metadata
license: cc-by-sa-4.0
tags:
- code
extra_gated_prompt: >-
You agree to not use the model to conduct experiments that cause harm to human
subjects or generate malicious code.
extra_gated_fields:
Company: text
Country: country
Specific date: date_picker
I want to use this model for:
type: select
options:
- Research
- Education
- label: Other
value: other
I agree to use this model for non-commercial use ONLY: checkbox
task_categories:
- text-generation
size_categories:
- 1M<n<10M
dataset_info:
- config_name: C
features:
- name: Source_Code
dtype: string
- name: IR_Original
dtype: string
splits:
- name: Perf_Optimized
num_bytes: 3383884149
num_examples: 341419
- name: Size_Optimized
num_bytes: 2528286566
num_examples: 341785
download_size: 1323447636
dataset_size: 5912170715
- config_name: C++
features:
- name: Source_Code
dtype: string
- name: IR_Original
dtype: string
splits:
- name: Perf_Optimized
num_bytes: 116351369851
num_examples: 2898509
- name: Size_Optimized
num_bytes: 92572469724
num_examples: 2916655
download_size: 51690627847
dataset_size: 208923839575
- config_name: D
features:
- name: Source_Code
dtype: string
- name: IR_Original
dtype: string
splits:
- name: Perf_Optimized
num_bytes: 2320830137
num_examples: 7000
- name: Size_Optimized
num_bytes: 3271276765
num_examples: 11054
download_size: 1316382832
dataset_size: 5592106902
- config_name: Fortran
features:
- name: Source_Code
dtype: string
- name: IR_Original
dtype: string
splits:
- name: Perf_Optimized
num_bytes: 357741835
num_examples: 6327
- name: Size_Optimized
num_bytes: 2320830137
num_examples: 7000
download_size: 563853972
dataset_size: 2678571972
- config_name: Go
features:
- name: Source_Code
dtype: string
- name: IR_Original
dtype: string
splits:
- name: Perf_Optimized
num_bytes: 819560767
num_examples: 3913
- name: Size_Optimized
num_bytes: 741733997
num_examples: 3925
download_size: 317182680
dataset_size: 1561294764
- config_name: Haskell
features:
- name: Source_Code
dtype: string
- name: IR_Original
dtype: string
splits:
- name: Perf_Optimized
num_bytes: 3838556743
num_examples: 27892
- name: Size_Optimized
num_bytes: 3667186152
num_examples: 28203
download_size: 1736729352
dataset_size: 7505742895
- config_name: Nim
features:
- name: Source_Code
dtype: string
- name: IR_Original
dtype: string
splits:
- name: Size_Optimized
num_bytes: 106424381
num_examples: 215
download_size: 22506456
dataset_size: 106424381
- config_name: Objective-C
features:
- name: Source_Code
dtype: string
- name: IR_Original
dtype: string
splits:
- name: Perf_Optimized
num_bytes: 1729045
num_examples: 283
- name: Size_Optimized
num_bytes: 1433377
num_examples: 283
download_size: 707508
dataset_size: 3162422
- config_name: Python
features:
- name: Source_Code
dtype: string
- name: IR_Original
dtype: string
splits:
- name: Perf_Optimized
num_bytes: 13118428652
num_examples: 154507
- name: Size_Optimized
num_bytes: 13118428652
num_examples: 154507
download_size: 6511950536
dataset_size: 26236857304
- config_name: Rust
features:
- name: Source_Code
dtype: string
- name: IR_Original
dtype: string
splits:
- name: Perf_Optimized
num_bytes: 5859467468
num_examples: 38323
- name: Size_Optimized
num_bytes: 8695405064
num_examples: 32720
download_size: 5326634011
dataset_size: 14554872532
- config_name: Swift
features:
- name: Source_Code
dtype: string
- name: IR_Original
dtype: string
splits:
- name: Perf_Optimized
num_bytes: 260013963
num_examples: 2003
- name: Size_Optimized
num_bytes: 266356839
num_examples: 2015
download_size: 144113584
dataset_size: 526370802
configs:
- config_name: C
data_files:
- split: Perf_Optimized
path: C/Perf_Optimized-*
- split: Size_Optimized
path: C/Size_Optimized-*
- config_name: C++
data_files:
- split: Perf_Optimized
path: C++/Perf_Optimized-*
- split: Size_Optimized
path: C++/Size_Optimized-*
- config_name: D
data_files:
- split: Perf_Optimized
path: D/Perf_Optimized-*
- split: Size_Optimized
path: D/Size_Optimized-*
- config_name: Fortran
data_files:
- split: Perf_Optimized
path: Fortran/Perf_Optimized-*
- split: Size_Optimized
path: Fortran/Size_Optimized-*
- config_name: Go
data_files:
- split: Perf_Optimized
path: Go/Perf_Optimized-*
- split: Size_Optimized
path: Go/Size_Optimized-*
- config_name: Haskell
data_files:
- split: Perf_Optimized
path: Haskell/Perf_Optimized-*
- split: Size_Optimized
path: Haskell/Size_Optimized-*
- config_name: Nim
data_files:
- split: Size_Optimized
path: Nim/Size_Optimized-*
- config_name: Objective-C
data_files:
- split: Perf_Optimized
path: Objective-C/Perf_Optimized-*
- split: Size_Optimized
path: Objective-C/Size_Optimized-*
- config_name: Python
data_files:
- split: Perf_Optimized
path: Python/Perf_Optimized-*
- split: Size_Optimized
path: Python/Size_Optimized-*
- config_name: Rust
data_files:
- split: Perf_Optimized
path: Rust/Perf_Optimized-*
- split: Size_Optimized
path: Rust/Size_Optimized-*
- config_name: Swift
data_files:
- split: Perf_Optimized
path: Swift/Perf_Optimized-*
- split: Size_Optimized
path: Swift/Size_Optimized-*
The dataset consists of source code and LLVM IR pairs generated from accepted and de-duped programming contest solutions. The dataset is divided into language configs and mode splits. The language can be one of C
, C++
, D
, Fortran
, Go
, Haskell
, Nim
, Objective-C
, Python
, Rust
and Swift
, indicating the source files' languages. The mode split indicates the compilation mode, which can be wither Size_Optimized
or Perf_Optimized
.
Once you have submitted an access request which has been approved, loading the dataset can be done as follows:
from datasets import load_dataset
dataset = load_dataset("UKPLab/SLTrans", "C", split="Size_Optimized")