| operators: | |
| - id: load_hf_python | |
| config: | |
| type: hf_source | |
| dataset: prithivMLmods/Coder-Stat | |
| split: train | |
| trust_remote_code: True | |
| - id: sample_dataset | |
| config: | |
| type: function | |
| function: data_strategies.commons.uniform_sample_limited | |
| function_config: | |
| num_samples: 125 # 1.25% of 10000 target from stage 1 | |
| input_ids: | |
| - load_hf_python | |
| - id: generate_instructions | |
| config: | |
| type: completions | |
| map: chat | |
| map_config: | |
| system_message: "You are a helpful assistant." | |
| user_message: | | |
| You are to generate a question or task for a language model based on the following error and code pairs. | |
| Error Type: {{original_status}} | |
| Code: {{original_src}} | |
| Include only the new question and task. Do not include anything like "Here is the instruction". Include | |
| the code in your question and make the task sound like what a human would ask a language model. | |
| output_column: instruction_seed | |
| model: gpt-4o-mini | |
| temperature: 1.0 | |
| input_ids: | |
| - sample_dataset | |
| - id: add_constant_column | |
| config: | |
| type: function | |
| function: data_strategies.commons.add_constant_columns | |
| function_config: | |
| response_seed: "" | |
| input_ids: | |
| - generate_instructions | |
| - id: rename_task_column | |
| config: | |
| type: function | |
| function: data_strategies.commons.keep_only_columns | |
| function_config: | |
| columns_to_keep: | |
| - response_seed | |
| - instruction_seed | |
| input_ids: | |
| - add_constant_column |