Andandd/amy-checkpoints

Amy Checkpoints (Encrypted)

All .pt checkpoint files are encrypted with AES-256-GCM.

Structure

• v1_cache_encrypted/ — Best models, V2 steps, EMA variants, SWA, training milestones (22 files)
• v1_desktop_encrypted/ — Full training run series from Desktop (174 files)
• decrypt_checkpoints.py — Decryption script

How to decrypt

1. Install dependencies:

   pip install cryptography huggingface_hub
   

2. Clone this repo:

   huggingface-cli download Andandd/amy-checkpoints --local-dir ./amy-checkpoints
   

3. Set your decryption key and run:

   export AMY_DECRYPT_KEY=<your-key-here>
   python decrypt_checkpoints.py ./amy-checkpoints/v1_cache_encrypted ./checkpoints_decrypted
   

Encryption details

• Algorithm: AES-256-GCM (authenticated encryption)
• Each file: [12-byte random nonce][ciphertext + 16-byte auth tag]
• Key is NOT stored in this repo

File inventory

v1_cache (C: drive, curated)

• amy_best.pt, amy_best_ema.pt — Best model + exponential moving average
• amy_best_old_507M.pt, amy_best_ema_old_507M.pt — Earlier 507M architecture
• amy_swa.pt — Stochastic weight average
• amy_v2_step_*.pt — V2 architecture training (steps 100, 200, 1000)
• amy_train_*/amy_emergency_* — Training run milestones (May 2026)
• amy_curriculum.pt, amy_spaced_rep.pt, amy_learnability_rw.pt, amy_selective_loss.pt — Training metadata

v1_desktop (full series)

• 174 checkpoints from April 2026 training runs (steps 1 through 16000)