Spaces:
Sleeping
Sleeping
| import yaml | |
| import gradio as gr | |
| from yourbench_space.utils import CONFIG_PATH | |
| def generate_config(hf_token: gr.OAuthToken | None, hf_org, model_name, provider, base_url, api_key, max_concurrent_requests): | |
| config = { | |
| "hf_configuration": { | |
| "token": hf_token, | |
| "private": True, | |
| "hf_organization": hf_org | |
| }, | |
| "model_list": [{ | |
| "model_name": model_name, | |
| "provider": provider, | |
| "base_url": base_url, | |
| "api_key": api_key, | |
| "max_concurrent_requests": max_concurrent_requests | |
| }], | |
| "model_roles": {role: [model_name] for role in [ | |
| "ingestion", "summarization", "single_shot_question_generation", | |
| "multi_hop_question_generation", "answer_generation", "judge_answers" | |
| ]}, | |
| "inference_config": {"max_concurrent_requests": 16}, | |
| "pipeline": { | |
| "ingestion": { | |
| "source_documents_dir": "/app/uploaded_files", | |
| "output_dir": "/app/ingested", | |
| "run": True | |
| }, | |
| "upload_ingest_to_hub": { | |
| "source_documents_dir": "/app/ingested", | |
| "hub_dataset_name": "test_ingested_documents", | |
| "local_dataset_path": "/app/ingested_dataset", | |
| "run": True | |
| }, | |
| "summarization": { | |
| "source_dataset_name": "test_ingested_documents", | |
| "output_dataset_name": "test_summaries", | |
| "local_dataset_path": "/results/test_summaries", | |
| "concat_existing_dataset": False, | |
| "run": True | |
| }, | |
| "chunking": { | |
| "source_dataset_name": "test_summaries", | |
| "output_dataset_name": "test_chunked_documents", | |
| "local_dataset_path": "/results/test_chunked_documents", | |
| "concat_existing_dataset": False, | |
| "chunking_configuration": { | |
| "l_min_tokens": 64, | |
| "l_max_tokens": 128, | |
| "tau_threshold": 0.3, | |
| "h_min": 2, | |
| "h_max": 4 | |
| }, | |
| "run": True | |
| }, | |
| "single_shot_question_generation": { | |
| "source_dataset_name": "test_chunked_documents", | |
| "output_dataset_name": "test_single_shot_questions", | |
| "local_dataset_path": "/results/test_single_shot_questions", | |
| "diversification_seed": "24 year old adult", | |
| "concat_existing_dataset": False, | |
| "run": True | |
| }, | |
| "multi_hop_question_generation": { | |
| "source_dataset_name": "test_chunked_documents", | |
| "output_dataset_name": "test_multi_hop_questions", | |
| "local_dataset_path": "/results/test_multi_hop_questions", | |
| "concat_existing_dataset": False, | |
| "run": True | |
| }, | |
| "answer_generation": { | |
| "run": True, | |
| "question_dataset_name": "test_single_shot_questions", | |
| "output_dataset_name": "test_answered_questions", | |
| "local_dataset_path": "/results/test_answered_questions", | |
| "concat_existing_dataset": False, | |
| "strategies": [{ | |
| "name": "zeroshot", | |
| "prompt": "ZEROSHOT_QA_USER_PROMPT", | |
| "model_name": model_name | |
| }, { | |
| "name": "gold", | |
| "prompt": "GOLD_QA_USER_PROMPT", | |
| "model_name": model_name | |
| }] | |
| }, | |
| "judge_answers": { | |
| "run": True, | |
| "source_judge_dataset_name": "test_answered_questions", | |
| "output_judged_dataset_name": "test_judged_comparisons", | |
| "local_dataset_path": "/results/test_judged_comparisons", | |
| "concat_existing_dataset": False, | |
| "comparing_strategies": [["zeroshot", "gold"]], | |
| "chunk_column_index": 0, | |
| "random_seed": 42 | |
| } | |
| } | |
| } | |
| return yaml.dump(config, default_flow_style=False) | |
| def save_config(yaml_text): | |
| with open(CONFIG_PATH, "w") as file: | |
| file.write(yaml_text) | |
| return "✅ Config saved!" | |