local_segment/sam2/sam2_video_predictor.py

# Copyright (c) Meta Platforms, Inc. and affiliates.
# All rights reserved.

# This source code is licensed under the license found in the
# LICENSE file in the root directory of this source tree.

import warnings
from collections import OrderedDict

import torch

from tqdm import tqdm

from sam2.modeling.sam2_base import NO_OBJ_SCORE, SAM2Base
from sam2.utils.misc import concat_points, fill_holes_in_mask_scores, load_video_frames


class SAM2VideoPredictor(SAM2Base):
    """The predictor class to handle user interactions and manage inference states."""

    def __init__(
        self,
        fill_hole_area=0,
        # whether to apply non-overlapping constraints on the output object masks
        non_overlap_masks=False,
        # whether to clear non-conditioning memory of the surrounding frames (which may contain outdated information) after adding correction clicks;
        # note that this would only apply to *single-object tracking* unless `clear_non_cond_mem_for_multi_obj` is also set to True)
        clear_non_cond_mem_around_input=False,
        # whether to also clear non-conditioning memory of the surrounding frames (only effective when `clear_non_cond_mem_around_input` is True).
        clear_non_cond_mem_for_multi_obj=False,
        **kwargs,
    ):
        super().__init__(**kwargs)
        self.fill_hole_area = fill_hole_area
        self.non_overlap_masks = non_overlap_masks
        self.clear_non_cond_mem_around_input = clear_non_cond_mem_around_input
        self.clear_non_cond_mem_for_multi_obj = clear_non_cond_mem_for_multi_obj

    @torch.inference_mode()
    def init_state(
        self,
        video_path,
        offload_video_to_cpu=False,
        offload_state_to_cpu=False,
        async_loading_frames=False,
    ):
        """Initialize an inference state."""
        compute_device = self.device  # device of the model
        images, video_height, video_width = load_video_frames(
            video_path=video_path,
            image_size=self.image_size,
            offload_video_to_cpu=offload_video_to_cpu,
            async_loading_frames=async_loading_frames,
            compute_device=compute_device,
        )
        inference_state = {}
        inference_state["images"] = images
        inference_state["num_frames"] = len(images)
        # whether to offload the video frames to CPU memory
        # turning on this option saves the GPU memory with only a very small overhead
        inference_state["offload_video_to_cpu"] = offload_video_to_cpu
        # whether to offload the inference state to CPU memory
        # turning on this option saves the GPU memory at the cost of a lower tracking fps
        # (e.g. in a test case of 768x768 model, fps dropped from 27 to 24 when tracking one object
        # and from 24 to 21 when tracking two objects)
        inference_state["offload_state_to_cpu"] = offload_state_to_cpu
        # the original video height and width, used for resizing final output scores
        inference_state["video_height"] = video_height
        inference_state["video_width"] = video_width
        inference_state["device"] = compute_device
        if offload_state_to_cpu:
            inference_state["storage_device"] = torch.device("cpu")
        else:
            inference_state["storage_device"] = compute_device
        # inputs on each frame
        inference_state["point_inputs_per_obj"] = {}
        inference_state["mask_inputs_per_obj"] = {}
        # visual features on a small number of recently visited frames for quick interactions
        inference_state["cached_features"] = {}
        # values that don't change across frames (so we only need to hold one copy of them)
        inference_state["constants"] = {}
        # mapping between client-side object id and model-side object index
        inference_state["obj_id_to_idx"] = OrderedDict()
        inference_state["obj_idx_to_id"] = OrderedDict()
        inference_state["obj_ids"] = []
        # A storage to hold the model's tracking results and states on each frame
        inference_state["output_dict"] = {
            "cond_frame_outputs": {},  # dict containing {frame_idx: <out>}
            "non_cond_frame_outputs": {},  # dict containing {frame_idx: <out>}
        }
        # Slice (view) of each object tracking results, sharing the same memory with "output_dict"
        inference_state["output_dict_per_obj"] = {}
        # A temporary storage to hold new outputs when user interact with a frame
        # to add clicks or mask (it's merged into "output_dict" before propagation starts)
        inference_state["temp_output_dict_per_obj"] = {}
        # Frames that already holds consolidated outputs from click or mask inputs
        # (we directly use their consolidated outputs during tracking)
        inference_state["consolidated_frame_inds"] = {
            "cond_frame_outputs": set(),  # set containing frame indices
            "non_cond_frame_outputs": set(),  # set containing frame indices
        }
        # metadata for each tracking frame (e.g. which direction it's tracked)
        inference_state["tracking_has_started"] = False
        inference_state["frames_already_tracked"] = {}
        # Warm up the visual backbone and cache the image feature on frame 0
        self._get_image_feature(inference_state, frame_idx=0, batch_size=1)
        return inference_state

    @classmethod
    def from_pretrained(cls, model_id: str, **kwargs) -> "SAM2VideoPredictor":
        """
        Load a pretrained model from the Hugging Face hub.

        Arguments:
          model_id (str): The Hugging Face repository ID.
          **kwargs: Additional arguments to pass to the model constructor.

        Returns:
          (SAM2VideoPredictor): The loaded model.
        """
        from sam2.build_sam import build_sam2_video_predictor_hf

        sam_model = build_sam2_video_predictor_hf(model_id, **kwargs)
        return sam_model

    def _obj_id_to_idx(self, inference_state, obj_id):
        """Map client-side object id to model-side object index."""
        obj_idx = inference_state["obj_id_to_idx"].get(obj_id, None)
        if obj_idx is not None:
            return obj_idx

        # This is a new object id not sent to the server before. We only allow adding
        # new objects *before* the tracking starts.
        allow_new_object = not inference_state["tracking_has_started"]
        if allow_new_object:
            # get the next object slot
            obj_idx = len(inference_state["obj_id_to_idx"])
            inference_state["obj_id_to_idx"][obj_id] = obj_idx
            inference_state["obj_idx_to_id"][obj_idx] = obj_id
            inference_state["obj_ids"] = list(inference_state["obj_id_to_idx"])
            # set up input and output structures for this object
            inference_state["point_inputs_per_obj"][obj_idx] = {}
            inference_state["mask_inputs_per_obj"][obj_idx] = {}
            inference_state["output_dict_per_obj"][obj_idx] = {
                "cond_frame_outputs": {},  # dict containing {frame_idx: <out>}
                "non_cond_frame_outputs": {},  # dict containing {frame_idx: <out>}
            }
            inference_state["temp_output_dict_per_obj"][obj_idx] = {
                "cond_frame_outputs": {},  # dict containing {frame_idx: <out>}
                "non_cond_frame_outputs": {},  # dict containing {frame_idx: <out>}
            }
            return obj_idx
        else:
            raise RuntimeError(
                f"Cannot add new object id {obj_id} after tracking starts. "
                f"All existing object ids: {inference_state['obj_ids']}. "
                f"Please call 'reset_state' to restart from scratch."
            )

    def _obj_idx_to_id(self, inference_state, obj_idx):
        """Map model-side object index to client-side object id."""
        return inference_state["obj_idx_to_id"][obj_idx]

    def _get_obj_num(self, inference_state):
        """Get the total number of unique object ids received so far in this session."""
        return len(inference_state["obj_idx_to_id"])

    @torch.inference_mode()
    def add_new_points_or_box(
        self,
        inference_state,
        frame_idx,
        obj_id,
        points=None,
        labels=None,
        clear_old_points=True,
        normalize_coords=True,
        box=None,
    ):
        """Add new points to a frame."""
        obj_idx = self._obj_id_to_idx(inference_state, obj_id)
        point_inputs_per_frame = inference_state["point_inputs_per_obj"][obj_idx]
        mask_inputs_per_frame = inference_state["mask_inputs_per_obj"][obj_idx]

        if (points is not None) != (labels is not None):
            raise ValueError("points and labels must be provided together")
        if points is None and box is None:
            raise ValueError("at least one of points or box must be provided as input")

        if points is None:
            points = torch.zeros(0, 2, dtype=torch.float32)
        elif not isinstance(points, torch.Tensor):
            points = torch.tensor(points, dtype=torch.float32)
        if labels is None:
            labels = torch.zeros(0, dtype=torch.int32)
        elif not isinstance(labels, torch.Tensor):
            labels = torch.tensor(labels, dtype=torch.int32)
        if points.dim() == 2:
            points = points.unsqueeze(0)  # add batch dimension
        if labels.dim() == 1:
            labels = labels.unsqueeze(0)  # add batch dimension

        # If `box` is provided, we add it as the first two points with labels 2 and 3
        # along with the user-provided points (consistent with how SAM 2 is trained).
        if box is not None:
            if not clear_old_points:
                raise ValueError(
                    "cannot add box without clearing old points, since "
                    "box prompt must be provided before any point prompt "
                    "(please use clear_old_points=True instead)"
                )
            if inference_state["tracking_has_started"]:
                warnings.warn(
                    "You are adding a box after tracking starts. SAM 2 may not always be "
                    "able to incorporate a box prompt for *refinement*. If you intend to "
                    "use box prompt as an *initial* input before tracking, please call "
                    "'reset_state' on the inference state to restart from scratch.",
                    category=UserWarning,
                    stacklevel=2,
                )
            if not isinstance(box, torch.Tensor):
                box = torch.tensor(box, dtype=torch.float32, device=points.device)
            box_coords = box.reshape(1, 2, 2)
            box_labels = torch.tensor([2, 3], dtype=torch.int32, device=labels.device)
            box_labels = box_labels.reshape(1, 2)
            points = torch.cat([box_coords, points], dim=1)
            labels = torch.cat([box_labels, labels], dim=1)

        if normalize_coords:
            video_H = inference_state["video_height"]
            video_W = inference_state["video_width"]
            points = points / torch.tensor([video_W, video_H]).to(points.device)
        # scale the (normalized) coordinates by the model's internal image size
        points = points * self.image_size
        points = points.to(inference_state["device"])
        labels = labels.to(inference_state["device"])

        if not clear_old_points:
            point_inputs = point_inputs_per_frame.get(frame_idx, None)
        else:
            point_inputs = None
        point_inputs = concat_points(point_inputs, points, labels)

        point_inputs_per_frame[frame_idx] = point_inputs
        mask_inputs_per_frame.pop(frame_idx, None)
        # If this frame hasn't been tracked before, we treat it as an initial conditioning
        # frame, meaning that the inputs points are to generate segments on this frame without
        # using any memory from other frames, like in SAM. Otherwise (if it has been tracked),
        # the input points will be used to correct the already tracked masks.
        is_init_cond_frame = frame_idx not in inference_state["frames_already_tracked"]
        # whether to track in reverse time order
        if is_init_cond_frame:
            reverse = False
        else:
            reverse = inference_state["frames_already_tracked"][frame_idx]["reverse"]
        obj_output_dict = inference_state["output_dict_per_obj"][obj_idx]
        obj_temp_output_dict = inference_state["temp_output_dict_per_obj"][obj_idx]
        # Add a frame to conditioning output if it's an initial conditioning frame or
        # if the model sees all frames receiving clicks/mask as conditioning frames.
        is_cond = is_init_cond_frame or self.add_all_frames_to_correct_as_cond
        storage_key = "cond_frame_outputs" if is_cond else "non_cond_frame_outputs"

        # Get any previously predicted mask logits on this object and feed it along with
        # the new clicks into the SAM mask decoder.
        prev_sam_mask_logits = None
        # lookup temporary output dict first, which contains the most recent output
        # (if not found, then lookup conditioning and non-conditioning frame output)
        prev_out = obj_temp_output_dict[storage_key].get(frame_idx)
        if prev_out is None:
            prev_out = obj_output_dict["cond_frame_outputs"].get(frame_idx)
            if prev_out is None:
                prev_out = obj_output_dict["non_cond_frame_outputs"].get(frame_idx)

        if prev_out is not None and prev_out["pred_masks"] is not None:
            device = inference_state["device"]
            prev_sam_mask_logits = prev_out["pred_masks"].to(device, non_blocking=True)
            # Clamp the scale of prev_sam_mask_logits to avoid rare numerical issues.
            prev_sam_mask_logits = torch.clamp(prev_sam_mask_logits, -32.0, 32.0)
        current_out, _ = self._run_single_frame_inference(
            inference_state=inference_state,
            output_dict=obj_output_dict,  # run on the slice of a single object
            frame_idx=frame_idx,
            batch_size=1,  # run on the slice of a single object
            is_init_cond_frame=is_init_cond_frame,
            point_inputs=point_inputs,
            mask_inputs=None,
            reverse=reverse,
            # Skip the memory encoder when adding clicks or mask. We execute the memory encoder
            # at the beginning of `propagate_in_video` (after user finalize their clicks). This
            # allows us to enforce non-overlapping constraints on all objects before encoding
            # them into memory.
            run_mem_encoder=False,
            prev_sam_mask_logits=prev_sam_mask_logits,
        )
        # Add the output to the output dict (to be used as future memory)
        obj_temp_output_dict[storage_key][frame_idx] = current_out

        # Resize the output mask to the original video resolution
        obj_ids = inference_state["obj_ids"]
        consolidated_out = self._consolidate_temp_output_across_obj(
            inference_state,
            frame_idx,
            is_cond=is_cond,
            run_mem_encoder=False,
            consolidate_at_video_res=True,
        )
        _, video_res_masks = self._get_orig_video_res_output(
            inference_state, consolidated_out["pred_masks_video_res"]
        )
        return frame_idx, obj_ids, video_res_masks

    def add_new_points(self, *args, **kwargs):
        """Deprecated method. Please use `add_new_points_or_box` instead."""
        return self.add_new_points_or_box(*args, **kwargs)

    @torch.inference_mode()
    def add_new_mask(
        self,
        inference_state,
        frame_idx,
        obj_id,
        mask,
    ):
        """Add new mask to a frame."""
        obj_idx = self._obj_id_to_idx(inference_state, obj_id)
        point_inputs_per_frame = inference_state["point_inputs_per_obj"][obj_idx]
        mask_inputs_per_frame = inference_state["mask_inputs_per_obj"][obj_idx]

        if not isinstance(mask, torch.Tensor):
            mask = torch.tensor(mask, dtype=torch.bool)
        assert mask.dim() == 2
        mask_H, mask_W = mask.shape
        mask_inputs_orig = mask[None, None]  # add batch and channel dimension
        mask_inputs_orig = mask_inputs_orig.float().to(inference_state["device"])

        # resize the mask if it doesn't match the model's image size
        if mask_H != self.image_size or mask_W != self.image_size:
            mask_inputs = torch.nn.functional.interpolate(
                mask_inputs_orig,
                size=(self.image_size, self.image_size),
                align_corners=False,
                mode="bilinear",
                antialias=True,  # use antialias for downsampling
            )
            mask_inputs = (mask_inputs >= 0.5).float()
        else:
            mask_inputs = mask_inputs_orig

        mask_inputs_per_frame[frame_idx] = mask_inputs
        point_inputs_per_frame.pop(frame_idx, None)
        # If this frame hasn't been tracked before, we treat it as an initial conditioning
        # frame, meaning that the inputs points are to generate segments on this frame without
        # using any memory from other frames, like in SAM. Otherwise (if it has been tracked),
        # the input points will be used to correct the already tracked masks.
        is_init_cond_frame = frame_idx not in inference_state["frames_already_tracked"]
        # whether to track in reverse time order
        if is_init_cond_frame:
            reverse = False
        else:
            reverse = inference_state["frames_already_tracked"][frame_idx]["reverse"]
        obj_output_dict = inference_state["output_dict_per_obj"][obj_idx]
        obj_temp_output_dict = inference_state["temp_output_dict_per_obj"][obj_idx]
        # Add a frame to conditioning output if it's an initial conditioning frame or
        # if the model sees all frames receiving clicks/mask as conditioning frames.
        is_cond = is_init_cond_frame or self.add_all_frames_to_correct_as_cond
        storage_key = "cond_frame_outputs" if is_cond else "non_cond_frame_outputs"

        current_out, _ = self._run_single_frame_inference(
            inference_state=inference_state,
            output_dict=obj_output_dict,  # run on the slice of a single object
            frame_idx=frame_idx,
            batch_size=1,  # run on the slice of a single object
            is_init_cond_frame=is_init_cond_frame,
            point_inputs=None,
            mask_inputs=mask_inputs,
            reverse=reverse,
            # Skip the memory encoder when adding clicks or mask. We execute the memory encoder
            # at the beginning of `propagate_in_video` (after user finalize their clicks). This
            # allows us to enforce non-overlapping constraints on all objects before encoding
            # them into memory.
            run_mem_encoder=False,
        )
        # Add the output to the output dict (to be used as future memory)
        obj_temp_output_dict[storage_key][frame_idx] = current_out

        # Resize the output mask to the original video resolution
        obj_ids = inference_state["obj_ids"]
        consolidated_out = self._consolidate_temp_output_across_obj(
            inference_state,
            frame_idx,
            is_cond=is_cond,
            run_mem_encoder=False,
            consolidate_at_video_res=True,
        )
        _, video_res_masks = self._get_orig_video_res_output(
            inference_state, consolidated_out["pred_masks_video_res"]
        )
        return frame_idx, obj_ids, video_res_masks

    def _get_orig_video_res_output(self, inference_state, any_res_masks):
        """
        Resize the object scores to the original video resolution (video_res_masks)
        and apply non-overlapping constraints for final output.
        """
        device = inference_state["device"]
        video_H = inference_state["video_height"]
        video_W = inference_state["video_width"]
        any_res_masks = any_res_masks.to(device, non_blocking=True)
        if any_res_masks.shape[-2:] == (video_H, video_W):
            video_res_masks = any_res_masks
        else:
            video_res_masks = torch.nn.functional.interpolate(
                any_res_masks,
                size=(video_H, video_W),
                mode="bilinear",
                align_corners=False,
            )
        if self.non_overlap_masks:
            video_res_masks = self._apply_non_overlapping_constraints(video_res_masks)
        return any_res_masks, video_res_masks

    def _consolidate_temp_output_across_obj(
        self,
        inference_state,
        frame_idx,
        is_cond,
        run_mem_encoder,
        consolidate_at_video_res=False,
    ):
        """
        Consolidate the per-object temporary outputs in `temp_output_dict_per_obj` on
        a frame into a single output for all objects, including
        1) fill any missing objects either from `output_dict_per_obj` (if they exist in
           `output_dict_per_obj` for this frame) or leave them as placeholder values
           (if they don't exist in `output_dict_per_obj` for this frame);
        2) if specified, rerun memory encoder after apply non-overlapping constraints
           on the object scores.
        """
        batch_size = self._get_obj_num(inference_state)
        storage_key = "cond_frame_outputs" if is_cond else "non_cond_frame_outputs"
        # Optionally, we allow consolidating the temporary outputs at the original
        # video resolution (to provide a better editing experience for mask prompts).
        if consolidate_at_video_res:
            assert not run_mem_encoder, "memory encoder cannot run at video resolution"
            consolidated_H = inference_state["video_height"]
            consolidated_W = inference_state["video_width"]
            consolidated_mask_key = "pred_masks_video_res"
        else:
            consolidated_H = consolidated_W = self.image_size // 4
            consolidated_mask_key = "pred_masks"

        # Initialize `consolidated_out`. Its "maskmem_features" and "maskmem_pos_enc"
        # will be added when rerunning the memory encoder after applying non-overlapping
        # constraints to object scores. Its "pred_masks" are prefilled with a large
        # negative value (NO_OBJ_SCORE) to represent missing objects.
        consolidated_out = {
            "maskmem_features": None,
            "maskmem_pos_enc": None,
            consolidated_mask_key: torch.full(
                size=(batch_size, 1, consolidated_H, consolidated_W),
                fill_value=NO_OBJ_SCORE,
                dtype=torch.float32,
                device=inference_state["storage_device"],
            ),
            "obj_ptr": torch.full(
                size=(batch_size, self.hidden_dim),
                fill_value=NO_OBJ_SCORE,
                dtype=torch.float32,
                device=inference_state["device"],
            ),
        }
        empty_mask_ptr = None
        for obj_idx in range(batch_size):
            obj_temp_output_dict = inference_state["temp_output_dict_per_obj"][obj_idx]
            obj_output_dict = inference_state["output_dict_per_obj"][obj_idx]
            out = obj_temp_output_dict[storage_key].get(frame_idx, None)
            # If the object doesn't appear in "temp_output_dict_per_obj" on this frame,
            # we fall back and look up its previous output in "output_dict_per_obj".
            # We look up both "cond_frame_outputs" and "non_cond_frame_outputs" in
            # "output_dict_per_obj" to find a previous output for this object.
            if out is None:
                out = obj_output_dict["cond_frame_outputs"].get(frame_idx, None)
            if out is None:
                out = obj_output_dict["non_cond_frame_outputs"].get(frame_idx, None)
            # If the object doesn't appear in "output_dict_per_obj" either, we skip it
            # and leave its mask scores to the default scores (i.e. the NO_OBJ_SCORE
            # placeholder above) and set its object pointer to be a dummy pointer.
            if out is None:
                # Fill in dummy object pointers for those objects without any inputs or
                # tracking outcomes on this frame (only do it under `run_mem_encoder=True`,
                # i.e. when we need to build the memory for tracking).
                if run_mem_encoder:
                    if empty_mask_ptr is None:
                        empty_mask_ptr = self._get_empty_mask_ptr(
                            inference_state, frame_idx
                        )
                    # fill object pointer with a dummy pointer (based on an empty mask)
                    consolidated_out["obj_ptr"][obj_idx : obj_idx + 1] = empty_mask_ptr
                continue
            # Add the temporary object output mask to consolidated output mask
            obj_mask = out["pred_masks"]
            consolidated_pred_masks = consolidated_out[consolidated_mask_key]
            if obj_mask.shape[-2:] == consolidated_pred_masks.shape[-2:]:
                consolidated_pred_masks[obj_idx : obj_idx + 1] = obj_mask
            else:
                # Resize first if temporary object mask has a different resolution
                resized_obj_mask = torch.nn.functional.interpolate(
                    obj_mask,
                    size=consolidated_pred_masks.shape[-2:],
                    mode="bilinear",
                    align_corners=False,
                )
                consolidated_pred_masks[obj_idx : obj_idx + 1] = resized_obj_mask
            consolidated_out["obj_ptr"][obj_idx : obj_idx + 1] = out["obj_ptr"]

        # Optionally, apply non-overlapping constraints on the consolidated scores
        # and rerun the memory encoder
        if run_mem_encoder:
            device = inference_state["device"]
            high_res_masks = torch.nn.functional.interpolate(
                consolidated_out["pred_masks"].to(device, non_blocking=True),
                size=(self.image_size, self.image_size),
                mode="bilinear",
                align_corners=False,
            )
            if self.non_overlap_masks_for_mem_enc:
                high_res_masks = self._apply_non_overlapping_constraints(high_res_masks)
            maskmem_features, maskmem_pos_enc = self._run_memory_encoder(
                inference_state=inference_state,
                frame_idx=frame_idx,
                batch_size=batch_size,
                high_res_masks=high_res_masks,
                is_mask_from_pts=True,  # these frames are what the user interacted with
            )
            consolidated_out["maskmem_features"] = maskmem_features
            consolidated_out["maskmem_pos_enc"] = maskmem_pos_enc

        return consolidated_out

    def _get_empty_mask_ptr(self, inference_state, frame_idx):
        """Get a dummy object pointer based on an empty mask on the current frame."""
        # A dummy (empty) mask with a single object
        batch_size = 1
        mask_inputs = torch.zeros(
            (batch_size, 1, self.image_size, self.image_size),
            dtype=torch.float32,
            device=inference_state["device"],
        )

        # Retrieve correct image features
        (
            _,
            _,
            current_vision_feats,
            current_vision_pos_embeds,
            feat_sizes,
        ) = self._get_image_feature(inference_state, frame_idx, batch_size)

        # Feed the empty mask and image feature above to get a dummy object pointer
        current_out = self.track_step(
            frame_idx=frame_idx,
            is_init_cond_frame=True,
            current_vision_feats=current_vision_feats,
            current_vision_pos_embeds=current_vision_pos_embeds,
            feat_sizes=feat_sizes,
            point_inputs=None,
            mask_inputs=mask_inputs,
            output_dict={},
            num_frames=inference_state["num_frames"],
            track_in_reverse=False,
            run_mem_encoder=False,
            prev_sam_mask_logits=None,
        )
        return current_out["obj_ptr"]

    @torch.inference_mode()
    def propagate_in_video_preflight(self, inference_state):
        """Prepare inference_state and consolidate temporary outputs before tracking."""
        # Tracking has started and we don't allow adding new objects until session is reset.
        inference_state["tracking_has_started"] = True
        batch_size = self._get_obj_num(inference_state)

        # Consolidate per-object temporary outputs in "temp_output_dict_per_obj" and
        # add them into "output_dict".
        temp_output_dict_per_obj = inference_state["temp_output_dict_per_obj"]
        output_dict = inference_state["output_dict"]
        # "consolidated_frame_inds" contains indices of those frames where consolidated
        # temporary outputs have been added (either in this call or any previous calls
        # to `propagate_in_video_preflight`).
        consolidated_frame_inds = inference_state["consolidated_frame_inds"]
        for is_cond in [False, True]:
            # Separately consolidate conditioning and non-conditioning temp outputs
            storage_key = "cond_frame_outputs" if is_cond else "non_cond_frame_outputs"
            # Find all the frames that contain temporary outputs for any objects
            # (these should be the frames that have just received clicks for mask inputs
            # via `add_new_points_or_box` or `add_new_mask`)
            temp_frame_inds = set()
            for obj_temp_output_dict in temp_output_dict_per_obj.values():
                temp_frame_inds.update(obj_temp_output_dict[storage_key].keys())
            consolidated_frame_inds[storage_key].update(temp_frame_inds)
            # consolidate the temporary output across all objects on this frame
            for frame_idx in temp_frame_inds:
                consolidated_out = self._consolidate_temp_output_across_obj(
                    inference_state, frame_idx, is_cond=is_cond, run_mem_encoder=True
                )
                # merge them into "output_dict" and also create per-object slices
                output_dict[storage_key][frame_idx] = consolidated_out
                self._add_output_per_object(
                    inference_state, frame_idx, consolidated_out, storage_key
                )
                clear_non_cond_mem = self.clear_non_cond_mem_around_input and (
                    self.clear_non_cond_mem_for_multi_obj or batch_size <= 1
                )
                if clear_non_cond_mem:
                    # clear non-conditioning memory of the surrounding frames
                    self._clear_non_cond_mem_around_input(inference_state, frame_idx)

            # clear temporary outputs in `temp_output_dict_per_obj`
            for obj_temp_output_dict in temp_output_dict_per_obj.values():
                obj_temp_output_dict[storage_key].clear()

        # edge case: if an output is added to "cond_frame_outputs", we remove any prior
        # output on the same frame in "non_cond_frame_outputs"
        for frame_idx in output_dict["cond_frame_outputs"]:
            output_dict["non_cond_frame_outputs"].pop(frame_idx, None)
        for obj_output_dict in inference_state["output_dict_per_obj"].values():
            for frame_idx in obj_output_dict["cond_frame_outputs"]:
                obj_output_dict["non_cond_frame_outputs"].pop(frame_idx, None)
        for frame_idx in consolidated_frame_inds["cond_frame_outputs"]:
            assert frame_idx in output_dict["cond_frame_outputs"]
            consolidated_frame_inds["non_cond_frame_outputs"].discard(frame_idx)

        # Make sure that the frame indices in "consolidated_frame_inds" are exactly those frames
        # with either points or mask inputs (which should be true under a correct workflow).
        all_consolidated_frame_inds = (
            consolidated_frame_inds["cond_frame_outputs"]
            | consolidated_frame_inds["non_cond_frame_outputs"]
        )
        input_frames_inds = set()
        for point_inputs_per_frame in inference_state["point_inputs_per_obj"].values():
            input_frames_inds.update(point_inputs_per_frame.keys())
        for mask_inputs_per_frame in inference_state["mask_inputs_per_obj"].values():
            input_frames_inds.update(mask_inputs_per_frame.keys())
        assert all_consolidated_frame_inds == input_frames_inds

    @torch.inference_mode()
    def propagate_in_video(
        self,
        inference_state,
        start_frame_idx=None,
        max_frame_num_to_track=None,
        reverse=False,
    ):
        """Propagate the input points across frames to track in the entire video."""
        self.propagate_in_video_preflight(inference_state)

        output_dict = inference_state["output_dict"]
        consolidated_frame_inds = inference_state["consolidated_frame_inds"]
        obj_ids = inference_state["obj_ids"]
        num_frames = inference_state["num_frames"]
        batch_size = self._get_obj_num(inference_state)
        if len(output_dict["cond_frame_outputs"]) == 0:
            raise RuntimeError("No points are provided; please add points first")
        clear_non_cond_mem = self.clear_non_cond_mem_around_input and (
            self.clear_non_cond_mem_for_multi_obj or batch_size <= 1
        )

        # set start index, end index, and processing order
        if start_frame_idx is None:
            # default: start from the earliest frame with input points
            start_frame_idx = min(output_dict["cond_frame_outputs"])
        if max_frame_num_to_track is None:
            # default: track all the frames in the video
            max_frame_num_to_track = num_frames
        if reverse:
            end_frame_idx = max(start_frame_idx - max_frame_num_to_track, 0)
            if start_frame_idx > 0:
                processing_order = range(start_frame_idx, end_frame_idx - 1, -1)
            else:
                processing_order = []  # skip reverse tracking if starting from frame 0
        else:
            end_frame_idx = min(
                start_frame_idx + max_frame_num_to_track, num_frames - 1
            )
            processing_order = range(start_frame_idx, end_frame_idx + 1)

        for frame_idx in tqdm(processing_order, desc="propagate in video"):
            # We skip those frames already in consolidated outputs (these are frames
            # that received input clicks or mask). Note that we cannot directly run
            # batched forward on them via `_run_single_frame_inference` because the
            # number of clicks on each object might be different.
            if frame_idx in consolidated_frame_inds["cond_frame_outputs"]:
                storage_key = "cond_frame_outputs"
                current_out = output_dict[storage_key][frame_idx]
                pred_masks = current_out["pred_masks"]
                if clear_non_cond_mem:
                    # clear non-conditioning memory of the surrounding frames
                    self._clear_non_cond_mem_around_input(inference_state, frame_idx)
            elif frame_idx in consolidated_frame_inds["non_cond_frame_outputs"]:
                storage_key = "non_cond_frame_outputs"
                current_out = output_dict[storage_key][frame_idx]
                pred_masks = current_out["pred_masks"]
            else:
                storage_key = "non_cond_frame_outputs"
                current_out, pred_masks = self._run_single_frame_inference(
                    inference_state=inference_state,
                    output_dict=output_dict,
                    frame_idx=frame_idx,
                    batch_size=batch_size,
                    is_init_cond_frame=False,
                    point_inputs=None,
                    mask_inputs=None,
                    reverse=reverse,
                    run_mem_encoder=True,
                )
                output_dict[storage_key][frame_idx] = current_out
            # Create slices of per-object outputs for subsequent interaction with each
            # individual object after tracking.
            self._add_output_per_object(
                inference_state, frame_idx, current_out, storage_key
            )
            inference_state["frames_already_tracked"][frame_idx] = {"reverse": reverse}

            # Resize the output mask to the original video resolution (we directly use
            # the mask scores on GPU for output to avoid any CPU conversion in between)
            _, video_res_masks = self._get_orig_video_res_output(
                inference_state, pred_masks
            )
            yield frame_idx, obj_ids, video_res_masks

    def _add_output_per_object(
        self, inference_state, frame_idx, current_out, storage_key
    ):
        """
        Split a multi-object output into per-object output slices and add them into
        `output_dict_per_obj`. The resulting slices share the same tensor storage.
        """
        maskmem_features = current_out["maskmem_features"]
        assert maskmem_features is None or isinstance(maskmem_features, torch.Tensor)

        maskmem_pos_enc = current_out["maskmem_pos_enc"]
        assert maskmem_pos_enc is None or isinstance(maskmem_pos_enc, list)

        output_dict_per_obj = inference_state["output_dict_per_obj"]
        for obj_idx, obj_output_dict in output_dict_per_obj.items():
            obj_slice = slice(obj_idx, obj_idx + 1)
            obj_out = {
                "maskmem_features": None,
                "maskmem_pos_enc": None,
                "pred_masks": current_out["pred_masks"][obj_slice],
                "obj_ptr": current_out["obj_ptr"][obj_slice],
            }
            if maskmem_features is not None:
                obj_out["maskmem_features"] = maskmem_features[obj_slice]
            if maskmem_pos_enc is not None:
                obj_out["maskmem_pos_enc"] = [x[obj_slice] for x in maskmem_pos_enc]
            obj_output_dict[storage_key][frame_idx] = obj_out

    @torch.inference_mode()
    def reset_state(self, inference_state):
        """Remove all input points or mask in all frames throughout the video."""
        self._reset_tracking_results(inference_state)
        # Remove all object ids
        inference_state["obj_id_to_idx"].clear()
        inference_state["obj_idx_to_id"].clear()
        inference_state["obj_ids"].clear()
        inference_state["point_inputs_per_obj"].clear()
        inference_state["mask_inputs_per_obj"].clear()
        inference_state["output_dict_per_obj"].clear()
        inference_state["temp_output_dict_per_obj"].clear()

    def _reset_tracking_results(self, inference_state):
        """Reset all tracking inputs and results across the videos."""
        for v in inference_state["point_inputs_per_obj"].values():
            v.clear()
        for v in inference_state["mask_inputs_per_obj"].values():
            v.clear()
        for v in inference_state["output_dict_per_obj"].values():
            v["cond_frame_outputs"].clear()
            v["non_cond_frame_outputs"].clear()
        for v in inference_state["temp_output_dict_per_obj"].values():
            v["cond_frame_outputs"].clear()
            v["non_cond_frame_outputs"].clear()
        inference_state["output_dict"]["cond_frame_outputs"].clear()
        inference_state["output_dict"]["non_cond_frame_outputs"].clear()
        inference_state["consolidated_frame_inds"]["cond_frame_outputs"].clear()
        inference_state["consolidated_frame_inds"]["non_cond_frame_outputs"].clear()
        inference_state["tracking_has_started"] = False
        inference_state["frames_already_tracked"].clear()

    def _get_image_feature(self, inference_state, frame_idx, batch_size):
        """Compute the image features on a given frame."""
        # Look up in the cache first
        image, backbone_out = inference_state["cached_features"].get(
            frame_idx, (None, None)
        )
        if backbone_out is None:
            # Cache miss -- we will run inference on a single image
            device = inference_state["device"]
            image = inference_state["images"][frame_idx].to(device).float().unsqueeze(0)
            backbone_out = self.forward_image(image)
            # Cache the most recent frame's feature (for repeated interactions with
            # a frame; we can use an LRU cache for more frames in the future).
            inference_state["cached_features"] = {frame_idx: (image, backbone_out)}

        # expand the features to have the same dimension as the number of objects
        expanded_image = image.expand(batch_size, -1, -1, -1)
        expanded_backbone_out = {
            "backbone_fpn": backbone_out["backbone_fpn"].copy(),
            "vision_pos_enc": backbone_out["vision_pos_enc"].copy(),
        }
        for i, feat in enumerate(expanded_backbone_out["backbone_fpn"]):
            expanded_backbone_out["backbone_fpn"][i] = feat.expand(
                batch_size, -1, -1, -1
            )
        for i, pos in enumerate(expanded_backbone_out["vision_pos_enc"]):
            pos = pos.expand(batch_size, -1, -1, -1)
            expanded_backbone_out["vision_pos_enc"][i] = pos

        features = self._prepare_backbone_features(expanded_backbone_out)
        features = (expanded_image,) + features
        return features

    def _run_single_frame_inference(
        self,
        inference_state,
        output_dict,
        frame_idx,
        batch_size,
        is_init_cond_frame,
        point_inputs,
        mask_inputs,
        reverse,
        run_mem_encoder,
        prev_sam_mask_logits=None,
    ):
        """Run tracking on a single frame based on current inputs and previous memory."""
        # Retrieve correct image features
        (
            _,
            _,
            current_vision_feats,
            current_vision_pos_embeds,
            feat_sizes,
        ) = self._get_image_feature(inference_state, frame_idx, batch_size)

        # point and mask should not appear as input simultaneously on the same frame
        assert point_inputs is None or mask_inputs is None
        current_out = self.track_step(
            frame_idx=frame_idx,
            is_init_cond_frame=is_init_cond_frame,
            current_vision_feats=current_vision_feats,
            current_vision_pos_embeds=current_vision_pos_embeds,
            feat_sizes=feat_sizes,
            point_inputs=point_inputs,
            mask_inputs=mask_inputs,
            output_dict=output_dict,
            num_frames=inference_state["num_frames"],
            track_in_reverse=reverse,
            run_mem_encoder=run_mem_encoder,
            prev_sam_mask_logits=prev_sam_mask_logits,
        )

        # optionally offload the output to CPU memory to save GPU space
        storage_device = inference_state["storage_device"]
        maskmem_features = current_out["maskmem_features"]
        if maskmem_features is not None:
            maskmem_features = maskmem_features.to(torch.bfloat16)
            maskmem_features = maskmem_features.to(storage_device, non_blocking=True)
        pred_masks_gpu = current_out["pred_masks"]
        # potentially fill holes in the predicted masks
        if self.fill_hole_area > 0:
            pred_masks_gpu = fill_holes_in_mask_scores(
                pred_masks_gpu, self.fill_hole_area
            )
        pred_masks = pred_masks_gpu.to(storage_device, non_blocking=True)
        # "maskmem_pos_enc" is the same across frames, so we only need to store one copy of it
        maskmem_pos_enc = self._get_maskmem_pos_enc(inference_state, current_out)
        # object pointer is a small tensor, so we always keep it on GPU memory for fast access
        obj_ptr = current_out["obj_ptr"]
        # make a compact version of this frame's output to reduce the state size
        compact_current_out = {
            "maskmem_features": maskmem_features,
            "maskmem_pos_enc": maskmem_pos_enc,
            "pred_masks": pred_masks,
            "obj_ptr": obj_ptr,
        }
        return compact_current_out, pred_masks_gpu

    def _run_memory_encoder(
        self, inference_state, frame_idx, batch_size, high_res_masks, is_mask_from_pts
    ):
        """
        Run the memory encoder on `high_res_masks`. This is usually after applying
        non-overlapping constraints to object scores. Since their scores changed, their
        memory also need to be computed again with the memory encoder.
        """
        # Retrieve correct image features
        _, _, current_vision_feats, _, feat_sizes = self._get_image_feature(
            inference_state, frame_idx, batch_size
        )
        maskmem_features, maskmem_pos_enc = self._encode_new_memory(
            current_vision_feats=current_vision_feats,
            feat_sizes=feat_sizes,
            pred_masks_high_res=high_res_masks,
            is_mask_from_pts=is_mask_from_pts,
        )

        # optionally offload the output to CPU memory to save GPU space
        storage_device = inference_state["storage_device"]
        maskmem_features = maskmem_features.to(torch.bfloat16)
        maskmem_features = maskmem_features.to(storage_device, non_blocking=True)
        # "maskmem_pos_enc" is the same across frames, so we only need to store one copy of it
        maskmem_pos_enc = self._get_maskmem_pos_enc(
            inference_state, {"maskmem_pos_enc": maskmem_pos_enc}
        )
        return maskmem_features, maskmem_pos_enc

    def _get_maskmem_pos_enc(self, inference_state, current_out):
        """
        `maskmem_pos_enc` is the same across frames and objects, so we cache it as
        a constant in the inference session to reduce session storage size.
        """
        model_constants = inference_state["constants"]
        # "out_maskmem_pos_enc" should be either a list of tensors or None
        out_maskmem_pos_enc = current_out["maskmem_pos_enc"]
        if out_maskmem_pos_enc is not None:
            if "maskmem_pos_enc" not in model_constants:
                assert isinstance(out_maskmem_pos_enc, list)
                # only take the slice for one object, since it's same across objects
                maskmem_pos_enc = [x[0:1].clone() for x in out_maskmem_pos_enc]
                model_constants["maskmem_pos_enc"] = maskmem_pos_enc
            else:
                maskmem_pos_enc = model_constants["maskmem_pos_enc"]
            # expand the cached maskmem_pos_enc to the actual batch size
            batch_size = out_maskmem_pos_enc[0].size(0)
            expanded_maskmem_pos_enc = [
                x.expand(batch_size, -1, -1, -1) for x in maskmem_pos_enc
            ]
        else:
            expanded_maskmem_pos_enc = None
        return expanded_maskmem_pos_enc

    def _clear_non_cond_mem_around_input(self, inference_state, frame_idx):
        """
        Remove the non-conditioning memory around the input frame. When users provide
        correction clicks, the surrounding frames' non-conditioning memories can still
        contain outdated object appearance information and could confuse the model.

        This method clears those non-conditioning memories surrounding the interacted
        frame to avoid giving the model both old and new information about the object.
        """
        r = self.memory_temporal_stride_for_eval
        frame_idx_begin = frame_idx - r * self.num_maskmem
        frame_idx_end = frame_idx + r * self.num_maskmem
        output_dict = inference_state["output_dict"]
        non_cond_frame_outputs = output_dict["non_cond_frame_outputs"]
        for t in range(frame_idx_begin, frame_idx_end + 1):
            non_cond_frame_outputs.pop(t, None)
            for obj_output_dict in inference_state["output_dict_per_obj"].values():
                obj_output_dict["non_cond_frame_outputs"].pop(t, None)