SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2001.01215 · 2020

A System for Real-Time Interactive Analysis of Deep Learning Training

arXiv · PDF · Open in the Atlas

Code that ran

9 of 17 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; 0 honoured the contract we drafted, 0 violated it and 9 have no contract verdict; 0 failures of our instrument; 8 no recorded run), measured 2026-11-21; 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/2001.01215.json · badge: /api/ran/2001.01215.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
microsoft/tensorwatch canonical (classifier label)
classifier basis: census
9 ran · 8 no recorded run
FunctionStatusWhere it lives
group_reduce Ran
ran on a synthesized input
microsoft/tensorwatch/tensorwatch/evaler_utils.py
code served (permissive licence) · get_code("6c7e3aab6467ff9c")
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 generator. 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 numpy as np
from typing import List, Any

torch.manual_seed(0)

# Create a simple input list
input_list = [
    {"id": 1, "value": 10},
    {"id": 2, "value": 20},
    {"id": 1, "value": 15},
    {"id": 2, "value": 25},
]

# Define key function - extracts the 'id' field
def key_func(item):
    return item["id"]

# Define value function - extracts the 'value' field
def val_func(item):
    return item["value"]

# Define reducer function - sums all values in a group
def reducer_func(values):
    return sum(list(values))

# Construct ARGS dictionary with all parameters for group_reduce
ARGS = {
    "l": input_list,
    "key_f": key_func,
    "val_f": val_func,
    "reducer": reducer_func,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type generator — a generator output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a generator 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
guess_image_dims Ran
ran on a synthesized input
microsoft/tensorwatch/tensorwatch/image_utils.py
code served (permissive licence) · get_code("05fe7faa77d1a5d2")
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 tuple, ndim 1, dtype family int; 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 math

torch.manual_seed(0)

# Test case 1: 1D tensor representing grayscale image (perfect square)
img_1d_grayscale = torch.randn(25)  # 5x5 grayscale

# Test case 2: 1D tensor representing RGB image (perfect square times 3)
img_1d_rgb = torch.randn(75)  # 5x5 RGB (25*3)

# Test case 3: Already 3D tensor (CHW format)
img_3d = torch.randn(3, 4, 5)  # 3 channels, 4 height, 5 width

ARGS = {
    "img": img_1d_rgb
}
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 tuple, ndim 1, dtype family int — a tuple output has no array shape of its own; none is stated
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
int2tensor Ran
ran on a synthesized input
microsoft/tensorwatch/tensorwatch/pytorch_utils.py
code served (permissive licence) · get_code("b6ef5930d34ebb10")
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 Tensor, ndim 1, dtype family int, shape (1,) 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

torch.manual_seed(0)

def int2tensor(val):
    return torch.LongTensor([val])

ARGS = {
    "val": 42
}
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 Tensor, ndim 1, dtype family int, shape (1,) — shape (1,) 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
pyt_ds2list Ran
ran on a synthesized input
microsoft/tensorwatch/tensorwatch/data_utils.py
code served (permissive licence) · get_code("bc61c240e64baea0")
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 list, ndim 2, 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

torch.manual_seed(0)

# Create a simple dataset-like object that supports indexing and len()
class SimpleDataset:
    def __init__(self):
        self.data = [
            (torch.randn(4), torch.randn(3)),
            (torch.randn(4), torch.randn(3)),
            (torch.randn(4), torch.randn(3)),
            (torch.randn(4), torch.randn(3)),
        ]
    
    def __len__(self):
        return len(self.data)
    
    def __iter__(self):
        return iter(self.data)
    
    def __getitem__(self, idx):
        return self.data[idx]

pyt_ds = SimpleDataset()

ARGS = {
    "pyt_ds": pyt_ds,
    "count": 2
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 2, dtype family float — a list output has no array shape of its own; none is stated
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
pyt_tensor2np Ran
ran on a synthesized input
microsoft/tensorwatch/tensorwatch/data_utils.py
code served (permissive licence) · get_code("4b2ad2d9ab9bb260")
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 (2, 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
import torch

torch.manual_seed(0)

# Test case 1: PyTorch tensor (2D)
tensor_2d = torch.randn(2, 4)

# Test case 2: PyTorch tensor (1D)
tensor_1d = torch.randn(1)

# Test case 3: None
none_val = None

# Test case 4: numpy array
numpy_array = np.array([[1.0, 2.0, 3.0], [4.0, 5.0, 6.0]])

# Test case 5: Python list
python_list = [1, 2, 3, 4]

# We'll use the 2D tensor as the primary test case
ARGS = {
    "pyt_tensor": tensor_2d
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — 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 (2, 4) — shape (2, 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
pyt_tuple2np Ran
ran on a synthesized input
microsoft/tensorwatch/tensorwatch/data_utils.py
code served (permissive licence) · get_code("758dbeb2636ae3c3")
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 tuple. 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 numpy as np

torch.manual_seed(0)

# Test case 1: PyTorch tensor (2D)
tensor_2d = torch.randn(2, 4)

# Test case 2: PyTorch tensor (1D)
tensor_1d = torch.randn(1, 3)

# Test case 3: NumPy array
np_array = np.array([1.0, 2.0, 3.0])

# Test case 4: None
none_value = None

# Test case 5: List
list_value = [1, 2, 3]

# Create a tuple with mixed types
pyt_tuple = (tensor_2d, tensor_1d, np_array, none_value, list_value)

ARGS = {
    "pyt_tuple": pyt_tuple
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is 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
stitch_horizontal Ran
ran on a synthesized input
microsoft/tensorwatch/tensorwatch/image_utils.py
code served (permissive licence) · get_code("f9c6d1845ac5da15")
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 3, dtype family float, shape (2, 15, 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

# Create sample images as numpy arrays with distinct dimensions
# Image 1: shape (2, 3, 4) - batch=2, height=3, width=4
image1 = np.random.rand(2, 3, 4)

# Image 2: shape (2, 5, 4) - batch=2, height=5, width=4
image2 = np.random.rand(2, 5, 4)

# Image 3: shape (2, 7, 4) - batch=2, height=7, width=4
image3 = np.random.rand(2, 7, 4)

images = [image1, image2, image3]

ARGS = {
    "images": images,
    "width_dim": 1
}
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 3, dtype family float, shape (2, 15, 4) — shape (2, 15, 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
to_imshow_array Ran
ran on a synthesized input
microsoft/tensorwatch/tensorwatch/image_utils.py
code served (permissive licence) · get_code("d06a7b46db306bc6")
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
from PIL import Image

# Test case 1: 1D array that can be reshaped to grayscale square
img_1d_gray = np.random.rand(16)  # 4x4 grayscale

# Test case 2: 1D array that can be reshaped to RGB
img_1d_rgb = np.random.rand(48)  # 4x4x3 RGB

# Test case 3: 2D array
img_2d = np.random.rand(8, 6)

# Test case 4: 3D array with 1 channel
img_3d_single = np.random.rand(1, 5, 4)

# Test case 5: 3D array with multiple channels (RGB)
img_3d_multi = np.random.rand(3, 5, 4)

# Test case 6: 4D array (batch of images)
img_4d = np.random.rand(2, 3, 5, 4)

# Test case 7: PIL Image
pil_img = Image.new('RGB', (4, 5), color='red')

ARGS = {
    "img": img_1d_gray,
    "width": None,
    "height": None
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, 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
to_tuples Ran
ran on a synthesized input
microsoft/tensorwatch/tensorwatch/evaler_utils.py
code served (permissive licence) · get_code("75217e6b03b24101")
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 generator. 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
torch.manual_seed(0)

ARGS = {
    "l": [1, 2, 3, 4, 5],
    "key_f": lambda x: x * 2,
    "val_f": lambda x: x ** 2,
}
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 generator — a generator output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a generator 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_model Not yet run
unverified (harvested, no recorded run)
microsoft/tensorwatch/tensorwatch/pytorch_utils.py
code served (permissive licence) · get_code("b89d16adfe8b4e3d")
How it ran

no invocation recorded.

image2batch Not yet run
unverified (harvested, no recorded run)
microsoft/tensorwatch/tensorwatch/imagenet_utils.py
code served (permissive licence) · get_code("fd376daff85e13cc")
How it ran

no invocation recorded.

model_stats Not yet run
unverified (harvested, no recorded run)
microsoft/tensorwatch/tensorwatch/model_graph/torchstat_utils.py
code served (permissive licence) · get_code("7d4bdc647c5d26ce")
How it ran

no invocation recorded.

model_stats2df Not yet run
unverified (harvested, no recorded run)
microsoft/tensorwatch/tensorwatch/model_graph/torchstat_utils.py
code served (permissive licence) · get_code("5171f2015e277d67")
How it ran

no invocation recorded.

predict Not yet run
unverified (harvested, no recorded run)
microsoft/tensorwatch/tensorwatch/imagenet_utils.py
code served (permissive licence) · get_code("d72c228526cca499")
How it ran

no invocation recorded.

probabilities2classes Not yet run
unverified (harvested, no recorded run)
microsoft/tensorwatch/tensorwatch/imagenet_utils.py
code served (permissive licence) · get_code("b1b7bb097a1cc013")
How it ran

no invocation recorded.

tensors2batch Not yet run
unverified (harvested, no recorded run)
microsoft/tensorwatch/tensorwatch/pytorch_utils.py
code served (permissive licence) · get_code("c61a459f1a7e6024")
How it ran

no invocation recorded.

topk Not yet run
unverified (harvested, no recorded run)
microsoft/tensorwatch/tensorwatch/evaler_utils.py
code served (permissive licence) · get_code("ff43ee11f61dfb71")
How it ran

no invocation recorded.

Also present in other papers' code

Of 17 samples lifted from this paper's linked repositories, 17 carry a function identity and 0 do not. Of the 17 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 17 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 2026-11-21.

Repositories linked to this paper

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

Abstract

Performing diagnosis or exploratory analysis during the training of deep learning models is challenging but often necessary for making a sequence of decisions guided by the incremental observations. Currently available systems for this purpose are limited to monitoring only the logged data that must be specified before the training process starts. Each time a new information is desired, a cycle of stop-change-restart is required in the training process. These limitations make interactive exploration and diagnosis tasks difficult, imposing long tedious iterations during the model development. We present a new system that enables users to perform interactive queries on live processes generating real-time information that can be rendered in multiple formats on multiple surfaces in the form of several desired visualizations simultaneously. To achieve this, we model various exploratory inspection and diagnostic tasks for deep learning training processes as specifications for streams using a map-reduce paradigm with which many data scientists are already familiar. Our design achieves generality and extensibility by defining composable primitives which is a fundamentally different approach than is used by currently available systems. The open source implementation of our system is available as TensorWatch project at https://github.com/microsoft/tensorwatch.

For agents

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

get_harvested_code_for_paper("2001.01215")
get_code_for_paper("2001.01215")
have("2001.01215")

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

curl https://syntology.ai/api/ran/2001.01215.json

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

[![Syntology run record](https://syntology.ai/api/ran/2001.01215.svg)](https://syntology.ai/paper/2001.01215)

Connect an agent — have() is free.