Dataset Viewer
Auto-converted to Parquet Duplicate
record_id
stringlengths
36
36
concentration_ppm
float64
100
50k
molecule
stringclasses
1 value
temperature_K
float64
291
302
pressure_atm
float64
0.95
1.07
path_length_m
float64
10
10
instrument_config_id
stringclasses
6 values
technique
stringclasses
2 values
raw_scan
listlengths
2k
2k
absorbance_clean
listlengths
2k
2k
00015808-fc65-447e-a883-ab3e6b3f07b2
6,563.855287
CO2
296.069147
0.998276
10
vi-co2-da-medium-18
TDLAS-DA
[ 0.9207104925837759, 0.9206189342611243, 0.9207104925837759, 0.9210767258743819, 0.9209851675517304, 0.9209851675517304, 0.9209851675517304, 0.9210767258743819, 0.9209851675517304, 0.9212598425196851, 0.9214429591649881, 0.9213514008423366, 0.9213514008423366, 0.9216260758102911, 0.921626...
[ 0.00028852656832499573, 0.00028990920732661, 0.0002913020999536553, 0.00029270534799011604, 0.0002941190544775005, 0.00029554332372794235, 0.00029697826134816884, 0.0002984239742558666, 0.000299880570697618, 0.00030134816026986276, 0.0003028268539374726, 0.0003043167640610106, 0.000305818004...
0002d12f-88a8-4284-b982-5844450d9f42
4,495.89305
CO2
295.582348
0.995829
10
vi-co2-da-medium-18
TDLAS-DA
[ 0.8403222852957334, 0.8405969602636879, 0.8408716352316425, 0.840780076908991, 0.8414209851675517, 0.8414209851675517, 0.8415125434902032, 0.8416041018128548, 0.8418787767808094, 0.8424281267167185, 0.8422450100714155, 0.8424281267167185, 0.84251968503937, 0.8424281267167185, 0.843160593...
[ 0.00022880316779755815, 0.0002299728232642478, 0.00023115165963993886, 0.00023233977293609824, 0.00023353726040962867, 0.0002347442205750579, 0.00023596075322721949, 0.0002371869594656089, 0.00023842294170273924, 0.00023966880369734328, 0.00024092465056570382, 0.00024219058880987053, 0.00024...
0008c8ae-d714-4dc0-8906-61e12f496135
314.395359
CO2
296.00698
1.0001
10
vi-co2-da-easy-17
TDLAS-DA
[0.8812085145342183,0.8812314030670634,0.8813916227969787,0.881528953994049,0.8816433966582742,0.881(...TRUNCATED)
[0.000011044243745761154,0.000011093543865734809,0.000011143185831357909,0.000011193172830688284,0.0(...TRUNCATED)
000c96e9-de3a-469d-9081-1ec77e8cc0c9
9,501.663713
CO2
296.302464
1.000973
10
vi-co2-da-medium-18
TDLAS-DA
[0.8939754623695294,0.8937923457242263,0.8942501373374839,0.8941585790148324,0.8943416956601355,0.89(...TRUNCATED)
[0.0007547866435894429,0.0007592022920231852,0.0007636571216019573,0.0007681515891536836,0.000772686(...TRUNCATED)
0036a490-43b7-41cc-ba02-390b003a3559
4,044.88283
CO2
296.000624
0.999307
10
vi-co2-da-easy-17
TDLAS-DA
[0.9129091325246051,0.9129549095902952,0.9132524605172808,0.9131837949187457,0.913344014648661,0.913(...TRUNCATED)
[0.00011565250552096021,0.00011613426545080969,0.00011661915649185125,0.00011710720612745251,0.00011(...TRUNCATED)
0049b6ec-40bb-4088-aa91-4f760d972ae8
8,092.803067
CO2
296.491708
1.009259
10
vi-co2-da-hard-19
TDLAS-DA
[0.8435897435897436,0.8435897435897436,0.8435897435897436,0.8435897435897436,0.8443223443223443,0.84(...TRUNCATED)
[0.0002541537068315472,0.00025522194452686316,0.0002562972096605389,0.0002573795646284793,0.00025846(...TRUNCATED)
00525de9-e7a2-4eac-97ad-c8c2c6589148
3,350.799058
CO2
295.994176
1.000142
10
vi-co2-da-easy-17
TDLAS-DA
[0.89482719157702,0.8949645227740902,0.8950102998397803,0.8951018539711605,0.8952162966353857,0.8953(...TRUNCATED)
[0.0002211283669645944,0.00022234801165435997,0.00022357792888551967,0.00022481823311631748,0.000226(...TRUNCATED)
005cbb11-87ac-4581-9c26-3063f75c41ff
24,467.821699
CO2
296.342479
0.99738
10
vi-co2-da-medium-18
TDLAS-DA
[0.8732832814502838,0.8732832814502838,0.8733748397729354,0.8735579564182384,0.8734663980955869,0.87(...TRUNCATED)
[0.001932880090039782,0.0019440513794807908,0.0019553209665145906,0.00196668998856925,0.001978159599(...TRUNCATED)
006ee9ce-366d-4872-a44d-9a9d1e613543
1,154.699752
CO2
296.028615
0.999932
10
vi-co2-da-medium-18
TDLAS-DA
[0.8945248123054386,0.8947994872733931,0.8949826039186962,0.8948910455960447,0.8951657205639992,0.89(...TRUNCATED)
[0.00002727081100453919,0.000027378825118538335,0.000027487505437677365,0.000027596857526208748,0.00(...TRUNCATED)
006f28f3-7dc4-4db3-830a-55047e6502fc
2,930.588735
CO2
296.002511
0.999989
10
vi-co2-da-easy-17
TDLAS-DA
[0.8910734721904326,0.8911192492561227,0.891256580453193,0.8913939116502633,0.8914625772487984,0.891(...TRUNCATED)
[0.0001338500162241488,0.00013450246015937124,0.00013515981364277459,0.00013582212606443942,0.000136(...TRUNCATED)
End of preview. Expand in Data Studio

SPEKTRAN CO2 v0 — AI Agent-Ready Benchmark Data

AI Agent Ready — load this dataset and run the full ML pipeline via natural language. Install spektran, then tell your AI coding agent: "Train a ridge baseline on CO2 data and show me the scores."

CO2 concentration regression benchmark from SPEKTRAN, the open-source TDLAS/NDIR simulation engine for optical gas sensing.

Why CO2?

CO2 is the most widely studied greenhouse gas in spectroscopy. Every lab doing laser-based gas sensing has CO2 cells and references. This dataset provides a standardized ML benchmark for CO2 concentration retrieval — the same task performed thousands of times daily in environmental monitoring, indoor air quality, and industrial process control.

Configs

Config Molecule Technique Task Splits
da CO2 TDLAS-DA T1 concentration + T3 cross-instrument train / val / test / test_heldout
wms CO2 TDLAS-WMS T4 WMS concentration train / val / test

Quick Start

from datasets import load_dataset

# DA concentration regression (T1/T3)
ds = load_dataset("spektran/spektran-co2-v0", "da")

# WMS concentration (T4)
ds_wms = load_dataset("spektran/spektran-co2-v0", "wms")

AI Agent Workflow

pip install spektran
spektran info --json                       # Project discovery
spektran list tasks --json                 # Available benchmark tasks
spektran train --baseline ridge --json     # Train baseline (auto-generates data)

Dataset Details

  • Molecule: CO2 (carbon dioxide)
  • Spectral band: 2ν1+ν3 combination band near 4978 cm⁻¹
  • Concentration range: 100–50,000 ppm (log-uniform)
  • Instruments: 4 virtual instruments (easy/medium/hard/held-out)
  • Physics: HITRAN demo lines, Voigt line profile, Beer-Lambert
  • Noise chain: scan nonlinearity, etalon fringes, baseline drift, white + 1/f noise, ADC quantization

Citation

@software{spektran,
  title   = {SPEKTRAN: Simulation Engine and ML Benchmark for Optical Gas Sensing},
  url     = {https://github.com/spektran/spektran},
  doi     = {10.5281/zenodo.21790394},
  version = {0.5.1},
  license = {Apache-2.0}
}

Links

Downloads last month
-