Instructions to use alexgrigoras/sdg_chronos_t5_small_dunnhumby with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use alexgrigoras/sdg_chronos_t5_small_dunnhumby with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("alexgrigoras/sdg_chronos_t5_small_dunnhumby", dtype="auto") - Notebooks
- Google Colab
- Kaggle
| { | |
| "model_name": "amazon/chronos-t5-small", | |
| "base_model_id": "amazon/chronos-t5-small", | |
| "context_length": 140, | |
| "prediction_length": 30, | |
| "num_bins": 4094, | |
| "value_range": [ | |
| -5.0, | |
| 5.0 | |
| ], | |
| "learning_rate": 2.5e-05, | |
| "train_steps": 300, | |
| "lora_rank": 32, | |
| "lora_alpha": 64, | |
| "batch_size": 2, | |
| "gradient_accumulation_steps": 8, | |
| "max_source_length": 768, | |
| "max_target_length": 256, | |
| "random_state": 42, | |
| "task_prefix": "generate synthetic retail demand future from historical context", | |
| "seasonality_strength": 0.6, | |
| "seasonal_period": 7, | |
| "seasonal_fallback_strength": 0.25, | |
| "zero_threshold_for_sparsity": 0.6, | |
| "prefer_backend": "lora", | |
| "use_special_tokens": true, | |
| "add_calendar_features": true, | |
| "warmup_ratio": 0.05, | |
| "weight_decay": 0.01, | |
| "privacy_reference_max_windows": 2000, | |
| "privacy_min_distance_quantile": 0.25, | |
| "privacy_distance_penalty": 4.0, | |
| "privacy_noise_strength": 0.1, | |
| "privacy_baseline_blend": 0.25, | |
| "privacy_training_jitter_prob": 0.6, | |
| "privacy_training_jitter_strength": 0.1, | |
| "privacy_deduplicate_examples": true, | |
| "privacy_filter_enabled": true, | |
| "privacy_filter_max_retries": 6 | |
| } |