Datasets:

Xin1818
/

AgriTaxon

Error code:   StreamingRowsError
Exception:    CastError
Message:      Couldn't cast
pair_id: string
track: string
qid: string
label: string
prediction: string
strata: list<item: string>
  child 0, item: string
items: list<item: struct<model: string, stratum: string>>
  child 0, item: struct<model: string, stratum: string>
      child 0, model: string
      child 1, stratum: string
human_label: int64
supplementary: struct<alias_validation_for_expert_review.xlsx: struct<records: int64, usable: int64, expert_alias:  (... 2064 chars omitted)
  child 0, alias_validation_for_expert_review.xlsx: struct<records: int64, usable: int64, expert_alias: int64, conflicts: null, qwen38v2: struct<TP: int (... 588 chars omitted)
      child 0, records: int64
      child 1, usable: int64
      child 2, expert_alias: int64
      child 3, conflicts: null
      child 4, qwen38v2: struct<TP: int64, FP: int64, FN: int64, TN: int64, precision: double, recall: double, accuracy: doub (... 43 chars omitted)
          child 0, TP: int64
          child 1, FP: int64
          child 2, FN: int64
          child 3, TN: int64
          child 4, precision: double
          child 5, recall: double
          child 6, accuracy: double
          child 7, kappa: double
          child 8, missing_verdicts: int64
      child 5, qwen38: struct<TP: int64, FP: int64, FN: int64, TN: int64, precision: double, recall: double, accuracy: doub (... 43 chars omitted)
          child 0, TP: int64
          child 1, FP: int64
          child 2, FN: int64
          child 3, TN: int64
          child 4, precisio
...
ld 0, item: double
              child 3, jeffreys_median: double
              child 4, nan_reps: int64
          child 8, FP: struct<point: double, ci95_bootstrap: list<item: double>, ci95_jeffreys: list<item: double>, jeffrey (... 34 chars omitted)
              child 0, point: double
              child 1, ci95_bootstrap: list<item: double>
                  child 0, item: double
              child 2, ci95_jeffreys: list<item: double>
                  child 0, item: double
              child 3, jeffreys_median: double
              child 4, nan_reps: int64
          child 9, FN: struct<point: double, ci95_bootstrap: list<item: double>, ci95_jeffreys: list<item: double>, jeffrey (... 34 chars omitted)
              child 0, point: double
              child 1, ci95_bootstrap: list<item: double>
                  child 0, item: double
              child 2, ci95_jeffreys: list<item: double>
                  child 0, item: double
              child 3, jeffreys_median: double
              child 4, nan_reps: int64
          child 10, TN: struct<point: double, ci95_bootstrap: list<item: double>, ci95_jeffreys: list<item: double>, jeffrey (... 34 chars omitted)
              child 0, point: double
              child 1, ci95_bootstrap: list<item: double>
                  child 0, item: double
              child 2, ci95_jeffreys: list<item: double>
                  child 0, item: double
              child 3, jeffreys_median: double
              child 4, nan_reps: int64
to
{'banner': Value('null'), 'judges': List(Value('string')), 'reference_judge': Value('string'), 'notes': List(Value('null')), 'llm_labels': {'pairs': Value('int64'), 'counts': {'1': Value('int64'), '0': Value('int64'), '?': Value('int64')}, 'errors': List(Value('null'))}, 'review': Value('null'), 'blind_subset': Value('null'), 'reference': {'llm (accepted by author)': Value('int64'), 'unresolved': Value('int64')}, 'adjudication_needed': List(Value('int64')), 'strata': List({'stratum': Value('string'), 'N_h': Value('int64'), 'n_labelled': Value('int64'), 'valid': Value('int64'), 'rate': Value('float64'), 'cp95': List(Value('float64'))}), 'estimates': {'qwen38v2': {'overall': {'precision': {'point': Value('float64'), 'ci95_bootstrap': List(Value('float64')), 'ci95_jeffreys': List(Value('float64')), 'jeffreys_median': Value('float64'), 'nan_reps': Value('int64')}, 'recall': {'point': Value('float64'), 'ci95_bootstrap': List(Value('float64')), 'ci95_jeffreys': List(Value('float64')), 'jeffreys_median': Value('float64'), 'nan_reps': Value('int64')}, 'accuracy': {'point': Value('float64'), 'ci95_bootstrap': List(Value('float64')), 'ci95_jeffreys': List(Value('float64')), 'jeffreys_median': Value('float64'), 'nan_reps': Value('int64')}, 'kappa': {'point': Value('float64'), 'ci95_bootstrap': List(Value('float64')), 'ci95_jeffreys': List(Value('float64')), 'jeffreys_median': Value('float64'), 'nan_reps': Value('int64')}, 'prevalence': {'point': Value('float64'), 'ci95_bootstrap': List(
...
 'accuracy': Value('float64'), 'kappa': Value('float64'), 'missing_verdicts': Value('int64')}, 'em_only': {'TP': Value('int64'), 'FP': Value('int64'), 'FN': Value('int64'), 'TN': Value('int64'), 'precision': Value('null'), 'recall': Value('float64'), 'accuracy': Value('float64'), 'kappa': Value('float64'), 'missing_verdicts': Value('int64')}}, 'both (deduplicated by qid)': {'records': Value('int64'), 'usable': Value('int64'), 'expert_alias': Value('int64'), 'conflicts': Value('int64'), 'qwen38v2': {'TP': Value('int64'), 'FP': Value('int64'), 'FN': Value('int64'), 'TN': Value('int64'), 'precision': Value('float64'), 'recall': Value('float64'), 'accuracy': Value('float64'), 'kappa': Value('float64'), 'missing_verdicts': Value('int64')}, 'qwen38': {'TP': Value('int64'), 'FP': Value('int64'), 'FN': Value('int64'), 'TN': Value('int64'), 'precision': Value('float64'), 'recall': Value('float64'), 'accuracy': Value('float64'), 'kappa': Value('float64'), 'missing_verdicts': Value('int64')}, 'legacy': {'TP': Value('int64'), 'FP': Value('int64'), 'FN': Value('int64'), 'TN': Value('int64'), 'precision': Value('float64'), 'recall': Value('float64'), 'accuracy': Value('float64'), 'kappa': Value('float64'), 'missing_verdicts': Value('int64')}, 'em_only': {'TP': Value('int64'), 'FP': Value('int64'), 'FN': Value('int64'), 'TN': Value('int64'), 'precision': Value('null'), 'recall': Value('float64'), 'accuracy': Value('float64'), 'kappa': Value('float64'), 'missing_verdicts': Value('int64')}}}}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/utils.py", line 147, in get_rows_or_raise
                  return get_rows(
                      dataset=dataset,
                  ...<4 lines>...
                      column_names=column_names,
                  )
                File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
                  return func(*args, **kwargs)
                File "/src/services/worker/src/worker/utils.py", line 127, in get_rows
                  rows_plus_one = list(itertools.islice(safe_iter(ds, dataset=dataset), rows_max_number + 1))
                File "/src/services/worker/src/worker/utils.py", line 483, in safe_iter
                  yield from ds.decode(False) if ds.features else ds
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2951, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2461, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2486, in _iter_arrow
                  for key, pa_table in self.ex_iterable._iter_arrow():
                                       ~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 547, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 430, in _iter_arrow
                  for key, pa_table in self.generate_tables_fn(**gen_kwags):
                                       ~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2378, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              pair_id: string
              track: string
              qid: string
              label: string
              prediction: string
              strata: list<item: string>
                child 0, item: string
              items: list<item: struct<model: string, stratum: string>>
                child 0, item: struct<model: string, stratum: string>
                    child 0, model: string
                    child 1, stratum: string
              human_label: int64
              supplementary: struct<alias_validation_for_expert_review.xlsx: struct<records: int64, usable: int64, expert_alias:  (... 2064 chars omitted)
                child 0, alias_validation_for_expert_review.xlsx: struct<records: int64, usable: int64, expert_alias: int64, conflicts: null, qwen38v2: struct<TP: int (... 588 chars omitted)
                    child 0, records: int64
                    child 1, usable: int64
                    child 2, expert_alias: int64
                    child 3, conflicts: null
                    child 4, qwen38v2: struct<TP: int64, FP: int64, FN: int64, TN: int64, precision: double, recall: double, accuracy: doub (... 43 chars omitted)
                        child 0, TP: int64
                        child 1, FP: int64
                        child 2, FN: int64
                        child 3, TN: int64
                        child 4, precision: double
                        child 5, recall: double
                        child 6, accuracy: double
                        child 7, kappa: double
                        child 8, missing_verdicts: int64
                    child 5, qwen38: struct<TP: int64, FP: int64, FN: int64, TN: int64, precision: double, recall: double, accuracy: doub (... 43 chars omitted)
                        child 0, TP: int64
                        child 1, FP: int64
                        child 2, FN: int64
                        child 3, TN: int64
                        child 4, precisio
              ...
              ld 0, item: double
                            child 3, jeffreys_median: double
                            child 4, nan_reps: int64
                        child 8, FP: struct<point: double, ci95_bootstrap: list<item: double>, ci95_jeffreys: list<item: double>, jeffrey (... 34 chars omitted)
                            child 0, point: double
                            child 1, ci95_bootstrap: list<item: double>
                                child 0, item: double
                            child 2, ci95_jeffreys: list<item: double>
                                child 0, item: double
                            child 3, jeffreys_median: double
                            child 4, nan_reps: int64
                        child 9, FN: struct<point: double, ci95_bootstrap: list<item: double>, ci95_jeffreys: list<item: double>, jeffrey (... 34 chars omitted)
                            child 0, point: double
                            child 1, ci95_bootstrap: list<item: double>
                                child 0, item: double
                            child 2, ci95_jeffreys: list<item: double>
                                child 0, item: double
                            child 3, jeffreys_median: double
                            child 4, nan_reps: int64
                        child 10, TN: struct<point: double, ci95_bootstrap: list<item: double>, ci95_jeffreys: list<item: double>, jeffrey (... 34 chars omitted)
                            child 0, point: double
                            child 1, ci95_bootstrap: list<item: double>
                                child 0, item: double
                            child 2, ci95_jeffreys: list<item: double>
                                child 0, item: double
                            child 3, jeffreys_median: double
                            child 4, nan_reps: int64
              to
              {'banner': Value('null'), 'judges': List(Value('string')), 'reference_judge': Value('string'), 'notes': List(Value('null')), 'llm_labels': {'pairs': Value('int64'), 'counts': {'1': Value('int64'), '0': Value('int64'), '?': Value('int64')}, 'errors': List(Value('null'))}, 'review': Value('null'), 'blind_subset': Value('null'), 'reference': {'llm (accepted by author)': Value('int64'), 'unresolved': Value('int64')}, 'adjudication_needed': List(Value('int64')), 'strata': List({'stratum': Value('string'), 'N_h': Value('int64'), 'n_labelled': Value('int64'), 'valid': Value('int64'), 'rate': Value('float64'), 'cp95': List(Value('float64'))}), 'estimates': {'qwen38v2': {'overall': {'precision': {'point': Value('float64'), 'ci95_bootstrap': List(Value('float64')), 'ci95_jeffreys': List(Value('float64')), 'jeffreys_median': Value('float64'), 'nan_reps': Value('int64')}, 'recall': {'point': Value('float64'), 'ci95_bootstrap': List(Value('float64')), 'ci95_jeffreys': List(Value('float64')), 'jeffreys_median': Value('float64'), 'nan_reps': Value('int64')}, 'accuracy': {'point': Value('float64'), 'ci95_bootstrap': List(Value('float64')), 'ci95_jeffreys': List(Value('float64')), 'jeffreys_median': Value('float64'), 'nan_reps': Value('int64')}, 'kappa': {'point': Value('float64'), 'ci95_bootstrap': List(Value('float64')), 'ci95_jeffreys': List(Value('float64')), 'jeffreys_median': Value('float64'), 'nan_reps': Value('int64')}, 'prevalence': {'point': Value('float64'), 'ci95_bootstrap': List(
              ...
               'accuracy': Value('float64'), 'kappa': Value('float64'), 'missing_verdicts': Value('int64')}, 'em_only': {'TP': Value('int64'), 'FP': Value('int64'), 'FN': Value('int64'), 'TN': Value('int64'), 'precision': Value('null'), 'recall': Value('float64'), 'accuracy': Value('float64'), 'kappa': Value('float64'), 'missing_verdicts': Value('int64')}}, 'both (deduplicated by qid)': {'records': Value('int64'), 'usable': Value('int64'), 'expert_alias': Value('int64'), 'conflicts': Value('int64'), 'qwen38v2': {'TP': Value('int64'), 'FP': Value('int64'), 'FN': Value('int64'), 'TN': Value('int64'), 'precision': Value('float64'), 'recall': Value('float64'), 'accuracy': Value('float64'), 'kappa': Value('float64'), 'missing_verdicts': Value('int64')}, 'qwen38': {'TP': Value('int64'), 'FP': Value('int64'), 'FN': Value('int64'), 'TN': Value('int64'), 'precision': Value('float64'), 'recall': Value('float64'), 'accuracy': Value('float64'), 'kappa': Value('float64'), 'missing_verdicts': Value('int64')}, 'legacy': {'TP': Value('int64'), 'FP': Value('int64'), 'FN': Value('int64'), 'TN': Value('int64'), 'precision': Value('float64'), 'recall': Value('float64'), 'accuracy': Value('float64'), 'kappa': Value('float64'), 'missing_verdicts': Value('int64')}, 'em_only': {'TP': Value('int64'), 'FP': Value('int64'), 'FN': Value('int64'), 'TN': Value('int64'), 'precision': Value('null'), 'recall': Value('float64'), 'accuracy': Value('float64'), 'kappa': Value('float64'), 'missing_verdicts': Value('int64')}}}}
              because column names don't match

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

AgriTaxon: Knowledge-grounded benchmarking of open-ended agricultural entity naming with multimodal foundation models

Xin Zeng, Benfeng Xu, Qian Chen, Qiang Wu, Wentao Zhang, Liguo Lang, Shancheng Fang, Huarui Wu

🌐 Project Page · 💻 GitHub

Overview

Multimodal foundation models are used as generative assistants: shown a photo of a crop, an animal, an insect, or a weed, they must answer with its name. AgriTaxon tests whether they can, in two settings on the same entities: multiple-choice selection (choose the label among four options) and open-ended naming (state the name without options). Models select far more names than they can state; AgriTaxon measures this seeing-without-naming gap.

Every entity is linked through Wikidata to an agricultural knowledge base (FAO Ecocrop, FAO DAD-IS, or the EPPO Global Database) and is labelled at species level or finer (for livestock, usually a breed).

Setting	Purpose	Entities	Images
AgriTaxon-Main	Default setting	6,049	6,049 Wiki images
AgriTaxon-Hard	Extended: entities that nearly all models fail	510	510 Wiki images
AgriTaxon-Wild	Extended: field images of the same entities	2,826	14,130 field images

Track	Entities	Knowledge base
Crop	1,960	FAO Ecocrop (Wikidata P4753)
Livestock	172	FAO DAD-IS (Wikidata P3380)
Pest	2,415	EPPO Global Database (Wikidata P3031)
Weed	1,502	EPPO Global Database (Wikidata P3031)
Total	6,049

Dataset Structure

├── annotations/{track}.jsonl          # AgriTaxon, open-ended naming
├── annotations/{track}_mc.jsonl       # AgriTaxon, multiple choice: options, answer, and `hard` (AgriTaxon-Hard)
├── images/{track}/                    # Wiki images, one per entity
├── splits/hard.json                   # AgriTaxon-Hard QIDs (510)
├── wild/                              # AgriTaxon-Wild: annotations, field images, per-photo attributions and licenses
├── coarse_subset/                     # 1,365 genus- and family-level entities, excluded from AgriTaxon (kept for reference)
├── construction/                      # the five construction steps, with removed ids and Wikidata ranks
├── judge/                             # LLM-judge prompt and its expert-annotated validation sample
├── outputs/                           # model outputs with scores
│   ├── main/<model>/{track}_{multiple-choice,open-ended}.jsonl      # 17 models
│   └── wild/k{1,3,5}/<model>/{track}_{multiple-choice,open-ended}.jsonl
└── metadata.json

Tracks are crop, livestock, pest, and weed. Annotation entries carry qid, label, image, track, the knowledge-base identifier (ecocropID, faoID, or eppoCode), and Wikipedia links when available; _mc files add options and answer. Output rows carry qid, subset (main or coarse; Wild rows omit it), track, setting, label, prediction, raw_response, and the score: correct for multiple choice; exact_match, judge_accepts, judge_reason, and correct for open-ended naming.

Every entity is labelled at species level or finer. Entities whose knowledge-base label is a genus or family are excluded and kept apart in coarse_subset/, because a photo shows one species and a genus label would score a correct species-level answer as wrong. construction/README.md documents how the 6,049 entities were selected.

Evaluation

MC accuracy: share of multiple-choice items answered with the correct option.
OE-EM: the prediction matches the label after normalization; a name more specific than the label (e.g. a subspecies of the labelled species) also counts.
OE-Acc: additionally credits aliases (common names, taxonomic synonyms, spelling variants) accepted by an LLM judge: the open-weight Qwen3.8-Flash-Next (FP8 checkpoint, reasoning disabled, temperature 0; prompt in judge/judge_prompt.md). On a stratified sample of 579 predictions, the judge agrees with expert annotation on 97.8% (OE-EM: 92.8%).

Summary scores are the unweighted mean over the four tracks.

Main Results

Accuracy (%), ranked by OE-Acc within each group. Hard is MC accuracy on AgriTaxon-Hard, which was selected against the first 14 models; Qwen3.8-Flash-Next, DeepSeek-V4.1-Flash and GPT-6 Luna were evaluated later (–).

Model	MC accuracy	OE-EM	OE-Acc	Hard
Proprietary models
gemini-3-pro-preview	85.8	38.7	47.3	13.7
doubao-seed-2-0-pro	82.6	36.8	45.8	10.8
doubao-seed-2-0-lite	80.8	32.4	40.5	10.8
gemini-3-flash-preview	85.8	25.0	39.5	11.4
gpt-5	81.4	23.0	31.6	12.9
gpt-5-mini	73.2	16.8	22.2	4.7
gpt-6-luna	73.1	6.9	12.9	–
claude-haiku-4-5	63.1	5.5	9.2	7.6
Open-source models
kimi-k2.5	76.6	23.6	32.0	3.7
qwen3.8-flash-next	76.4	18.9	25.8	–
deepseek-v4.1-flash	78.6	18.6	23.3	–
qwen3.5-397b-a17b	74.0	16.3	21.3	14.1
qwen3-vl-235b-a22b	70.7	15.1	20.2	3.5
glm-4.6v	65.9	12.3	17.1	2.2
glm-4.6v-flashx	61.7	8.6	12.5	5.9
qwen3-vl-30b-a3b	62.3	7.0	11.7	3.1
qwen3.5-35b-a3b	70.0	5.2	10.6	7.8

AgriTaxon-Wild Results

Model	Images	MC accuracy	OE-EM	OE-Acc
doubao-seed-2-0-lite	Wiki image	81.0	37.9	48.4
	K=1	80.6	31.0	40.3
	K=3	88.1	47.0	56.3
	K=5	89.4	49.7	61.1
qwen3.5-35b-a3b	Wiki image	75.5	7.1	23.1
	K=1	66.2	7.2	15.2
	K=3	75.5	16.0	25.6
	K=5	76.7	18.5	28.8

One field image is less informative than the Wiki image; several field images overtake it.

Usage

pip install huggingface_hub
huggingface-cli download Xin1818/AgriTaxon --repo-type dataset --local-dir AgriTaxon

import json
crop = [json.loads(l) for l in open("AgriTaxon/annotations/crop.jsonl")]
hard = set(json.load(open("AgriTaxon/splits/hard.json"))["qids"])
print(len(crop), "crop entities;", len(hard), "AgriTaxon-Hard entities")

Licensing and Attribution

Benchmark annotations and metadata: CC BY 4.0.
Wiki images (images/): Wikimedia Commons images under their original licenses, predominantly CC BY and CC BY-SA.
Wild images (wild/images/): iNaturalist photographs under their individual licenses, predominantly CC BY-NC. Per-photo licenses and credits are provided in wild/attributions.jsonl and summarized in wild/LICENSES.md.
Source code, prompts, and project documentation: MIT License.
Knowledge-base identifiers: Wikidata QIDs are available under CC0; FAO and EPPO identifiers are used for reference linking only.

Wild license distribution:

License	Photos	License	Photos
CC BY-NC	11,837	CC BY-NC-SA	196
CC BY	1,391	CC BY-NC-ND	163
CC0	360	CC BY-SA	143
CC BY-ND	40

Because AgriTaxon-Wild includes NonCommercial and NoDerivatives images, the repository metadata uses a mixed license. Users must consult the per-photo license and attribution record before reuse. In particular, the Wild subset should be treated as non-commercial, and the 203 NoDerivatives photographs must not be modified.

Downloads last month: 521

Total file size:

18.9 GB