SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2409.09016 · NeurIPS · 2024

Closed-Loop Visuomotor Control with Generative Expectation for Robotic Manipulation

Ping Luo, Yi Ma, Junchi Yan, Hongyang Li, Li Chen, Guyue Zhou, Shanghai Lab, Conv, Qingwen Bu, Yanchao Yang, Hku, Multimodal Encoder, and 3 more

arXiv · PDF · Open in the Atlas

Code that ran

13 of 14 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (9 of them constructed an object rather than computing a result; 1 honoured the contract we drafted, 0 violated it and 11 have no contract verdict; 1 failure of our instrument; 1 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2409.09016.json · badge: /api/ran/2409.09016.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
OpenDriveLab/CLOVER marked official in Papers with Code
Papers with Code archive
13 ran · 1 no recorded run
FunctionStatusWhere it lives
Attention Ran
ran on a synthesized input
OpenDriveLab/CLOVER/FeedbackPolicy/models/policy.py
code served (permissive licence) · get_code("08c3841b48475b05")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, ndim 3, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

# INIT_ARGS: parameters for Attention.__init__
INIT_ARGS = {
    'embed_dim': 8,
    'n_heads': 2,
    'dropout': 0.1
}

# ARGS: parameters for Attention.forward
# B=2 (batch), N=5 (sequence length), C=8 (embed_dim)
x = torch.randn(2, 5, 8)

# Optional mask: can be 2D (bsz, seq) or 4D (bsz, n_heads, seq, seq)
# Using 2D mask for simplicity
mask = torch.ones(2, 5)
mask[0, 3:] = 0  # Mask out last 2 positions in first batch element
mask[1, 4:] = 0  # Mask out last position in second batch element

ARGS = {
    'x': x,
    'mask': mask
}
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
Imports
einops, numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 3, dtype family float — no output shape is stated: observe_signature.py observed this class with its constructor sizes canonicalised where it could, which the node does not record, so its shape is not stated as this setup block's
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
Block Ran
ran on a synthesized input · constructed an object; no result was computed
OpenDriveLab/CLOVER/FeedbackPolicy/models/policy.py
code served (permissive licence) · get_code("92c6e1689d974314")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, Block: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# Define ARGS dict with all parameters for Block.__init__
ARGS = {
    "embed_dim": 8,
    "n_heads": 2,
    "mlp_ratio": 4.0,
    "dropout": 0.0,
    "do_rms_norm": False,
    "do_swish_glu": False,
    "do_layer_scale": False,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
einops, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Block — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a Block output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
ConvFuser Ran
ran on a synthesized input · constructed an object; no result was computed
OpenDriveLab/CLOVER/FeedbackPolicy/models/policy.py
code served (permissive licence) · get_code("f432d25a96b1dca6")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, ConvFuser: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn
from einops import rearrange

torch.manual_seed(0)

class ConvFuser(nn.Module):

    def __init__(self, in_channels: int, out_channels: int, patch_num: int) -> None:
        super().__init__()
        self.in_channels = in_channels
        self.out_channels = out_channels
        self.patch_num = patch_num
        self.conv = nn.Conv2d(in_channels, out_channels, kernel_size=3, padding=1)
        self.channel_selection = nn.Sequential(nn.Linear(out_channels, out_channels), nn.Sigmoid())

    def forward(self, inputs_rgb: torch.Tensor, inputs_depth: torch.Tensor) -> torch.Tensor:
        inputs = torch.cat([inputs_rgb, inputs_depth], dim=-1)
        inputs = rearrange(inputs, 'b (h w) d -> b d h w', h=self.patch_num, w=self.patch_num)
        feature = self.conv(inputs)
        feature = rearrange(feature, 'b d h w -> b (h w) d')
        selection_weights = self.channel_selection(feature.mean(dim=1))
        feature = feature * selection_weights.unsqueeze(1)
        return feature

in_channels_val = 8
out_channels_val = 16
patch_num_val = 4

model = ConvFuser(in_channels=in_channels_val, out_channels=out_channels_val, patch_num=patch_num_val)

batch_size = 2
patch_total = patch_num_val * patch_num_val
rgb_channels = 3
depth_channels = in_channels_val - rgb_channels

inputs_rgb = torch.randn(batch_size, patch_total, rgb_channels)
inputs_depth = torch.randn(batch_size, patch_total, depth_channels)

ARGS = {
    "in_channels": in_channels_val,
    "out_channels": out_channels_val,
    "patch_num": patch_num_val,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
einops, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ConvFuser — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a ConvFuser output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
LayerScale Ran
ran on a synthesized input · constructed an object; no result was computed
OpenDriveLab/CLOVER/FeedbackPolicy/models/policy.py
code served (permissive licence) · get_code("07517d004d8d3758")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, LayerScale: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

ARGS = {
    "dim": 4,
    "init_values": 0.1
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type LayerScale — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a LayerScale output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
MAPAttention Ran
ran on a synthesized input
OpenDriveLab/CLOVER/FeedbackPolicy/models/policy.py
code served (permissive licence) · get_code("7d33cd646ac2ec81")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned Tensor, ndim 3, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

# INIT_ARGS for MAPAttention.__init__
INIT_ARGS = {
    'embed_dim': 8,
    'n_heads': 2
}

# ARGS for MAPAttention.forward
# seed: (B_s, K, C_s) - batch of seed vectors
# x: (B_x, N, C_x) - batch of pool inputs
# attention_mask: optional mask of shape (B, N)
ARGS = {
    'seed': torch.randn(2, 3, 8),  # batch=2, K=3 seed vectors, embed_dim=8
    'x': torch.randn(2, 5, 8),      # batch=2, N=5 pool inputs, embed_dim=8
    'attention_mask': torch.ones(2, 5)  # batch=2, N=5
}
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 3, dtype family float — no output shape is stated: observe_signature.py observed this class with its constructor sizes canonicalised where it could, which the node does not record, so its shape is not stated as this setup block's
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
MLPActionGripperHead Ran
ran on a synthesized input · constructed an object; no result was computed
OpenDriveLab/CLOVER/FeedbackPolicy/models/policy.py
code served (permissive licence) · get_code("8988ff4b78a1a749")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, MLPActionGripperHead: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# MLPActionGripperHead only takes hidden_size as a parameter
# Looking at the forward method, it expects input x to be compatible with nn.Linear(hidden_size, 1024)
# So x should have shape (batch_size, hidden_size)

hidden_size = 256

ARGS = {
    "hidden_size": hidden_size
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type MLPActionGripperHead — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a MLPActionGripperHead output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
MLPActionVelocityHead_Tanh Ran
ran on a synthesized input · constructed an object; no result was computed
OpenDriveLab/CLOVER/FeedbackPolicy/models/policy.py
code served (permissive licence) · get_code("57554a8e59f95d85")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, MLPActionVelocityHead_Tanh: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

ARGS = {
    "hidden_size": 256
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type MLPActionVelocityHead_Tanh — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a MLPActionVelocityHead_Tanh output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
PatchEmbed Ran
ran on a synthesized input · constructed an object; no result was computed
OpenDriveLab/CLOVER/FeedbackPolicy/models/policy.py
code served (permissive licence) · get_code("8164f7520779e070")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, PatchEmbed: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn
from einops import rearrange

torch.manual_seed(0)

class PatchEmbed(nn.Module):

    def __init__(self, resolution: int, patch_size: int, embed_dim: int, in_channels: int=3, flatten: bool=True):
        super().__init__()
        self.resolution, self.patch_size = ((resolution, resolution), (patch_size, patch_size))
        self.grid_size = (self.resolution[0] // self.patch_size[0], self.resolution[1] // self.patch_size[1])
        self.num_patches = self.grid_size[0] * self.grid_size[1]
        self.flatten = flatten
        self.proj = nn.Conv2d(in_channels, embed_dim, kernel_size=self.patch_size, stride=self.patch_size)

    def forward(self, patches: torch.Tensor) -> torch.Tensor:
        patch_embeddings = self.proj(patches)
        if self.flatten:
            return rearrange(patch_embeddings, 'bsz embed patch_h patch_w -> bsz (patch_h patch_w) embed')
        return patch_embeddings

ARGS = {
    "resolution": 8,
    "patch_size": 4,
    "embed_dim": 16,
    "in_channels": 3,
    "flatten": True
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
einops, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type PatchEmbed — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a PatchEmbed output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
RMSNorm Ran
ran on a synthesized input · constructed an object; no result was computed
OpenDriveLab/CLOVER/FeedbackPolicy/models/policy.py
code served (permissive licence) · get_code("008a90cb086e88c6")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, RMSNorm: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# RMSNorm expects:
# - dim: int (dimension for the scale parameter and g parameter)
# - eps: float (optional, default 1e-08)
# 
# The forward method expects x to be a tensor where we can compute norm over the last dimension

# Create a small input tensor for testing
# Shape: (batch=2, sequence=3, dim=4)
x = torch.randn(2, 3, 4)

ARGS = {
    'dim': 4,
    'eps': 1e-08
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type RMSNorm — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a RMSNorm output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
SwishGLU Ran
ran on a synthesized input · constructed an object; no result was computed
OpenDriveLab/CLOVER/FeedbackPolicy/models/policy.py
code served (permissive licence) · get_code("52bc3049bf0e3ed8")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, SwishGLU: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# SwishGLU takes in_dim and out_dim as parameters
# in_dim: input dimension
# out_dim: output dimension

in_dim = 4
out_dim = 3

ARGS = {
    "in_dim": in_dim,
    "out_dim": out_dim,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type SwishGLU — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a SwishGLU output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
TokenAggregation Ran
ran on a synthesized input · constructed an object; no result was computed
OpenDriveLab/CLOVER/FeedbackPolicy/models/policy.py
code served (permissive licence) · get_code("9ed32659b7758f44")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, TokenAggregation: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn
from einops import repeat

torch.manual_seed(0)

# Define ARGS dict with all parameters for TokenAggregation.__init__
ARGS = {
    'n_latents': 2,
    'embed_dim': 8,
    'n_heads': 2,
    'mlp_ratio': 4.0,
    'do_rms_norm': True,
    'do_swish_glu': True,
}

# Additional setup to verify the module works with forward pass
if __name__ == "__main__":
    # Instantiate the module
    token_agg = nn.TokenAggregation(**ARGS)
    
    # Create minimal tensors for forward pass
    batch_size = 2
    seq_len = 3
    embed_dim = ARGS['embed_dim']
    n_latents = ARGS['n_latents']
    
    # x: input tensor (batch_size, seq_len, embed_dim)
    x = torch.randn(batch_size, seq_len, embed_dim)
    
    # latents: (n_latents, embed_dim) - will be repeated to (batch_size, n_latents, embed_dim)
    latents = torch.randn(n_latents, embed_dim)
    
    # Optional mask: (batch_size, seq_len) with 1s and 0s
    mask = torch.ones(batch_size, seq_len)
    
    # Call forward
    output = token_agg(x, latents=latents, mask=mask)
    print(f"Output shape: {output.shape}")
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
einops, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type TokenAggregation — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a TokenAggregation output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_1D_sine_cosine Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
OpenDriveLab/CLOVER/FeedbackPolicy/models/policy.py
code served (permissive licence) · get_code("460db904f1c5c885")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 2, dtype family float, shape (4, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

dim = 4
pos = np.array([0, 1, 2, 3], dtype=np.float32)

ARGS = {
    "dim": dim,
    "pos": pos,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (4, 4) — shape (4, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_2D_position_embeddings Ran
ran, honoured the contract we drafted
OpenDriveLab/CLOVER/FeedbackPolicy/models/policy.py
code served (permissive licence) · get_code("b7dffc56b907b1e2")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 2, dtype family float, shape (17, 8) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

ARGS = {
    "embed_dim": 8,
    "grid_size": 4,
    "cls_token": True,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (17, 8) — shape (17, 8) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
FeedbackDrivenPolicy Not yet run
unverified (harvested, no recorded run)
OpenDriveLab/CLOVER/FeedbackPolicy/models/policy.py
code served (permissive licence) · get_code("0e9447c615fe1a69")
How it ran

no invocation recorded.

Also present in other papers' code

Of 14 samples lifted from this paper's linked repositories, 14 carry a function identity and 0 do not. Of the 14 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 0 later) and 13 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
get_1D_sine_cosine
brick of this paper's assembly: helper
1 other paper
1 dated earlier
2301.02555 (2023-01)
"No, to the Right" -- Online Language Corrections for Robotic Manipulation via Shared Autonomy

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Despite significant progress in robotics and embodied AI in recent years, deploying robots for long-horizon tasks remains a great challenge. Majority of prior arts adhere to an open-loop philosophy and lack real-time feedback, leading to error accumulation and undesirable robustness. A handful of approaches have endeavored to establish feedback mechanisms leveraging pixel-level differences or pre-trained visual representations, yet their efficacy and adaptability have been found to be constrained. Inspired by classic closed-loop control systems, we propose CLOVER, a closed-loop visuomotor control framework that incorporates feedback mechanisms to improve adaptive robotic control. CLOVER consists of a text-conditioned video diffusion model for generating visual plans as reference inputs, a measurable embedding space for accurate error quantification, and a feedback-driven controller that refines actions from feedback and initiates replans as needed. Our framework exhibits notable advancement in real-world robotic tasks and achieves state-of-the-art on CALVIN benchmark, improving by 8% over previous open-loop counterparts. Code and checkpoints are maintained at https://github.com/OpenDriveLab/CLOVER. Preprint. * Equal contribution. ♮ Project lead(s).

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2409.09016")
get_code_for_paper("2409.09016")
have("2409.09016")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2409.09016.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2409.09016.svg)](https://syntology.ai/paper/2409.09016)

Connect an agent — have() is free.