KFAC-OBD-Llada-Instruct

BashCache 's Collections

updated about 15 hours ago

GSAI-ML/LLaDA-8B-Instruct structurally pruned with KFAC-OBD (SNR-weighted) at 20/30/40% sparsity, with LoRA adapters.

BashCache/temp-llada-instruct-20-snr-weighted

7B • Updated 24 days ago • 160

Note 20% structured-pruned LLaDA-8B-Instruct (KFAC-OBD, SNR-weighted)
BashCache/temp-llada-instruct-20-snr-weighted-adapters-v2

Text Generation • Updated 21 days ago • 39

Note LoRA adapter recovering the 20% pruned model; load on top of BashCache/temp-llada-instruct-20-snr-weighted
BashCache/temp-llada-instruct-30-snr-weighted

6B • Updated 24 days ago • 55

Note 30% structured-pruned LLaDA-8B-Instruct (KFAC-OBD, SNR-weighted)
BashCache/temp-llada-instruct-30-snr-weighted-adapters-v2

Text Generation • Updated 9 days ago • 21

Note LoRA adapter recovering the 30% pruned model; load on top of BashCache/temp-llada-instruct-30-snr-weighted
BashCache/temp-llada-instruct-40-snr-weighted

6B • Updated 24 days ago • 113

Note 40% structured-pruned LLaDA-8B-Instruct (KFAC-OBD, SNR-weighted)
BashCache/temp-llada-instruct-40-snr-weighted-adapters-v2

Text Generation • Updated 21 days ago • 35

Note LoRA adapter recovering the 40% pruned model; load on top of BashCache/temp-llada-instruct-40-snr-weighted