SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2501.19098 · 2025

$\infty$-Video: A Training-Free Approach to Long Video Understanding via Continuous-Time Memory Consolidation

arXiv · PDF · Open in the Atlas

Code that ran

2 of 3 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 1 honoured the contract we drafted, 0 violated it and 0 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/2501.19098.json · badge: /api/ran/2501.19098.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
deep-spin/infinite-video canonical (classifier label)
classifier basis: classifier_2025
2 ran · 1 no recorded run
FunctionStatusWhere it lives
check_answer_egoschema Ran
ran, honoured the contract we drafted
deep-spin/infinite-video/infty-VideoChat2/eval_code/run_nextqa_mistral.py
code served (permissive licence) · get_code("5b14ccbef0d4360a")
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 int, ndim 0, dtype family int, shape () 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 torch

# Construct valid arguments for check_answer_egoschema
pred = "(A) This is the correct answer"
qid = "q1"
ans_dict = {
    "q1": {
        "content": "This is the correct answer.",
        "answer": "(A)"
    }
}

ARGS = {
    "pred": pred,
    "qid": qid,
    "ans_dict": ans_dict
}
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 int, ndim 0, dtype family int, shape () — shape () 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
process_qa Ran
ran; our contract draft was wrong, not the code
deep-spin/infinite-video/infty-Video-LLaMA/eval_code/eval/run_inference_inf_video_llama_egochema.py
code served (permissive licence) · get_code("9ed8b954fac93d9d")
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 str. 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 MockVideoEncoder(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(3, 4)
    
    def forward(self, x):
        return self.linear(x)

class MockModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.encoder = MockVideoEncoder()
        self.long_memory_buffer = []
        self.temp_short_memory = []
        self.short_memory_buffer = []
    
    def encode_videoQformer_visual(self, video_list):
        # Returns (img_list, something_else)
        img_tensor = torch.randn(2, 4)
        return img_tensor, None

class MockChat(nn.Module):
    def __init__(self):
        super().__init__()
        self.model = MockModel()
    
    def upload_video_without_audio(self, video_list, num_frames=8):
        # Returns (msg, img_list)
        msg = "Video uploaded"
        img_list = [torch.randn(2, 4), torch.randn(2, 4)]
        return msg, img_list
    
    def answer(self, img_list, input_text, options, msg, num_beams, temperature, max_new_tokens, max_length):
        # Returns a list with one element (the answer)
        return ["The answer is option A"]

chat = MockChat()
video_list = [torch.randn(3, 224, 224), torch.randn(3, 224, 224)]
question = "What is happening in the video?"
options = ["Option A", "Option B", "Option C"]
num_beams = 3
temperature = 0.7
num_frames = 8
task = "video_llama"

ARGS = {
    "chat": chat,
    "video_list": video_list,
    "question": question,
    "options": options,
    "num_beams": num_beams,
    "temperature": temperature,
    "num_frames": num_frames,
    "task": task
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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_video_fragments Not yet run
unverified (harvested, no recorded run)
deep-spin/infinite-video/infty-Video-LLaMA/eval_code/eval/run_inference_inf_video_llama_egochema.py
code served (permissive licence) · get_code("88959200d487da89")
How it ran

no invocation recorded.

Also present in other papers' code

Of 3 samples lifted from this paper's linked repositories, 3 carry a function identity and 0 do not. Of the 3 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 3 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.

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

Current video-language models struggle with long-video understanding due to limited context lengths and reliance on sparse frame subsampling, often leading to information loss. This paper introduces $\infty$-Video, which can process arbitrarily long videos through a continuous-time long-term memory (LTM) consolidation mechanism. Our framework augments video Q-formers by allowing them to process unbounded video contexts efficiently and without requiring additional training. Through continuous attention, our approach dynamically allocates higher granularity to the most relevant video segments, forming "sticky" memories that evolve over time. Experiments with Video-LLaMA and VideoChat2 demonstrate improved performance in video question-answering tasks, showcasing the potential of continuous-time LTM mechanisms to enable scalable and training-free comprehension of long videos.

For agents

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

get_harvested_code_for_paper("2501.19098")
get_code_for_paper("2501.19098")
have("2501.19098")

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

curl https://syntology.ai/api/ran/2501.19098.json

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

[![Syntology run record](https://syntology.ai/api/ran/2501.19098.svg)](https://syntology.ai/paper/2501.19098)

Connect an agent — have() is free.