bsmith925/cnake-charmer

---
dataset_info:
- config_name: candidates
features:
- name: blob_id
dtype: string
- name: filename
dtype: string
- name: path
dtype: string
- name: repo_name
dtype: string
- name: length_bytes
dtype: int64
- name: src_encoding
dtype: string
- name: content
dtype: string
- name: split
dtype: string
- name: detected_licenses
list: string
- name: license_type
dtype: string
- name: star_events_count
dtype: int64
splits:
- name: train
num_bytes: 4315753
num_examples: 1000
download_size: 1648199
dataset_size: 4315753
- config_name: curated
features:
- name: problem_id
dtype: string
- name: category
dtype: string
- name: difficulty
dtype: string
- name: func_name
dtype: string
- name: description
dtype: string
- name: python_code
dtype: string
- name: cython_code
dtype: string
- name: test_code
dtype: string
- name: test_cases_json
dtype: string
- name: benchmark_args
dtype: string
- name: speedup
dtype: float64
- name: py_avg_sec
dtype: float64
- name: cy_avg_sec
dtype: float64
splits:
- name: train
num_bytes: 2605863
num_examples: 692
download_size: 961059
dataset_size: 2605863
- config_name: full
features:
- name: blob_id
dtype: string
- name: filename
dtype: string
- name: path
dtype: string
- name: repo_name
dtype: string
- name: length_bytes
dtype: int64
- name: src_encoding
dtype: string
- name: content
dtype: string
- name: split
dtype: 'null'
- name: detected_licenses
list: string
- name: license_type
dtype: string
- name: star_events_count
dtype: int64
- name: extension
dtype: string
- name: language
dtype: string
- name: is_generated
dtype: bool
- name: is_vendor
dtype: bool
splits:
- name: train
num_bytes: 69762578
num_examples: 21799
download_size: 28178150
dataset_size: 69762578
- config_name: sft
features:
- name: messages
list:
- name: content
dtype: string
- name: name
dtype: string
- name: role
dtype: string
- name: tool_call_id
dtype: string
- name: tool_calls
list:
- name: function
struct:
- name: arguments
struct:
- name: '0'
dtype: string
- name: annotate
dtype: bool
- name: benchmark
dtype: bool
- name: code
dtype: string
- name: nogil
dtype: string
- name: test
dtype: bool
- name: name
dtype: string
- name: id
dtype: string
- name: type
dtype: string
- name: model
dtype: string
- name: prompt_id
dtype: string
- name: problem_id
dtype: string
- name: func_name
dtype: string
- name: category
dtype: string
- name: difficulty
dtype: string
- name: num_iterations
dtype: int64
- name: sft_score
dtype: float64
- name: speedup
dtype: float64
- name: thinking
dtype: bool
splits:
- name: train
num_bytes: 3905832
num_examples: 410
download_size: 3886177
dataset_size: 3905832
- config_name: traces
features:
- name: problem_id
dtype: string
- name: model
dtype: string
- name: prompt_id
dtype: string
- name: attempt
dtype: int64
- name: timestamp
dtype: string
- name: func_name
dtype: string
- name: category
dtype: string
- name: difficulty
dtype: string
- name: thinking
dtype: bool
- name: version
dtype: string
- name: num_iterations
dtype: int64
- name: tools_used
dtype: string
- name: final_code
dtype: 'null'
- name: reward
dtype: float64
- name: steps
dtype: string
- name: trailing_reasoning
dtype: string
- name: metrics
dtype: string
splits:
- name: train
num_bytes: 325971
num_examples: 1497
download_size: 56111
dataset_size: 325971
configs:
- config_name: candidates
data_files:
- split: train
path: candidates/train-*
- config_name: curated
data_files:
- split: train
path: curated/train-*
- config_name: full
data_files:
- split: train
path: full/train-*
- config_name: sft
data_files:
- split: train
path: sft/train-*
- config_name: traces
data_files:
- split: train
path: traces/train-*
---