gemma-2-2b-it-sculpt-experimental/README.md

license, library_name, pipeline_tag, language, base_model, tags, datasets, model-index

license	library_name	pipeline_tag	language	base_model	tags	datasets	model-index
apache-2.0	transformers	text-generation	en	google/gemma-2-2b-it	dystrio sculpt pruned compressed efficient dense runtime-agnostic no-custom-kernels hf-drop-in drop-in-replacement smaller faster gemma	wikitext	name results Dystrio Sculpt (gemma-2-2b-it Experimental) task dataset metrics type text-generation name type WikiText-103 (validation) wikitext name type value perplexity perplexity 48.9699 name type value ppl_ratio ppl_ratio 1.8995

dystrio/gemma-2-2b-it-sculpt-experimental

30% smaller, +23% faster prefill, drop-in replacement. No custom kernels. No runtime changes.

Dystrio Sculpt structurally compresses transformer models, producing dense models that load with standard transformers — no custom code, no new ops, no deployment friction.

This is the Experimental tier of gemma 2 2b it.

Quick Start

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("dystrio/gemma-2-2b-it-sculpt-experimental", torch_dtype="bfloat16", device_map="auto")
tokenizer = AutoTokenizer.from_pretrained("dystrio/gemma-2-2b-it-sculpt-experimental")

inputs = tokenizer("The future of AI inference is", return_tensors="pt").to(model.device)
outputs = model.generate(**inputs, max_new_tokens=100)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Benchmark Results

All tiers compiled from gemma 2 2b it on A100 80GB, bf16:

Model	PPL	PPL Ratio	Weights (GB)	Chat Prefill TPS	RAG TTFT p95 (ms)	Decode TPS
Baseline	25.7807	1.0	4.869591	21611.9	70.251	59.6
sculpt-default	20.5854	0.7985	4.441124	23065.3	69.007	60.0
sculpt-production	22.4118	0.8693	4.226891	23404.0	66.554	60.7
sculpt-throughput	29.8372	1.1573	3.969811	24330.1	64.529	59.3
sculpt-experimental	48.9699	1.8995	3.412804	26496.2	60.97	59.5

Key Metrics (this model)

Metric	Value
Weights memory	3.412804 GB (30% smaller)
PPL ratio	1.8995
Chat prefill TPS	26496.2 (+23%)
RAG TTFT p95	60.97 ms (-13%)
Decode TPS	59.5 (flat)
Parameters	1.83B

All Sculpt Tiers

Tier	HuggingFace	Size	PPL Ratio	Use Case
default	dystrio/gemma-2-2b-it-sculpt-default	4.441124 GB	0.7985	Zero-regret: quality preserved, smaller footprint
production	dystrio/gemma-2-2b-it-sculpt-production	4.226891 GB	0.8693	Practical savings with modest quality tradeoff
throughput	dystrio/gemma-2-2b-it-sculpt-throughput	3.969811 GB	1.1573	Maximum usable compression for speed/edge
experimental	dystrio/gemma-2-2b-it-sculpt-experimental 👈 this model	3.412804 GB	1.8995	Boundary exploration, maximum structural compression

What is Dystrio Sculpt?

Dystrio Sculpt compiles transformer models into smaller, faster variants. Output models:

• Are dense (not sparse) — standard architecture, fewer parameters
• Load with standard HuggingFace Transformers — no custom code needed
• Require no custom kernels and no runtime changes
• Work as a one-step compile before deployment
• Stack with quantization (AWQ, GPTQ, GGUF) for compound savings

Compatibility

• ✅ HuggingFace Transformers
• ✅ vLLM
• ✅ TGI (Text Generation Inference)
• ✅ llama.cpp / GGUF conversion
• ✅ AWQ / GPTQ quantization
• ✅ Any framework that loads standard safetensors

Benchmark Environment

• GPU: NVIDIA A100-SXM4-80GB
• dtype: bf16
• Torch: 2.10.0+cu128
• Transformers: 5.3.0
• Deterministic: True
• Single-GPU, standard HuggingFace Transformers, no custom kernels.

Metric Definitions

• PPL ratio: WikiText-103 perplexity relative to baseline. <1.0 = quality improved.
• Prefill TPS: Tokens per second during prompt encoding (higher = faster).
• TTFT p95: Time to first token at 95th percentile (lower = faster).
• Decode TPS: Tokens per second during generation (higher = faster).
• Weights (GB): Model parameter memory (deterministic, runtime-independent).

Citation

@misc{dystrio_sculpt_2026,
  title={Dystrio Sculpt: Structural Compilation for Transformer LLMs},
  author={Dystrio},
  year={2026},
  url={https://huggingface.co/dystrio}
}

Downstream Benchmarks (lm-eval)

Evaluated with lm-eval-harness on A100-80GB, bf16, zero-shot.

Benchmark	Baseline	This Model	Delta
ARC-Challenge	0.5094	0.2645	-0.2449
HellaSwag	0.5375	0.3577	-0.1798
MMLU	0.5691	0.2304	-0.3387
TruthfulQA MC2	0.5322	0.5040	-0.0282