Views
No views yet
| Model | W_res decomposition | Analog training | Size |
|---|---|---|---|
cheb/ | Chebyshev LRA (min-max) | ✅ 12ep | 3.3G |
pissa/ | SVD (Frobenius optimal) | ✅ 12ep | 3.3G |
lora_analog/ | None (pretrained) | ✅ 12ep | 3.3G |
lora/ | None (pretrained) | ❌ digital | 2.0G |
cheb_init.pt (1.4G) — Chebyshev LRA W_res for cheb model. eq_init_v3.pt (832M) — RMSNorm-folded gate/up variant.1conda create -n pissa-dora python=3.10
2conda activate pissa-dora
3pip install torch transformers peft datasets accelerate
4pip install aihwkit # IBM analog hardware simulator
5pip install huggingface_hub # for model downloadpython eval_noise.py --model cheb --noise 1.0--model: cheb | pissa | lora_analog | lora (default: cheb)--noise: 0 (digital) | 1.0 | 2.0 (default: 1.0)--runs N: repeat N times for statistics (default: 1)python eval_drift.py --model cheb --time 1year --runs 3--time: t0 | 1min | 1hr | 1day | 1week | 1month | 1year | 10year--runs N: independent programming + drift repeats (default: 3)| Model | digital | n=1.0 | n=2.0 | 10yr drift |
|---|---|---|---|---|
| Cheb | 32.5% | 28.1% | 20.8% | 16.5% |
| PiSSA | 31.3% | 24.7% | 20.2% | 15.3% |
| LoRA+analog | 21.6% | 20.2% | 14.3% | 10.4% |
| LoRA (digital) | 27.1% | 1.1% | 1.0% | 1.0% |