Instructions to use prodigyhuh/atomicvision-hard-recall-micro-boost-lora with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use prodigyhuh/atomicvision-hard-recall-micro-boost-lora with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3-1.7B") model = PeftModel.from_pretrained(base_model, "prodigyhuh/atomicvision-hard-recall-micro-boost-lora") - Notebooks
- Google Colab
- Kaggle
AtomicVision Hard Recall Micro Boost LoRA
This adapter materializes the promoted checkpoint-1 checkpoint from the
targeted hard recall micro-repair continuation.
Parent
- Base adapter: prodigyhuh/atomicvision-medium-fidelity-boost-lora
- Source HF job: 69ed269fd70108f37acdef6d
- Source commit:
3838f9048bce4c6bc81e57f5c0dab00980c7fa08
Held-out strict eval (seed_start=10000, episodes=32)
| Difficulty | Reward | F1 | MAE | Strict | Normalized | Done | Submit |
|---|---|---|---|---|---|---|---|
| medium | 4.5065 | 0.7891 | 0.02712 | 1.00 | 1.00 | 1.00 | 1.00 |
| hard | 4.7148 | 0.8207 | 0.02552 | 1.00 | 1.00 | 1.00 | 1.00 |
This run preserves perfect strict execution and slightly improves the hard slice over the previous best published adapter without regressing medium.
- Downloads last month
- -