BERT-Base (cur_compressor) — NeuralOps Compressed & LoRA-Healed Model Zoo

This repository provides a physically compressed and LoRA-healed checkpoint of google-bert/bert-base-uncased fine-tuned on GLUE SST-2, compressed using the cur_compressor (CURCompressor) engine from the NeuralOps Library at Pareto-selected rank k=128 (2.33x compression ratio, -57.1% parameter reduction).

Compression & Architecture Profile

CURCompressor: CURCompressor neural operator. CUR Compressor: replaces Linear and Conv1D layers with CUR decomposition. Uses leverage score sampling or energy sampling for column/row selection. Registered in COMPRESSOR_REGISTRY as 'cur' and 'cur_compressor'.

Property	Specification
Registry Key (`COMPRESSOR_REGISTRY`)	`cur_compressor`
NeuralOps Class & Family	`CURCompressor` (`low_rank`)
Source Implementation	`neuralops/compression/low_rank/cur_compressor.py`
Base Architecture	`google-bert/bert-base-uncased` (`109,483,778` parameters, 12 layers, $D=768$)
Showcase Rank (`k`)	`128` (`LoRA Recovery Applied: True`)
Compressed Parameter Count	`46,962,434` (`-57.11%` reduction, `2.331x` physical compression ratio)
Teacher Baseline SST-2 Acc	`91.63%`
Showcase Final SST-2 Acc	`72.48%` (Zero-Shot: `50.92%`, Parity Delta: `-19.15%`)

Multi-Rank Pareto Evaluation (`k=256, 128, 64, 32`)

Evaluated physically on NVIDIA GeForce RTX 3090 (24GB) on the GLUE SST-2 validation split (872 samples). Whenever zero-shot compression drops accuracy by > 2.0% alongside > 1.10x reduction, NeuralOps applies targeted r=8 LoRA healing (50 steps, lr=2e-4, 45s budget):

Rank (`k`)	Physical Ratio	Compressed Params (Reduction)	Zero-Shot Acc	Healed Acc	Parity vs Teacher (`91.63%`)
`256`	`1.53x`	`71,735,042` (`-34.5%`)	`56.54%`	`49.08%` (LoRA)	`-42.55%`
`128` ⭐	`2.33x`	`46,962,434` (`-57.1%`)	`50.92%`	`72.48%` (LoRA)	`-19.15%`
`64`	`3.09x`	`35,460,866` (`-67.6%`)	`49.31%`	`49.08%` (LoRA)	`-42.55%`
`32`	`3.66x`	`29,931,266` (`-72.7%`)	`49.08%`	`50.92%` (LoRA)	`-40.71%`

(⭐ indicates the serialized checkpoint packaged in this Hugging Face repository.)

Packaged Artifacts

pytorch_model.bin: Full serialized compressed BERT state dictionary (with low-rank factors / quantization buffers / LoRA healing adapters baked in).
config.json: Model configuration tagged with compression_method: "cur_compressor" and compression_rank: 128.
compression_summary.json: Complete machine-readable telemetry across all evaluated ranks [256, 128, 64, 32].
Tokenizer Assets: Standard bert-base-uncased vocabulary (vocab.txt, tokenizer.json, tokenizer_config.json).

Quick Start (Reconstruction & Inference)

import torch
from huggingface_hub import hf_hub_download
from transformers import AutoTokenizer, BertForSequenceClassification
from neuralops.core.compression_pipeline import compress_model

repo_id = "tahamajs/bert-comp-cur-compressor-sst2"
tokenizer = AutoTokenizer.from_pretrained(repo_id)

# Reconstruct compressed NeuralOps topology and load serialized compressed + healed weights
model = BertForSequenceClassification.from_pretrained("google-bert/bert-base-uncased", num_labels=2)
model = compress_model(model, {"compressors": [{"method": "cur_compressor", "rank": 128}]})

ckpt_path = hf_hub_download(repo_id=repo_id, filename="pytorch_model.bin")
state_dict = torch.load(ckpt_path, map_location="cpu")
model.load_state_dict(state_dict, strict=False)
model.eval()

inputs = tokenizer("NeuralOps achieves strong compression without sacrificing downstream accuracy.", return_tensors="pt")
with torch.no_grad():
    logits = model(**inputs).logits
    pred = logits.argmax(dim=-1).item()
print("Predicted SST-2 Sentiment (0=Negative, 1=Positive):", pred)

Citation

@software{neuralops2026compressorzoo,
  title   = {NeuralOps: Multi-Rank Compressed BERT-SST2 Zoo with Pareto LoRA Recovery},
  author  = {Majlesi, Taha},
  year    = {2026},
  url     = {https://github.com/tahamajs/NeuralOps}
}