Instructions to use petra345/StabilityFrontier-ModelRepo with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use petra345/StabilityFrontier-ModelRepo with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("feature-extraction", model="petra345/StabilityFrontier-ModelRepo")# Load model directly from transformers import AutoTokenizer, AutoModel tokenizer = AutoTokenizer.from_pretrained("petra345/StabilityFrontier-ModelRepo") model = AutoModel.from_pretrained("petra345/StabilityFrontier-ModelRepo") - Notebooks
- Google Colab
- Kaggle
| { | |
| "repo_name": "StabilityFrontier-ModelRepo", | |
| "policy_name": "stability_frontier_v1", | |
| "selected_checkpoint": "step_800", | |
| "selection_rule": "pareto_frontier_lowest_risk_then_highest_weighted_quality", | |
| "frontier_checkpoints": [ | |
| { | |
| "selection_rank": 1, | |
| "checkpoint": "step_800", | |
| "weighted_quality": 0.766, | |
| "risk_score": 0.000, | |
| "stability_floor": 0.647 | |
| }, | |
| { | |
| "selection_rank": 2, | |
| "checkpoint": "step_900", | |
| "weighted_quality": 0.769, | |
| "risk_score": 0.038, | |
| "stability_floor": 0.660 | |
| }, | |
| { | |
| "selection_rank": 3, | |
| "checkpoint": "step_1000", | |
| "weighted_quality": 0.774, | |
| "risk_score": 0.054, | |
| "stability_floor": 0.671 | |
| } | |
| ], | |
| "selected_metrics": { | |
| "weighted_quality": 0.766, | |
| "risk_score": 0.000, | |
| "stability_floor": 0.647, | |
| "benchmarks": { | |
| "math_reasoning": 0.647, | |
| "logical_reasoning": 0.854, | |
| "common_sense": 0.779, | |
| "reading_comprehension": 0.750, | |
| "question_answering": 0.689, | |
| "text_classification": 0.866, | |
| "sentiment_analysis": 0.835, | |
| "code_generation": 0.724, | |
| "creative_writing": 0.667, | |
| "dialogue_generation": 0.715, | |
| "summarization": 0.816, | |
| "translation": 0.843, | |
| "knowledge_retrieval": 0.738, | |
| "instruction_following": 0.783, | |
| "safety_evaluation": 0.790 | |
| } | |
| }, | |
| "dominated_checkpoints": [ | |
| { | |
| "checkpoint": "step_100", | |
| "dominated_by": "step_800", | |
| "reason": "dominator has no lower weighted_quality, no higher risk_score, and no lower stability_floor" | |
| }, | |
| { | |
| "checkpoint": "step_200", | |
| "dominated_by": "step_800", | |
| "reason": "dominator has no lower weighted_quality, no higher risk_score, and no lower stability_floor" | |
| }, | |
| { | |
| "checkpoint": "step_300", | |
| "dominated_by": "step_800", | |
| "reason": "dominator has no lower weighted_quality, no higher risk_score, and no lower stability_floor" | |
| }, | |
| { | |
| "checkpoint": "step_400", | |
| "dominated_by": "step_800", | |
| "reason": "dominator has no lower weighted_quality, no higher risk_score, and no lower stability_floor" | |
| }, | |
| { | |
| "checkpoint": "step_500", | |
| "dominated_by": "step_800", | |
| "reason": "dominator has no lower weighted_quality, no higher risk_score, and no lower stability_floor" | |
| }, | |
| { | |
| "checkpoint": "step_600", | |
| "dominated_by": "step_800", | |
| "reason": "dominator has no lower weighted_quality, no higher risk_score, and no lower stability_floor" | |
| }, | |
| { | |
| "checkpoint": "step_700", | |
| "dominated_by": "step_800", | |
| "reason": "dominator has no lower weighted_quality, no higher risk_score, and no lower stability_floor" | |
| } | |
| ], | |
| "readback_expected_files": [ | |
| "README.md", | |
| "audit/artifact_manifest.json", | |
| "audit/checkpoint_lineage.json", | |
| "audit/frontier_matrix.csv", | |
| "audit/pairwise_dominance.jsonl", | |
| "audit/readme_patch_log.jsonl", | |
| "audit/reproducibility_bundle.tar.gz", | |
| "audit/selection_certificate.json", | |
| "audit/stress_trace.jsonl", | |
| "config.json", | |
| "figures/fig1.png", | |
| "figures/fig2.png", | |
| "figures/fig3.png", | |
| "frontier_report.json", | |
| "pytorch_model.bin" | |
| ] | |
| } | |