SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2502.13508 · ICLR · 2025

VLAS: Vision-Language-Action Model With Speech Instructions For Customized Robot Manipulation

Han Zhao, Min Zhang, Wei Zhao, Donglin Wang, Shuanghao Bai, Pengxiang Ding, Zhefei Gong

arXiv · PDF · Open in the Atlas

Code that ran

6 of 11 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (2 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 3 have no contract verdict; 3 failures of our instrument; 5 no recorded run), measured 2026-11-22; 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/2502.13508.json · badge: /api/ran/2502.13508.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
whichwhichgone/VLAS reimplementation (classifier label)
classifier basis: classifier_2025
5 ran · 5 no recorded run
copy not recorded community
not in the paper's declared repositories
1 ran · 0 no recorded run
FunctionStatusWhere it lives
AsrLlavaConfig Ran
ran on a synthesized input · constructed an object; no result was computed
whichwhichgone/VLAS/llava/model/language_model/asr_llava.py
code served (permissive licence) · get_code("3d3e1a0dec8eb859")
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, AsrLlavaConfig: 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 transformers import LlamaConfig

torch.manual_seed(0)

class AsrLlavaConfig(LlamaConfig):
    model_type = 'asr_llava'

ARGS = {
    'vocab_size': 32000,
    'hidden_size': 512,
    'intermediate_size': 1024,
    'num_hidden_layers': 4,
    'num_attention_heads': 8,
    'num_key_value_heads': 8,
    'hidden_act': 'silu',
    'initializer_range': 0.02,
    'initializer_factor': 1.0,
    'rms_norm_eps': 1e-6,
    'use_cache': True,
    'pad_token_id': 0,
    'bos_token_id': 1,
    'eos_token_id': 2,
    'pretraining_tp': 1,
    'tie_word_embeddings': False,
    'rope_theta': 10000.0,
    'rope_scaling': None,
    'attention_bias': False,
    'attention_dropout': 0.0,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, transformers — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type AsrLlavaConfig — 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 AsrLlavaConfig 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
DownsampleAudioProjector Ran
ran on a synthesized input · constructed an object; no result was computed
whichwhichgone/VLAS/llava/model/language_model/asr_llava.py
code served (permissive licence) · get_code("c15657d0345ff9ea")
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, DownsampleAudioProjector: 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)

class Config:
    def __init__(self):
        self.mm_hidden_size_audio = 8
        self.hidden_size = 16

config = Config()
reduce_factor = 5

ARGS = {
    "config": config,
    "reduce_factor": reduce_factor,
}
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 DownsampleAudioProjector — 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 DownsampleAudioProjector 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
build_audio_projector Ran
ran; our contract draft was wrong, not the code
whichwhichgone/VLAS/llava/model/language_model/asr_llava.py
code served (permissive licence) · get_code("66730018fe001b2c")
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 DownsampleAudioProjector. 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)

class Config:
    def __init__(self):
        self.mm_hidden_size_audio = 8
        self.hidden_size = 16

config = Config()

ARGS = {
    "config": config,
    "delay_load": False,
}
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 DownsampleAudioProjector — a DownsampleAudioProjector output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DownsampleAudioProjector 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
build_audio_tower Ran
ran on a synthesized input
whichwhichgone/VLAS/llava/model/language_model/asr_llava.py
code served (permissive licence) · get_code("eff7947bd6160de0")
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) and returned WhisperAudioTower. 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 unittest.mock import Mock, MagicMock

torch.manual_seed(0)

# Create a minimal config object that has the required attributes
class MinimalConfig:
    def __init__(self):
        self.mm_audio_tower = 'openai/whisper-tiny'
        self.unfreeze_mm_audio_tower = False

# Create a minimal args object
audio_tower_cfg = MinimalConfig()

ARGS = {
    'audio_tower_cfg': audio_tower_cfg
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch, transformers — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type WhisperAudioTower — a WhisperAudioTower output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a WhisperAudioTower 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_llm_data Ran
ran; our contract draft was wrong, not the code
whichwhichgone/VLAS/playground/calvin_data/calvin2json.py
code served (permissive licence) · get_code("6c406b8d0051cfe9")
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) and returned dict. 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 numpy as np
from pathlib import Path

# Create sample data for the function parameters
instruction = "Pick up the red cube and place it on the table"
task = "object_manipulation"
split = "train"
sample = "trajectory_001/frame_042.jpg"

# Create next_actions as a list of numpy arrays
next_actions = [
    np.array([[0.1, 0.2], [0.3, 0.4]]),
    np.array([[0.5, 0.6], [0.7, 0.8]])
]

# Create robot_obs as a numpy array
robot_obs = np.array([
    [1.0, 2.0, 3.0, 4.0],
    [5.0, 6.0, 7.0, 8.0],
    [9.0, 10.0, 11.0, 12.0]
])

ARGS = {
    "instruction": instruction,
    "task": task,
    "split": split,
    "sample": sample,
    "next_actions": next_actions,
    "robot_obs": robot_obs
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
unpad_image Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
this paper's copy was not recorded; identical code first harvested from haotian-liu/LLaVA
pointer only · get_code("55c32993da87759b")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

AsrLlavaForCausalLM Not yet run
unverified (harvested, no recorded run)
whichwhichgone/VLAS/llava/model/language_model/asr_llava.py
code served (permissive licence) · get_code("d5366fdc0ee9e59c")
How it ran

no invocation recorded.

AsrLlavaMetaForCausalLM Not yet run
unverified (harvested, no recorded run)
whichwhichgone/VLAS/llava/model/language_model/asr_llava.py
code served (permissive licence) · get_code("c30216d7b3f4633d")
How it ran

no invocation recorded.

AsrLlavaMetaModel Not yet run
unverified (harvested, no recorded run)
whichwhichgone/VLAS/llava/model/language_model/asr_llava.py
code served (permissive licence) · get_code("b2da05619cf2fae9")
How it ran

no invocation recorded.

AsrLlavaModel Not yet run
unverified (harvested, no recorded run)
whichwhichgone/VLAS/llava/model/language_model/asr_llava.py
code served (permissive licence) · get_code("047512cce4444df5")
How it ran

no invocation recorded.

WhisperAudioTower Not yet run
unverified (harvested, no recorded run)
whichwhichgone/VLAS/llava/model/language_model/asr_llava.py
code served (permissive licence) · get_code("45f0b796bdb06c4f")
How it ran

no invocation recorded.

Also present in other papers' code

Of 11 samples lifted from this paper's linked repositories, 11 carry a function identity and 0 do not. Of the 11 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, 1 later) and 10 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
unpad_image 69 other papers
44 dated earlier, 20 later, 5 same year, order not determinable · 5 only as an assembly brick
2304.08485 (2023-04)
Visual Instruction Tuning
2309.09958 (2023-09)
An Empirical Study of Scaling Instruct-Tuned Large Multimodal Models
2310.03744 (2023-10)
Improved Baselines with Visual Instruction Tuning
and 66 more not named here

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

Vision-language-action models (VLAs) have become increasingly popular in robot manipulation for their end-to-end design and remarkable performance. However, existing VLAs rely heavily on vision-language models (VLMs) that only support text-based instructions, neglecting the more natural speech modality for human-robot interaction. Traditional speech integration methods usually involves a separate speech recognition system, which complicates the model and introduces error propagation. Moreover, the transcription procedure would lose nonsemantic information in the raw speech, such as voiceprint, which may be crucial for robots to successfully complete customized tasks. To overcome above challenges, we propose VLAS, a novel end-to-end VLA that integrates speech recognition directly into the robot policy model. VLAS allows the robot to understand spoken commands through inner speech-text alignment and produces corresponding actions to fulfill the task. We also present two new datasets, SQA and CSI, to support a three-stage tuning process for speech instructions, which empowers VLAS with the ability of multimodal interaction across text, image, speech, and robot actions. Taking a step further, a voice retrieval-augmented generation (RAG) paradigm is designed to enable our model to effectively handle tasks that require individual-specific knowledge. Our extensive experiments show that VLAS can effectively accomplish robot manipulation tasks with diverse speech commands, offering a seamless and customized interaction experience.

For agents

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

get_harvested_code_for_paper("2502.13508")
get_code_for_paper("2502.13508")
have("2502.13508")

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

curl https://syntology.ai/api/ran/2502.13508.json

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

[![Syntology run record](https://syntology.ai/api/ran/2502.13508.svg)](https://syntology.ai/paper/2502.13508)

Connect an agent — have() is free.