---
license: other
task_categories:
- text-retrieval
tags:
- mteb
- code
- python
configs:
- config_name: AppsRetrieval-corpus
  data_files:
  - split: train
    path: AppsRetrieval/train/corpus.parquet
- config_name: AppsRetrieval-queries
  data_files:
  - split: train
    path: AppsRetrieval/train/queries.parquet
- config_name: AppsRetrieval-qrels
  data_files:
  - split: train
    path: AppsRetrieval/train/qrels.parquet
- config_name: COIRCodeSearchNetRetrieval-corpus
  data_files:
  - split: train
    path: COIRCodeSearchNetRetrieval/train/corpus.parquet
  - split: valid
    path: COIRCodeSearchNetRetrieval/valid/corpus.parquet
- config_name: COIRCodeSearchNetRetrieval-queries
  data_files:
  - split: train
    path: COIRCodeSearchNetRetrieval/train/queries.parquet
  - split: valid
    path: COIRCodeSearchNetRetrieval/valid/queries.parquet
- config_name: COIRCodeSearchNetRetrieval-qrels
  data_files:
  - split: train
    path: COIRCodeSearchNetRetrieval/train/qrels.parquet
  - split: valid
    path: COIRCodeSearchNetRetrieval/valid/qrels.parquet
- config_name: CodeSearchNetCCRetrieval-corpus
  data_files:
  - split: train
    path: CodeSearchNetCCRetrieval/train/corpus.parquet
  - split: valid
    path: CodeSearchNetCCRetrieval/valid/corpus.parquet
- config_name: CodeSearchNetCCRetrieval-queries
  data_files:
  - split: train
    path: CodeSearchNetCCRetrieval/train/queries.parquet
  - split: valid
    path: CodeSearchNetCCRetrieval/valid/queries.parquet
- config_name: CodeSearchNetCCRetrieval-qrels
  data_files:
  - split: train
    path: CodeSearchNetCCRetrieval/train/qrels.parquet
  - split: valid
    path: CodeSearchNetCCRetrieval/valid/qrels.parquet
- config_name: CodeTransOceanContest-corpus
  data_files:
  - split: train
    path: CodeTransOceanContest/train/corpus.parquet
  - split: valid
    path: CodeTransOceanContest/valid/corpus.parquet
- config_name: CodeTransOceanContest-queries
  data_files:
  - split: train
    path: CodeTransOceanContest/train/queries.parquet
  - split: valid
    path: CodeTransOceanContest/valid/queries.parquet
- config_name: CodeTransOceanContest-qrels
  data_files:
  - split: train
    path: CodeTransOceanContest/train/qrels.parquet
  - split: valid
    path: CodeTransOceanContest/valid/qrels.parquet
- config_name: CodeTransOceanDL-corpus
  data_files:
  - split: train
    path: CodeTransOceanDL/train/corpus.parquet
  - split: valid
    path: CodeTransOceanDL/valid/corpus.parquet
- config_name: CodeTransOceanDL-queries
  data_files:
  - split: train
    path: CodeTransOceanDL/train/queries.parquet
  - split: valid
    path: CodeTransOceanDL/valid/queries.parquet
- config_name: CodeTransOceanDL-qrels
  data_files:
  - split: train
    path: CodeTransOceanDL/train/qrels.parquet
  - split: valid
    path: CodeTransOceanDL/valid/qrels.parquet
- config_name: CosQA-corpus
  data_files:
  - split: train
    path: CosQA/train/corpus.parquet
  - split: valid
    path: CosQA/valid/corpus.parquet
- config_name: CosQA-queries
  data_files:
  - split: train
    path: CosQA/train/queries.parquet
  - split: valid
    path: CosQA/valid/queries.parquet
- config_name: CosQA-qrels
  data_files:
  - split: train
    path: CosQA/train/qrels.parquet
  - split: valid
    path: CosQA/valid/qrels.parquet
---

Train/valid splits that accompany the Python subsets of MTEB(Code, v1), decontaminated against the eval sets in [TeraflopAI/mteb-code-python](https://huggingface.co/datasets/TeraflopAI/mteb-code-python): documents whose text matches a test answer (after whitespace normalization) are removed from the corpora, along with the train/valid pairs that point at them.

Configs per task: `<task>-corpus`, `<task>-queries`, `<task>-qrels` (MTEB retrieval format).

| Task | MTEB split | Source | Revision | License |
|---|---|---|---|---|
| AppsRetrieval | train | [CoIR-Retrieval/apps](https://huggingface.co/datasets/CoIR-Retrieval/apps) | `f22508f` | mit |
| COIRCodeSearchNetRetrieval | train, valid | [CoIR-Retrieval/CodeSearchNet](https://huggingface.co/datasets/CoIR-Retrieval/CodeSearchNet) | `4adc7bc` | mit |
| CodeSearchNetCCRetrieval | train, valid | [CoIR-Retrieval/CodeSearchNet-ccr](https://huggingface.co/datasets/CoIR-Retrieval/CodeSearchNet-ccr) | `6e1effa` | mit |
| CodeTransOceanContest | train, valid | [CoIR-Retrieval/codetrans-contest](https://huggingface.co/datasets/CoIR-Retrieval/codetrans-contest) | `20da4eb` | apache-2.0 |
| CodeTransOceanDL | train, valid | [CoIR-Retrieval/codetrans-dl](https://huggingface.co/datasets/CoIR-Retrieval/codetrans-dl) | `281562c` | apache-2.0 |
| CosQA | train, valid | [CoIR-Retrieval/cosqa](https://huggingface.co/datasets/CoIR-Retrieval/cosqa) | `bc5efb7` | mit |