>=1.0.0,<2.0.0",
     "requests>=2.32.0,<3.0.0",
 

src/lerobot/configs/policies.py

@@ -163,8 +163,10 @@ class PreTrainedConfig(draccus.ChoiceRegistry, HubMixin, abc.ABC):  # type: igno
         return None
 
     def _save_pretrained(self, save_directory: Path) -> None:
-        with open(save_directory / CONFIG_NAME, "w") as f, draccus.config_type("json"):
+        # Encode against the base class so draccus includes the choice "type" key,
-            draccus.dump(self, f, indent=4)
+        # which `from_pretrained` needs to resolve the concrete subclass.
+        with open(save_directory / CONFIG_NAME, "w") as f:
+            json.dump(draccus.encode(self, PreTrainedConfig), f, indent=4)
 
     @classmethod
     def from_pretrained(

src/lerobot/configs/rewards.py

@@ -103,8 +103,10 @@ class RewardModelConfig(draccus.ChoiceRegistry, HubMixin, abc.ABC):
         pass
 
     def _save_pretrained(self, save_directory: Path) -> None:
-        with open(save_directory / CONFIG_NAME, "w") as f, draccus.config_type("json"):
+        # Encode against the base class so draccus includes the choice "type" key,
-            draccus.dump(self, f, indent=4)
+        # which `from_pretrained` needs to resolve the concrete subclass.
+        with open(save_directory / CONFIG_NAME, "w") as f:
+            json.dump(draccus.encode(self, RewardModelConfig), f, indent=4)
 
     @classmethod
     def from_pretrained(

src/lerobot/configs/train.py

@@ -194,7 +194,11 @@ class TrainPipelineConfig(HubMixin):
             )
 
         if Path(config_path).resolve().exists():
-            policy_dir = Path(config_path).parent
+            # `config_path` may point at the checkpoint's train_config.json or at its
+            # pretrained_model/ directory (both documented above) — resolve either to
+            # the pretrained_model/ directory.
+            config_path_obj = Path(config_path)
+            policy_dir = config_path_obj.parent if config_path_obj.is_file() else config_path_obj
             self.checkpoint_path = policy_dir.parent
         elif self.job.is_remote:
             return

src/lerobot/policies/evo1/modeling_evo1.py

@@ -42,6 +42,9 @@ class Evo1Policy(PreTrainedPolicy):
     config_class = Evo1Config
     name = "evo1"
 
+    def supports_rtc(self) -> bool:
+        return True
+
     def __init__(self, config: Evo1Config, *, vlm_hub_kwargs: dict | None = None, **kwargs):
         super().__init__(config)
         config.validate_features()

src/lerobot/policies/groot/modeling_groot.py

@@ -68,6 +68,9 @@ class GrootPolicy(PreTrainedPolicy):
     name = "groot"
     config_class = GrootConfig
 
+    def supports_rtc(self) -> bool:
+        return True
+
     def __init__(self, config: GrootConfig, **kwargs):
         """Initialize Groot policy wrapper."""
         require_package("transformers", extra="groot")

src/lerobot/policies/molmoact2/modeling_molmoact2.py

@@ -520,6 +520,9 @@ class MolmoAct2Policy(PreTrainedPolicy):
     config_class = MolmoAct2Config
     name = "molmoact2"
 
+    def supports_rtc(self) -> bool:
+        return self.config.inference_action_mode == "continuous"
+
     def __init__(
         self,
         config: MolmoAct2Config,

src/lerobot/policies/pi0/modeling_pi0.py

@@ -749,6 +749,9 @@ class PI0Policy(PreTrainedPolicy):
     config_class = PI0Config
     name = "pi0"
 
+    def supports_rtc(self) -> bool:
+        return True
+
     def __init__(
         self,
         config: PI0Config,

src/lerobot/policies/pi05/modeling_pi05.py

@@ -714,6 +714,9 @@ class PI05Policy(PreTrainedPolicy):
     config_class = PI05Config
     name = "pi05"
 
+    def supports_rtc(self) -> bool:
+        return True
+
     def __init__(
         self,
         config: PI05Config,

src/lerobot/policies/pretrained.py

@@ -249,6 +249,10 @@ class PreTrainedPolicy(nn.Module, HubMixin, abc.ABC):
         """
         raise NotImplementedError
 
+    def supports_rtc(self) -> bool:
+        """Whether this policy implements Real-Time Chunking inference semantics."""
+        return False
+
     # TODO(aliberts, rcadene): split into 'forward' and 'compute_loss'?
     @abc.abstractmethod
     def forward(self, batch: dict[str, Tensor]) -> tuple[Tensor, dict | None]:

src/lerobot/policies/smolvla/modeling_smolvla.py

@@ -145,6 +145,9 @@ class SmolVLAPolicy(PreTrainedPolicy):
     config_class = SmolVLAConfig
     name = "smolvla"
 
+    def supports_rtc(self) -> bool:
+        return True
+
     def __init__(
         self,
         config: SmolVLAConfig,

src/lerobot/policies/smolvla/smolvlm_with_expert.py

@@ -168,14 +168,23 @@ class SmolVLMWithExpertModel(nn.Module):
                 last_layers.append(self.num_vlm_layers - 2)
             frozen_layers = [
                 "lm_head",
-                "text_model.model.norm.weight",
+                "text_model.norm.weight",
             ]
             for layer in last_layers:
-                frozen_layers.append(f"text_model.model.layers.{layer}.")
+                frozen_layers.append(f"text_model.layers.{layer}.")
 
+            unmatched_patterns = set(frozen_layers)
             for name, params in self.vlm.named_parameters():
-                if any(k in name for k in frozen_layers):
+                matched_patterns = [k for k in frozen_layers if k in name]
+                if matched_patterns:
                     params.requires_grad = False
+                    unmatched_patterns.difference_update(matched_patterns)
+            if unmatched_patterns:
+                raise RuntimeError(
+                    "Some frozen layer patterns matched no VLM parameters, so the corresponding layers "
+                    "would silently remain trainable (parameter naming may have changed in transformers): "
+                    f"{sorted(unmatched_patterns)}"
+                )
         # To avoid unused params issue with distributed training
         for name, params in self.lm_expert.named_parameters():
             if "lm_head" in name:

src/lerobot/rl/algorithms/configs.py

@@ -16,6 +16,7 @@ from __future__ import annotations
 
 import abc
 import builtins
+import json
 import logging
 import os
 from dataclasses import dataclass, field
@@ -78,8 +79,10 @@ class RLAlgorithmConfig(draccus.ChoiceRegistry, HubMixin, abc.ABC):
 
     def _save_pretrained(self, save_directory: Path) -> None:
         """Serialize this config as ``config.json`` inside ``save_directory``."""
-        with open(save_directory / CONFIG_NAME, "w") as f, draccus.config_type("json"):
+        # Encode against the base class so draccus includes the choice "type" key,
-            draccus.dump(self, f, indent=4)
+        # which `from_pretrained` needs to resolve the concrete subclass.
+        with open(save_directory / CONFIG_NAME, "w") as f:
+            json.dump(draccus.encode(self, RLAlgorithmConfig), f, indent=4)
 
     @classmethod
     def from_pretrained(

src/lerobot/robots/unitree_g1/config_unitree_g1.py

@@ -68,10 +68,6 @@ class UnitreeG1Config(RobotConfig):
     # Compensates for gravity on the unitree's arms using the arm ik solver
     gravity_compensation: bool = False
 
-    # Locomotion controller class name, e.g. "GrootLocomotionController",
+    # Lower-body controller class name, e.g. "GrootLocomotionController" or
-    # "HolosomaLocomotionController", or "SonicWholeBodyController". None disables it.
+    # "HolosomaLocomotionController". None disables it.
-    # Selecting "SonicWholeBodyController" implicitly switches the robot to the 64-D
-    # latent-token action/observation interface (``motion_token.{i}.pos`` action and a
-    # ``motion_token_state.{i}.pos`` state echo) so ``lerobot-rollout`` can drive a
-    # policy trained on SONIC motion tokens (e.g. nepyope/sonic_walk).
     controller: str | None = None

src/lerobot/robots/unitree_g1/controllers/__init__.py

@@ -1,27 +0,0 @@
-#!/usr/bin/env python
-
-# Copyright 2025 The HuggingFace Inc. team. All rights reserved.
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#     http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# distributed under the License is distributed on an "AS IS" BASIS,
-# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and
-# limitations under the License.
-
-"""Unitree G1 locomotion controllers (Groot, Holosoma, SONIC)."""
-
-from .gr00t_locomotion import GrootLocomotionController
-from .holosoma_locomotion import HolosomaLocomotionController
-from .sonic_whole_body import SonicWholeBodyController
-
-__all__ = [
-    "GrootLocomotionController",
-    "HolosomaLocomotionController",
-    "SonicWholeBodyController",
-]

src/lerobot/robots/unitree_g1/controllers/sonic_whole_body.py

@@ -1,360 +0,0 @@
-#!/usr/bin/env python
-
-# Copyright 2025 The HuggingFace Inc. team. All rights reserved.
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#     http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# distributed under the License is distributed on an "AS IS" BASIS,
-# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and
-# limitations under the License.
-
-"""SONIC decoder whole-body controller for the Unitree G1 (token-only).
-
-Pure-Python/ONNX re-implementation of the *decode* half of NVIDIA's SONIC deploy stack.
-The encoder is intentionally absent: a token-output VLA (e.g. ``nepyope/sonic_walk``)
-supplies the 64-D latent ``motion_token`` directly each tick, and the SONIC **decoder**
-maps ``token + recent proprioception history`` to a residual action that is scaled and
-added onto the standing pose (``default_angles``) to produce 50 Hz joint-position targets
-for the robot's PD controller.
-
-Index spaces: joints exist in two orderings — **IsaacLab** (policy/training order) and
-**MuJoCo** (deploy order). ``ISAACLAB_TO_MUJOCO`` / ``MUJOCO_TO_ISAACLAB`` (in g1_utils)
-convert between them. Quaternions are scalar-first ``(w, x, y, z)``.
-"""
-
-from __future__ import annotations
-
-import json
-import logging
-
-import numpy as np
-import onnx
-import onnxruntime as ort
-from huggingface_hub import hf_hub_download
-
-from ..g1_utils import (
-    ISAACLAB_TO_MUJOCO,
-    MUJOCO_TO_ISAACLAB,
-    G1_29_JointIndex,
-    get_gravity_orientation,
-)
-
-logger = logging.getLogger(__name__)
-
-# ── Constants (hardware-validated; see the NVIDIA SONIC deploy reference) ──────
-CONTROL_DT = 0.02  # 50 Hz control period (s)
-TOKEN_DIM = 64  # decoder latent size
-
-# SONIC decoder checkpoint: NVIDIA's decoder ONNX re-packaged with its deploy constants
-# (kp/kd PD gains, the standing pose default_angles, and the residual action_scale) embedded
-# in the ONNX metadata; see upload_sonic_decoder.py for provisioning. The runtime loads the
-# model *and* all of these straight from the checkpoint (the Holosoma convention), so no
-# motor-physics math happens at deploy time.
-DEFAULT_SONIC_REPO_ID = "lerobot/sonic_decoder"
-DECODER_FILENAME = "model_decoder.onnx"
-DECODER_INPUT_DIM = 994  # token(64) + 10-frame proprio history + gravity
-
-
-def load_sonic_decoder(repo_id: str = DEFAULT_SONIC_REPO_ID):
-    """Load the SONIC decoder ONNX and its baked-in deploy constants from the checkpoint.
-
-    Returns ``(decoder_session, kp, kd, default_angles, action_scale, neutral_token)``. The
-    gains/pose/scale are (29,) float32 in IsaacLab joint order and ``neutral_token`` is the
-    (64,) float32 idle latent -- all read from the ONNX ``metadata_props`` rather than
-    recomputed/hardcoded at deploy time (mirrors ``holosoma_locomotion.load_policy``).
-    """
-    decoder_path = hf_hub_download(repo_id=repo_id, filename=DECODER_FILENAME)
-    so = ort.SessionOptions()
-    so.log_severity_level = 3  # quiet ORT logs
-    session = ort.InferenceSession(decoder_path, sess_options=so)
-    dec_dim = int(session.get_inputs()[0].shape[1])
-    if dec_dim != DECODER_INPUT_DIM:
-        raise RuntimeError(f"Unexpected decoder input dim {dec_dim} (expected {DECODER_INPUT_DIM})")
-
-    meta = {p.key: p.value for p in onnx.load(decoder_path, load_external_data=False).metadata_props}
-    required = ("kp", "kd", "default_angles", "action_scale", "neutral_token")
-    missing = [k for k in required if k not in meta]
-    if missing:
-        raise ValueError(
-            f"SONIC decoder ONNX at {repo_id} is missing metadata {missing}; "
-            "re-run upload_sonic_decoder.py to (re)provision the checkpoint."
-        )
-    arr = {k: np.array(json.loads(meta[k]), dtype=np.float32) for k in required}
-    logger.info("Loaded SONIC deploy constants from %s (%d joints)", repo_id, len(arr["kp"]))
-    return session, arr["kp"], arr["kd"], arr["default_angles"], arr["action_scale"], arr["neutral_token"]
-
-
-# Action-feature prefix for the latent-token interface (see _extract_token_from_action).
-TOKEN_ACTION_PREFIX = "motion_token"  # nosec B105 - feature-key prefix, not a secret
-# Proprio-state prefix for the token interface: the robot echoes the last commanded token
-# here so ``lerobot-rollout`` aggregates it into a 64-D ``observation.state``.
-TOKEN_STATE_PREFIX = "motion_token_state"  # nosec B105 - feature-key prefix, not a secret
-
-
-def token_action_key(i: int) -> str:
-    """Action-dict key for the i-th component of the 64-D SONIC latent token.
-
-    The ``.pos`` suffix is required so the value flows through ``lerobot-rollout``, which
-    only routes ``.pos`` scalar features onto the policy action vector.
-    """
-    return f"{TOKEN_ACTION_PREFIX}.{i}.pos"
-
-
-def token_state_key(i: int) -> str:
-    """Observation key for the i-th component of the 64-D SONIC latent token state."""
-    return f"{TOKEN_STATE_PREFIX}.{i}.pos"
-
-
-# Startup blend duration: over the first control ticks, linearly interpolate every joint
-# from the robot's initial measured pose into the policy's commanded target, so control
-# eases in without a snap on the first command.
-INIT_RAMP_S = 3.0
-
-
-def _extract_token_from_action(action: dict | None) -> np.ndarray | None:
-    """Reassemble a dense (64,) latent token from ``motion_token.{i}`` keys, or None.
-
-    The token-only interface: the caller supplies the 64-D encoder latent directly (e.g. a
-    token-output VLA's action), which the decoder consumes with the encoder bypassed.
-    Requires the full dense token; a partial one is ignored (returns None).
-    """
-    if not action:
-        return None
-    keys = [token_action_key(i) for i in range(TOKEN_DIM)]
-    if any(key not in action for key in keys):
-        return None
-    return np.fromiter((float(action[key]) for key in keys), dtype=np.float32, count=TOKEN_DIM)
-
-
-class SonicDecoder:
-    """Runs the SONIC decoder ONNX model and owns the proprioception history.
-
-    Each tick it appends the latest robot state to 10-frame history buffers, then maps the
-    supplied 64-D ``token`` + that history to a residual action added onto ``default_angles``.
-    The encoder is bypassed entirely (token supplied by the policy). ``default_angles`` and
-    ``action_scale`` are (29,) float32 in IsaacLab order, loaded from the checkpoint.
-    """
-
-    def __init__(self, decoder, default_angles, action_scale):
-        self.decoder = decoder
-        self.decoder_input = decoder.get_inputs()[0].name
-        self.default_angles = np.asarray(default_angles, np.float32)
-        self.action_scale = np.asarray(action_scale, np.float32)
-        self.default_angles_mj = self.default_angles[MUJOCO_TO_ISAACLAB]
-        self.token = np.zeros(TOKEN_DIM, np.float32)
-        self.last_action_mj = np.zeros(29, np.float32)
-        self.h_q_mj = [np.zeros(29, np.float32)] * 10
-        self.h_dq_mj = [np.zeros(29, np.float32)] * 10
-        self.h_ang = [np.zeros(3, np.float32)] * 10
-        self.h_act_mj = [np.zeros(29, np.float32)] * 10
-        self.h_quat = [np.array([1, 0, 0, 0], np.float32)] * 10
-
-    def reset(self):
-        """Clear the token and 10-frame proprioception history.
-
-        ``UnitreeG1.reset()`` relies on this so the first decoder outputs of a new episode
-        are not contaminated by the previous episode's state.
-        """
-        self.token = np.zeros(TOKEN_DIM, np.float32)
-        self.last_action_mj = np.zeros(29, np.float32)
-        self.h_q_mj = [np.zeros(29, np.float32)] * 10
-        self.h_dq_mj = [np.zeros(29, np.float32)] * 10
-        self.h_ang = [np.zeros(3, np.float32)] * 10
-        self.h_act_mj = [np.zeros(29, np.float32)] * 10
-        self.h_quat = [np.array([1, 0, 0, 0], np.float32)] * 10
-
-    def update_history(self, q, dq, ang, quat):
-        """Push the latest proprioception (pos/vel/gyro/orientation) into the 10-frame buffers."""
-        quat = quat / (np.linalg.norm(quat) + 1e-8)
-        # Reorder IsaacLab-order state into the MuJoCo order the decoder consumes. This
-        # permutation direction is validated against the deployed SONIC ONNX; don't flip it.
-        q_mj = q[MUJOCO_TO_ISAACLAB]
-        dq_mj = dq[MUJOCO_TO_ISAACLAB]
-        self.h_q_mj = [q_mj - self.default_angles_mj] + self.h_q_mj[:-1]
-        self.h_dq_mj = [dq_mj] + self.h_dq_mj[:-1]
-        self.h_ang = [ang.copy()] + self.h_ang[:-1]
-        self.h_act_mj = [self.last_action_mj.copy()] + self.h_act_mj[:-1]
-        self.h_quat = [quat.copy()] + self.h_quat[:-1]
-
-    def build_decoder_obs(self):
-        """Assemble the 994-D decoder input: token + 10-frame proprioception history + gravity."""
-        obs = np.zeros(994, np.float32)
-        off = 0
-        obs[off : off + 64] = self.token
-        off += 64
-        for h, sz in [
-            (list(reversed(self.h_ang)), 3),
-            (list(reversed(self.h_q_mj)), 29),
-            (list(reversed(self.h_dq_mj)), 29),
-            (list(reversed(self.h_act_mj)), 29),
-        ]:
-            for f in range(10):
-                obs[off : off + sz] = h[f]
-                off += sz
-        for q in reversed(self.h_quat):
-            obs[off : off + 3] = get_gravity_orientation(q)
-            off += 3
-        assert off == 994, f"Decoder obs mismatch: {off}"
-        return obs
-
-    def step(self, robot_obs, token, debug=False):
-        """One control tick: read robot obs, decode the supplied token -> joint targets.
-
-        Args:
-            robot_obs: dict with ``<joint>.q``/``.dq`` and ``imu.*`` fields.
-            token: 64-D latent supplied by the policy (encoder bypassed).
-            debug: log action/delta norms.
-
-        Returns:
-            dict of ``<joint>.q`` target positions (rad) in IsaacLab joint order.
-        """
-        self.token = np.asarray(token, np.float32)
-        jnames = [m.name for m in G1_29_JointIndex]
-        q = np.array(
-            [
-                robot_obs.get(f"{n}.q", self.default_angles[m.value])
-                for m, n in zip(G1_29_JointIndex, jnames, strict=False)
-            ],
-            np.float32,
-        )
-        dq = np.array([robot_obs.get(f"{n}.dq", 0.0) for n in jnames], np.float32)
-        quat = np.array(
-            [
-                robot_obs.get("imu.quat.w", 1),
-                robot_obs.get("imu.quat.x", 0),
-                robot_obs.get("imu.quat.y", 0),
-                robot_obs.get("imu.quat.z", 0),
-            ],
-            np.float32,
-        )
-        ang = np.array([robot_obs.get(f"imu.gyro.{a}", 0) for a in "xyz"], np.float32)
-        self.update_history(q, dq, ang, quat)
-        action_mj = (
-            self.decoder.run(None, {self.decoder_input: self.build_decoder_obs().reshape(1, -1)})[0]
-            .squeeze()
-            .astype(np.float32)
-        )
-        self.last_action_mj = action_mj.copy()
-        target = self.default_angles + action_mj[ISAACLAB_TO_MUJOCO] * self.action_scale
-        if debug:
-            delta = target - q
-            logger.debug(
-                "token_norm=%.4f action_norm=%.4f delta_max=%.4f delta_rms=%.4f",
-                np.linalg.norm(self.token),
-                np.linalg.norm(action_mj),
-                np.max(np.abs(delta)),
-                np.sqrt(np.mean(delta**2)),
-            )
-        return {f"{m.name}.q": float(target[m.value]) for m in G1_29_JointIndex}
-
-
-class SonicRuntime:
-    """Loads the SONIC decoder ONNX model and owns the decode controller.
-
-    Token-only deploy: the encoder is bypassed; each tick the decoder consumes a 64-D
-    latent token supplied directly by the policy.
-    """
-
-    def __init__(self):
-        decoder_sess, self.kp, self.kd, default_angles, action_scale, neutral_token = load_sonic_decoder()
-        self.default_angles = default_angles
-        self.neutral_token = neutral_token
-        self.controller = SonicDecoder(decoder_sess, default_angles, action_scale)
-
-    @property
-    def pipeline(self):
-        return self.controller
-
-    def reset(self):
-        self.controller.reset()
-
-    def shutdown(self):
-        pass
-
-
-class SonicWholeBodyController:
-    """Full-body SONIC controller for UnitreeG1's background controller thread."""
-
-    control_dt = CONTROL_DT
-    full_body = True
-
-    def __init__(self):
-        logger.info("Loading SONIC whole-body controller...")
-        self._runtime = SonicRuntime()
-        self.kp = self._runtime.kp
-        self.kd = self._runtime.kd
-        self.controller = self._runtime.controller
-        self._default_angles = self._runtime.default_angles
-        self._neutral_token = self._runtime.neutral_token
-
-        # Startup blend: ease from the robot's initial pose into the first commanded policy
-        # targets over INIT_RAMP_S (captured on the first control tick).
-        self._init_ramp_steps = max(1, round(INIT_RAMP_S / CONTROL_DT))
-        self._init_step = 0
-        self._start_pose: dict[str, float] = {}
-
-        # Token-interface state. The controller holds a stable *neutral* token until the first
-        # real token arrives, and afterwards holds the *last* token received between ticks (the
-        # async controller runs ~50 Hz while a token VLA streams ~30 Hz).
-        self._last_token: np.ndarray | None = None
-
-        logger.info("SONIC ready (decoder, 64-D token command path)")
-
-    def _startup_blend(self, obs: dict, out: dict) -> dict:
-        """Ease into policy control at startup: for the first ``INIT_RAMP_S`` seconds,
-        interpolate between the robot's pose captured on the first tick and the policy's
-        live commanded target, so the handoff has no snap.
-
-        ``out`` is the policy's ``<joint>.q`` target dict for this tick; the blend ratio
-        climbs 0->1 over the ramp, after which the raw policy target passes through.
-        """
-        if self._init_step >= self._init_ramp_steps or not out:
-            return out
-        if self._init_step == 0:
-            # Capture the robot's actual pose as the interpolation start point.
-            self._start_pose = {
-                f"{m.name}.q": float(obs.get(f"{m.name}.q", self._default_angles[m.value]))
-                for m in G1_29_JointIndex
-            }
-        self._init_step += 1
-        ratio = min(1.0, self._init_step / self._init_ramp_steps)
-        blended = {
-            k: self._start_pose.get(k, float(tgt)) * (1.0 - ratio) + float(tgt) * ratio
-            for k, tgt in out.items()
-        }
-        if self._init_step >= self._init_ramp_steps:
-            logger.info("SONIC startup blend complete -> full policy control")
-        return blended
-
-    def run_step(self, action: dict, obs: dict) -> dict:
-        if not obs:
-            return {}
-
-        # Token-only interface (token-output VLA): a dense 64-D ``motion_token.{i}`` command
-        # is decoded directly, encoder bypassed.
-        token = _extract_token_from_action(action)
-        if token is not None:
-            self._last_token = token
-        elif self._last_token is None:
-            # No token has arrived yet: hold the checkpoint's neutral token, which the decoder
-            # maps to a stable, natural standing pose.
-            self._last_token = self._neutral_token.copy()
-        # Either a fresh token this tick or the last one received (held between the ~30 Hz
-        # token stream and the ~50 Hz control loop).
-        return self._startup_blend(obs, self.controller.step(obs, self._last_token))
-
-    def reset(self):
-        self._runtime.reset()
-        self._init_step = 0  # re-run the startup blend after a reset
-        self._start_pose = {}
-        # Drop the held token so the neutral token is re-seeded after a reset.
-        self._last_token = None
-
-    def shutdown(self):
-        self._runtime.shutdown()

src/lerobot/robots/unitree_g1/g1_utils.py

@@ -23,47 +23,6 @@ import numpy as np
 
 NUM_MOTORS = 29
 
-# Joint-order permutations between the two 29-DoF layouts used across the G1 stack:
-# IsaacLab (policy/training order) and MuJoCo (deploy order). ``a[ISAACLAB_TO_MUJOCO]``
-# reorders an IsaacLab-ordered vector into MuJoCo order, and vice-versa.
-ISAACLAB_TO_MUJOCO = np.array(
-    [
-        0,
-        3,
-        6,
-        9,
-        13,
-        17,
-        1,
-        4,
-        7,
-        10,
-        14,
-        18,
-        2,
-        5,
-        8,
-        11,
-        15,
-        19,
-        21,
-        23,
-        25,
-        27,
-        12,
-        16,
-        20,
-        22,
-        24,
-        26,
-        28,
-    ],
-    dtype=np.int32,
-)
-# The two orderings are inverses of each other, so derive one from the other (argsort) to
-# guarantee they can never drift out of sync.
-MUJOCO_TO_ISAACLAB = np.argsort(ISAACLAB_TO_MUJOCO).astype(np.int32)
-
 REMOTE_AXES = ("remote.lx", "remote.ly", "remote.rx", "remote.ry")
 REMOTE_BUTTONS = tuple(f"remote.button.{i}" for i in range(16))
 REMOTE_KEYS = REMOTE_AXES + REMOTE_BUTTONS
@@ -109,9 +68,8 @@ def make_locomotion_controller(name: str | None):
     if name is None:
         return None
     controllers = {
-        "GrootLocomotionController": "lerobot.robots.unitree_g1.controllers.gr00t_locomotion",
+        "GrootLocomotionController": "lerobot.robots.unitree_g1.gr00t_locomotion",
-        "HolosomaLocomotionController": "lerobot.robots.unitree_g1.controllers.holosoma_locomotion",
+        "HolosomaLocomotionController": "lerobot.robots.unitree_g1.holosoma_locomotion",
-        "SonicWholeBodyController": "lerobot.robots.unitree_g1.controllers.sonic_whole_body",
     }
     module_path = controllers.get(name)
     if module_path is None:

src/lerobot/robots/unitree_g1/gr00t_locomotion.py

@@ -21,7 +21,7 @@ import numpy as np
 import onnxruntime as ort
 from huggingface_hub import hf_hub_download
 
-from ..g1_utils import (
+from .g1_utils import (
     REMOTE_AXES,
     REMOTE_BUTTONS,
     G1_29_JointIndex,

src/lerobot/robots/unitree_g1/holosoma_locomotion.py

@@ -22,7 +22,7 @@ import onnx
 import onnxruntime as ort
 from huggingface_hub import hf_hub_download
 
-from ..g1_utils import (
+from .g1_utils import (
     REMOTE_AXES,
     G1_29_JointArmIndex,
     G1_29_JointIndex,

src/lerobot/robots/unitree_g1/unitree_g1.py

@@ -34,6 +34,7 @@ from .config_unitree_g1 import UnitreeG1Config
 from .g1_kinematics import G1_29_ArmIK
 from .g1_utils import (
     REMOTE_AXES,
+    REMOTE_KEYS,
     G1_29_JointArmIndex,
     G1_29_JointIndex,
     default_remote_input,
@@ -147,54 +148,22 @@ class UnitreeG1(Robot):
 
         self.arm_ik = G1_29_ArmIK() if config.gravity_compensation else None
 
-        # Lower-body / whole-body controller loaded dynamically
+        # Lower-body controller loaded dynamically
         self.controller: LocomotionController | None = make_locomotion_controller(config.controller)
+
         # Controller thread state
         self._controller_thread = None
-        # When set, the controller loop stops publishing low commands so reset() can
-        # drive the joints directly without two publishers fighting (single-publisher).
-        self._controller_paused = threading.Event()
         self._controller_action_lock = threading.Lock()
         self.controller_input = default_remote_input()
         self.controller_output = {}
 
-        # Token-mode state: last 64-D SONIC latent token commanded by the policy,
-        # echoed back as ``observation.state`` so a token-output VLA closes the loop
-        # on its own previous token. Implicit whenever the SONIC whole-body controller
-        # is active. Seeded to zeros; the controller's startup blend eases joints in.
-        self._last_token: np.ndarray | None = None
-        if self._sonic_token:
-            from .controllers.sonic_whole_body import TOKEN_DIM
-
-            self._last_token = np.zeros(TOKEN_DIM, dtype=np.float32)
-
-    @property
-    def _sonic_token(self) -> bool:
-        """Whether the SONIC whole-body decoder is active.
-
-        A SONIC controller consumes a 64-D latent motion token as its action and echoes
-        the last commanded token as ``observation.state``. Keyed purely off the selected
-        controller so the token interface is implicit -- no separate config flag.
-        """
-        return self.config.controller == "SonicWholeBodyController"
-
     def _subscribe_lowstate(self):  # polls robot state @ 250Hz
         while not self._shutdown_event.is_set():
             start_time = time.time()
 
             # Step simulation if in simulation mode
             if self.config.is_simulation and self.sim_env is not None:
-                try:
                 self.sim_env.step()
-                except ValueError as e:
-                    # Startup race: the sim thread can step once before reset() has
-                    # written a valid base pose, giving a zero-norm pelvis quaternion
-                    # (scipy>=1.11 raises instead of normalizing). Skip and retry so
-                    # the thread survives instead of dying and freezing the sim.
-                    if "zero norm" not in str(e).lower():
-                        raise
-                    time.sleep(self.control_dt)
-                    continue
 
             msg = self.lowstate_subscriber.Read()
             if msg is not None:
@@ -262,38 +231,15 @@ class UnitreeG1(Robot):
                 features[f"{cam}_depth"] = (cfg.height, cfg.width, 1)
         return features
 
-    @property
-    def _token_state_ft(self) -> dict[str, type]:
-        """64-D SONIC latent-token proprio state (``motion_token_state.{i}.pos``).
-
-        Exposed only when a SONIC whole-body controller is active; aggregated by the
-        rollout into a 64-D ``observation.state`` (the last token the policy commanded).
-        """
-        if not self._sonic_token:
-            return {}
-        from .controllers.sonic_whole_body import TOKEN_DIM, token_state_key
-
-        return {token_state_key(i): float for i in range(TOKEN_DIM)}
-
     @cached_property
     def observation_features(self) -> dict[str, type | tuple]:
-        return {**self._motors_ft, **self._token_state_ft, **self._cameras_ft}
+        return {**self._motors_ft, **self._cameras_ft}
 
     @cached_property
     def action_features(self) -> dict[str, type]:
-        # No controller configured at all: raw 29-DoF joint teleop.
         if self.controller is None:
             return {f"{G1_29_JointIndex(motor).name}.q": float for motor in G1_29_JointIndex}
 
-        # Token-output VLA (SONIC decoder): advertise a 64-D latent-token action space
-        # (``motion_token.{i}.pos``) so ``lerobot-rollout`` maps a 64-D policy output
-        # straight onto the decoder, bypassing the encoder.
-        if self._sonic_token:
-            from .controllers.sonic_whole_body import TOKEN_DIM, token_action_key
-
-            return {token_action_key(i): float for i in range(TOKEN_DIM)}
-
-        # Locomotion controllers (GR00T / Holosoma): arm joint targets + joystick axes.
         arm_features = {f"{G1_29_JointArmIndex(motor).name}.q": float for motor in G1_29_JointArmIndex}
         remote_features = dict.fromkeys(REMOTE_AXES, float)
         return {**arm_features, **remote_features}
@@ -309,11 +255,6 @@ class UnitreeG1(Robot):
         while not self._shutdown_event.is_set():
             start_time = time.time()
 
-            # Paused during reset() so the reset routine is the sole low-cmd publisher.
-            if self._controller_paused.is_set():
-                time.sleep(control_dt)
-                continue
-
             with self._lowstate_lock:
                 lowstate = self._lowstate
 
@@ -330,12 +271,8 @@ class UnitreeG1(Robot):
                 with self._controller_action_lock:
                     controller_input = dict(self.controller_input)
 
-                # Full-body controllers (SONIC) consume the full observation dict; others
+                # Run controller step
-                # take the raw lowstate. get_observation() is the single lowstate -> obs builder.
+                controller_action = self.controller.run_step(controller_input, lowstate)
-                controller_state = (
-                    self.get_observation() if getattr(self.controller, "full_body", False) else lowstate
-                )
-                controller_action = self.controller.run_step(controller_input, controller_state)
 
                 # Write controller output snapshot
                 with self._controller_action_lock:
@@ -406,9 +343,6 @@ class UnitreeG1(Robot):
 
         self.kp = np.array(self.config.kp, dtype=np.float32)
         self.kd = np.array(self.config.kd, dtype=np.float32)
-        if self.controller is not None and hasattr(self.controller, "kp"):
-            self.kp = np.array(self.controller.kp, dtype=np.float32)
-            self.kd = np.array(self.controller.kd, dtype=np.float32)
 
         for joint in G1_29_JointIndex:
             self.msg.motor_cmd[joint].mode = 1
@@ -457,10 +391,6 @@ class UnitreeG1(Robot):
             if self._controller_thread.is_alive():
                 logger.warning("Controller thread did not stop cleanly")
 
-        # Release controller resources (e.g. SONIC decoder sessions).
-        if self.controller is not None and hasattr(self.controller, "shutdown"):
-            self.controller.shutdown()
-
         # Close simulation environment
         if self.config.is_simulation and self.sim_env is not None:
             try:
@@ -531,15 +461,6 @@ class UnitreeG1(Robot):
         if lowstate.wireless_remote:
             obs["wireless_remote"] = lowstate.wireless_remote
 
-        # Token mode: echo the last commanded latent token as observation.state so a
-        # token-output VLA closes the loop on its own previous token.
-        if self._sonic_token:
-            from .controllers.sonic_whole_body import token_state_key
-
-            token = self._last_token if self._last_token is not None else []
-            for i, v in enumerate(token):
-                obs[token_state_key(i)] = float(v)
-
         # Cameras - read images from ZMQ cameras
         for cam_name, cam in self._cameras.items():
             if getattr(cam, "use_rgb", True):
@@ -552,22 +473,9 @@ class UnitreeG1(Robot):
     def send_action(self, action: RobotAction) -> RobotAction:
         action_to_publish = action
         if self.controller is not None:
-            # SONIC decoder: pull the 64-D latent token out of the action and remember it
-            # for the observation.state echo. The controller thread reads it back from
-            # controller_input (populated below) and decodes it into a 29-DoF command.
-            if self._sonic_token:
-                from .controllers.sonic_whole_body import _extract_token_from_action
-
-                token = _extract_token_from_action(action)
-                if token is not None:
-                    self._last_token = token
-            self._update_controller_action(action)
-            # Full-body controllers (SONIC) own the whole 29-DoF command; nothing to
-            # publish here (the controller thread is the sole publisher).
-            if getattr(self.controller, "full_body", False):
-                return action
             # Controller thread owns legs/waist. Here we only update joystick inputs
             # and publish arm targets from the teleoperator.
+            self._update_controller_action(action)
             arm_prefixes = tuple(j.name for j in G1_29_JointArmIndex)
             action_to_publish = {
                 key: value
@@ -595,17 +503,11 @@ class UnitreeG1(Robot):
         return action
 
     def _update_controller_action(self, action: RobotAction) -> None:
-        """Update controller input state from an incoming teleop action.
+        """Update controller input state from incoming teleop action."""
-
-        Controller-agnostic: every value-carrying key (locomotion ``remote.*`` axes or
-        SONIC ``motion_token.*`` values) is forwarded verbatim into ``controller_input``
-        and each controller extracts only the keys it understands. The robot deliberately
-        does not enumerate any controller's key schema here.
-        """
         with self._controller_action_lock:
-            for key, value in action.items():
+            for key in REMOTE_KEYS:
-                if isinstance(key, str) and value is not None:
+                if key in action:
-                    self.controller_input[key] = value
+                    self.controller_input[key] = action[key]
 
     @property
     def is_calibrated(self) -> bool:
@@ -635,18 +537,6 @@ class UnitreeG1(Robot):
         if default_positions is None:
             default_positions = np.array(self.config.default_positions, dtype=np.float32)
 
-        # Full-body controllers (SONIC) own the whole 29-DoF command and ignore
-        # ``<joint>.q`` in send_action(), so reset() must publish the default pose
-        # directly. Pause the background controller first so the two aren't both writing
-        # low commands while the robot moves to the default pose.
-        full_body = getattr(self.controller, "full_body", False)
-        paused = False
-        if full_body and self._controller_thread is not None:
-            self._controller_paused.set()
-            paused = True
-            time.sleep(control_dt)  # let any in-flight controller tick settle
-
-        try:
         if self.config.is_simulation and self.sim_env is not None:
             self.sim_env.reset()
             self.publish_lowcmd(
@@ -675,11 +565,6 @@ class UnitreeG1(Robot):
                     interp_pos = init_dof_pos[motor.value] * (1 - alpha) + target_pos * alpha
                     action_dict[f"{motor.name}.q"] = float(interp_pos)
 
-                    # Full-body controllers no-op in send_action(); publish the pose
-                    # directly (arm-only controllers keep the send_action() path).
-                    if full_body:
-                        self.publish_lowcmd(action_dict)
-                    else:
                 self.send_action(action_dict)
 
                 # Maintain constant control rate
@@ -687,12 +572,8 @@ class UnitreeG1(Robot):
                 sleep_time = max(0, control_dt - elapsed)
                 time.sleep(sleep_time)
 
-            # Reset controller internal state (gait phase, obs history, etc.) before
+        # Reset controller internal state (gait phase, obs history, etc.)
-            # resuming so its buffers reflect the post-reset pose.
         if self.controller is not None and hasattr(self.controller, "reset"):
             self.controller.reset()
-        finally:
-            if paused:
-                self._controller_paused.clear()
 
         logger.info("Reached default position")

src/lerobot/rollout/context.py

@@ -57,6 +57,7 @@ from .inference import (
     SyncInferenceConfig,
     create_inference_engine,
 )
+from .inference.rtc import supports_rtc_inference
 from .robot_wrapper import ThreadSafeRobot
 
 if TYPE_CHECKING or _peft_available:
@@ -226,6 +227,12 @@ def build_rollout_context(
     policy = _load_pretrained_policy(policy_config)
 
     if is_rtc:
+        if not supports_rtc_inference(policy):
+            raise ValueError(
+                f"RTC inference is not supported by policy type '{policy_config.type}': "
+                "the policy must implement RTC semantics and predict_action_chunk must accept "
+                "inference_delay and prev_chunk_left_over. Use '--inference.type=sync' instead."
+            )
         policy.config.rtc_config = cfg.inference.rtc
         if hasattr(policy, "init_rtc_processor"):
             policy.init_rtc_processor()

src/lerobot/rollout/inference/rtc.py

@@ -22,6 +22,7 @@ way via ``notify_observation``.
 
 from __future__ import annotations
 
+import inspect
 import logging
 import math
 import time
@@ -62,6 +63,23 @@ _RTC_JOIN_TIMEOUT_S: float = 3.0
 # ---------------------------------------------------------------------------
 
 
+def supports_rtc_inference(policy: PreTrainedPolicy) -> bool:
+    """Whether a policy declares RTC support and accepts the RTC call shape."""
+    supports_rtc = getattr(policy, "supports_rtc", None)
+    if not callable(supports_rtc) or not supports_rtc():
+        return False
+
+    try:
+        inspect.signature(policy.predict_action_chunk).bind(
+            object(),
+            inference_delay=0,
+            prev_chunk_left_over=None,
+        )
+    except (TypeError, ValueError):
+        return False
+    return True
+
+
 def _normalize_prev_actions_length(prev_actions: torch.Tensor, target_steps: int) -> torch.Tensor:
     """Pad or truncate RTC prefix actions to a fixed length for stable compiled inference."""
     if prev_actions.ndim != 2:

src/lerobot/scripts/lerobot_train.py

@@ -348,7 +348,6 @@ def train(cfg: TrainPipelineConfig, accelerator: "Accelerator | None" = None):
         preprocessor_overrides = {
             "device_processor": {"device": device.type},
             "normalizer_processor": {
-                "stats": dataset.meta.stats,
                 "features": {**policy.config.input_features, **policy.config.output_features},
                 "norm_map": policy.config.normalization_mapping,
             },
@@ -356,11 +355,17 @@ def train(cfg: TrainPipelineConfig, accelerator: "Accelerator | None" = None):
         }
         postprocessor_overrides = {
             "unnormalizer_processor": {
-                "stats": dataset.meta.stats,
                 "features": policy.config.output_features,
                 "norm_map": policy.config.normalization_mapping,
             },
         }
+        # On resume, the checkpoint's saved processor stats are authoritative: they may have
+        # been adapted by the policy (e.g. EVO1 pads state/action stats to max_state_dim),
+        # and force-feeding raw dataset stats over them crashes normalization (#4006).
+        # This mirrors the `dataset_stats` kwarg above, which is also skipped on resume.
+        if not cfg.resume:
+            preprocessor_overrides["normalizer_processor"]["stats"] = dataset.meta.stats
+            postprocessor_overrides["unnormalizer_processor"]["stats"] = dataset.meta.stats
         if getattr(active_cfg, "use_relative_actions", False):
             preprocessor_overrides["relative_actions_processor"] = {
                 "enabled": True,

src/lerobot/teleoperators/bi_so_leader/bi_so_leader.py

@@ -67,7 +67,15 @@ class BiSOLeader(BimanualMixin, Teleoperator):
 
     @cached_property
     def feedback_features(self) -> dict[str, type]:
-        return {}
+        # Bimanual teleop has feedback (can be actuated for handover).
+        # Return the same structure as action_features for consistency with left/right arms.
+        left_arm_features = self.left_arm.feedback_features
+        right_arm_features = self.right_arm.feedback_features
+
+        return {
+            **{f"left_{k}": v for k, v in left_arm_features.items()},
+            **{f"right_{k}": v for k, v in right_arm_features.items()},
+        }
 
     def setup_motors(self) -> None:
         self.left_arm.setup_motors()
@@ -87,6 +95,43 @@ class BiSOLeader(BimanualMixin, Teleoperator):
 
         return action_dict
 
+    def enable_torque(self) -> None:
+        """Enable torque on both leader arms for smooth handover."""
+        self.left_arm.enable_torque()
+        self.right_arm.enable_torque()
+
+    def disable_torque(self) -> None:
+        """Disable torque on both leader arms to allow human control."""
+        self.left_arm.disable_torque()
+        self.right_arm.disable_torque()
+
+    @check_if_not_connected
     def send_feedback(self, feedback: dict[str, float]) -> None:
-        # TODO: Implement force feedback
+        """Route bimanual feedback to left and right arms with proper prefix stripping.
-        raise NotImplementedError
+
+        Receives feedback dict with keys like: left_shoulder_pan.pos, right_shoulder_pan.pos, ...
+        Splits and routes to each arm by removing the prefix.
+
+        This enables DAgger smooth handover: when transitioning from policy control to human
+        intervention, both leader arms are commanded to the follower's current pose to avoid
+        discontinuities.
+        """
+        # Split feedback by arm prefix
+        left_feedback = {}
+        right_feedback = {}
+
+        for key, value in feedback.items():
+            if key.startswith("left_"):
+                # Strip "left_" prefix and pass to left arm
+                stripped_key = key[5:]  # len("left_") == 5
+                left_feedback[stripped_key] = value
+            elif key.startswith("right_"):
+                # Strip "right_" prefix and pass to right arm
+                stripped_key = key[6:]  # len("right_") == 6
+                right_feedback[stripped_key] = value
+
+        # Send to each arm
+        if left_feedback:
+            self.left_arm.send_feedback(left_feedback)
+        if right_feedback:
+            self.right_arm.send_feedback(right_feedback)

src/lerobot/utils/random_utils.py

@@ -85,6 +85,8 @@ def serialize_torch_rng_state() -> dict[str, torch.Tensor]:
     torch_rng_state_dict = {"torch_rng_state": torch.get_rng_state()}
     if torch.cuda.is_available():
         torch_rng_state_dict["torch_cuda_rng_state"] = torch.cuda.get_rng_state()
+    if torch.backends.mps.is_available():
+        torch_rng_state_dict["torch_mps_rng_state"] = torch.mps.get_rng_state()
     return torch_rng_state_dict
 
 
@@ -95,6 +97,8 @@ def deserialize_torch_rng_state(rng_state_dict: dict[str, torch.Tensor]) -> None
     torch.set_rng_state(rng_state_dict["torch_rng_state"])
     if torch.cuda.is_available() and "torch_cuda_rng_state" in rng_state_dict:
         torch.cuda.set_rng_state(rng_state_dict["torch_cuda_rng_state"])
+    if torch.backends.mps.is_available() and "torch_mps_rng_state" in rng_state_dict:
+        torch.mps.set_rng_state(rng_state_dict["torch_mps_rng_state"])
 
 
 def serialize_rng_state() -> dict[str, torch.Tensor]:

tests/configs/test_resume_from_hub.py

@@ -66,3 +66,27 @@ def test_from_pretrained_raises_when_no_root_config_and_no_checkpoints(monkeypat
 
     with pytest.raises(FileNotFoundError, match="train_config.json not found"):
         TrainPipelineConfig.from_pretrained("user/empty-repo")
+
+
+@pytest.mark.parametrize("pass_dir", [False, True])
+def test_resolve_resume_checkpoint_accepts_file_or_pretrained_model_dir(tmp_path, monkeypatch, pass_dir):
+    """`--config_path` may point at the checkpoint's train_config.json or at its
+    pretrained_model/ directory; both must resolve `policy.pretrained_path` to the
+    pretrained_model/ directory (regression test for the directory case, which
+    previously resolved one level too high and failed on model.safetensors)."""
+    pretrained_dir = tmp_path / "checkpoints" / "000002" / "pretrained_model"
+    pretrained_dir.mkdir(parents=True)
+    (pretrained_dir / "train_config.json").touch()
+    target = pretrained_dir if pass_dir else pretrained_dir / "train_config.json"
+
+    from lerobot.policies.act.configuration_act import ACTConfig
+
+    cfg = tc.draccus.parse(TrainPipelineConfig, args=["--dataset.repo_id", "u/d"])
+    cfg.policy = ACTConfig()
+    cfg.resume = True
+    monkeypatch.setattr(tc.parser, "parse_arg", lambda name: str(target) if name == "config_path" else None)
+
+    cfg._resolve_resume_checkpoint()
+
+    assert cfg.policy.pretrained_path == pretrained_dir
+    assert cfg.checkpoint_path == pretrained_dir.parent

tests/utils/test_random_utils.py

@@ -73,6 +73,17 @@ def test_serialize_deserialize_torch_rng(fixed_seed):
     assert val2 == val3
 
 
+@pytest.mark.skipif(not torch.backends.mps.is_available(), reason="MPS not available")
+def test_serialize_deserialize_torch_rng_mps(fixed_seed):
+    _ = torch.rand(1, device="mps").item()
+    st = serialize_torch_rng_state()
+    assert "torch_mps_rng_state" in st
+    val2 = torch.rand(1, device="mps").item()
+    deserialize_torch_rng_state(st)
+    val3 = torch.rand(1, device="mps").item()
+    assert val2 == val3
+
+
 def test_serialize_deserialize_rng(fixed_seed):
     # Generate one from each library
     _ = random.random()

uv.lock

@@ -1,5 +1,5 @@
 version = 1
-revision = 3
+revision = 2
 requires-python = ">=3.12"
 resolution-markers = [
     "(python_full_version >= '3.15' and platform_machine == 'AMD64' and sys_platform == 'linux') or (python_full_version >= '3.15' and platform_machine == 'x86_64' and sys_platform == 'linux')",
@@ -1359,18 +1359,17 @@ sdist = { url = "https://files.pythonhosted.org/packages/a2/55/8f8cab2afd404cf57
 
 [[package]]
 name = "draccus"
-version = "0.10.0"
+version = "0.11.6"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "mergedeep" },
     { name = "pyyaml" },
-    { name = "pyyaml-include" },
     { name = "toml" },
     { name = "typing-inspect" },
 ]
-sdist = { url = "https://files.pythonhosted.org/packages/4e/e2/f5012fda17ee5d1eaf3481b6ca3e11dffa5348e5e08ab745538fdc8041bb/draccus-0.10.0.tar.gz", hash = "sha256:8dd08304219becdcd66cd16058ba98e9c3e6b7bfe48ccb9579dae39f8d37ae19", size = 62243, upload-time = "2025-02-05T07:27:48.182Z" }
+sdist = { url = "https://files.pythonhosted.org/packages/df/b0/cc399719e0cf6fea451f155798e19ea8c5e9c838b701d1565f29ea626c18/draccus-0.11.6.tar.gz", hash = "sha256:d134f576a1f4febd93c6b200df7f92e5febffe9efdc0a5f381bf50c2e0568a39", size = 67940, upload-time = "2026-06-12T13:21:19.999Z" }
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/c4/9a/a83083b230d352ee5d205757b74006dbe084448ca45e3bc5ca99215b1e55/draccus-0.10.0-py3-none-any.whl", hash = "sha256:90243418ae0e9271c390a59cafb6acfd37001193696ed36fcc8525f791a83282", size = 71783, upload-time = "2025-02-05T07:27:46.1Z" },
+    { url = "https://files.pythonhosted.org/packages/18/f1/d56bef4563d1cfaf55dd58de298a456587b9fe0a85dabb55768d44ace8f7/draccus-0.11.6-py3-none-any.whl", hash = "sha256:1cf3f37c64766e0f17d757099b388e762f1d22c772cc2376b4e366b515fb5737", size = 85449, upload-time = "2026-06-12T13:21:18.83Z" },
 ]
 
 [[package]]
@@ -3289,7 +3288,7 @@ requires-dist = [
     { name = "deepdiff", marker = "extra == 'deepdiff-dep'", specifier = ">=7.0.1,<9.0.0" },
     { name = "diffusers", marker = "extra == 'diffusers-dep'", specifier = ">=0.38.0,<0.40.0" },
     { name = "dm-tree", marker = "extra == 'groot'", specifier = ">=0.1.8,<1.0.0" },
-    { name = "draccus", specifier = "==0.10.0" },
+    { name = "draccus", specifier = ">=0.11.6,<0.12.0" },
     { name = "dynamixel-sdk", marker = "extra == 'dynamixel'", specifier = ">=3.7.31,<3.9.0" },
     { name = "einops", specifier = ">=0.8.0,<0.9.0" },
     { name = "faker", marker = "extra == 'sarm'", specifier = ">=33.0.0,<35.0.0" },
@@ -5636,18 +5635,6 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/f1/12/de94a39c2ef588c7e6455cfbe7343d3b2dc9d6b6b2f40c4c6565744c873d/pyyaml-6.0.3-cp314-cp314t-win_arm64.whl", hash = "sha256:ebc55a14a21cb14062aa4162f906cd962b28e2e9ea38f9b4391244cd8de4ae0b", size = 149341, upload-time = "2025-09-25T21:32:56.828Z" },
 ]
 
-[[package]]
-name = "pyyaml-include"
-version = "1.4.1"
-source = { registry = "https://pypi.org/simple" }
-dependencies = [
-    { name = "pyyaml" },
-]
-sdist = { url = "https://files.pythonhosted.org/packages/7f/be/2d07ad85e3d593d69640876a8686eae2c533db8cb7bf298d25c421b4d2d5/pyyaml-include-1.4.1.tar.gz", hash = "sha256:1a96e33a99a3e56235f5221273832464025f02ff3d8539309a3bf00dec624471", size = 20592, upload-time = "2024-03-25T14:56:43.748Z" }
-wheels = [
-    { url = "https://files.pythonhosted.org/packages/d5/ca/6a2cc3a73170d10b5af1f1613baa2ed1f8f46f62dd0bfab2bffd2c2fe260/pyyaml_include-1.4.1-py3-none-any.whl", hash = "sha256:323c7f3a19c82fbc4d73abbaab7ef4f793e146a13383866831631b26ccc7fb00", size = 19079, upload-time = "2024-03-25T14:56:41.274Z" },
-]
-
 [[package]]
 name = "pyzmq"
 version = "27.1.0"