References:
Use the following command to load this dataset in TFDS:
ds = tfds.load('huggingface:code_x_glue_tc_text_to_code')
- Description:
CodeXGLUE text-to-code dataset, available at https://github.com/microsoft/CodeXGLUE/tree/main/Text-Code/text-to-code
We use concode dataset which is a widely used code generation dataset from Iyer's EMNLP 2018 paper Mapping Language to Code in Programmatic Context. See paper for details.
- License: No known license
- Version: 0.0.0
- Splits:
Split | Examples |
---|---|
'test' |
2000 |
'train' |
100000 |
'validation' |
2000 |
- Features:
{
"id": {
"dtype": "int32",
"id": null,
"_type": "Value"
},
"nl": {
"dtype": "string",
"id": null,
"_type": "Value"
},
"code": {
"dtype": "string",
"id": null,
"_type": "Value"
}
}