New to Claude Skills? Learn how to install them →

Mdavila7 on GitHub

Model Pruning

Free

Compress LLMs and speed up inference with pruning techniques.

Get this skill

Free · Opens the source repo

What Model Pruning does

Model Pruning is a skill designed for developers and data scientists looking to optimize large language models (LLMs) by reducing their size and enhancing inference speed. By employing advanced pruning techniques such as Wanda and SparseGPT, users can achieve significant model compression—up to 60%—with minimal accuracy loss, typically under 1%. This is particularly beneficial for deploying models on constrained hardware environments like mobile and edge devices, where memory and processing power are limited.

The skill encompasses various pruning strategies, including unstructured pruning, structured pruning, and N:M sparsity. These methods allow for fine-tuning the model's architecture without the need for retraining, making it easier to integrate into existing workflows. The one-shot pruning approach enables users to quickly compress models while maintaining their performance, which is critical in scenarios where rapid deployment is necessary.

Wanda, one of the key techniques included in this skill, utilizes a combination of weight magnitudes and input activations to determine which weights to prune, resulting in more efficient models that retain their functionality. SparseGPT, on the other hand, leverages second-order information to enhance the pruning process, providing a more nuanced approach to model compression. Both methods are supported by detailed examples and code snippets, making it straightforward for users to implement them in their projects.

This skill is ideal for machine learning practitioners who need to optimize their models for deployment without sacrificing performance. Whether you're working on a research project, developing a commercial application, or simply experimenting with LLMs, Model Pruning provides the tools necessary to enhance your models' efficiency and speed.

When to use it

Use Model Pruning when you need to compress LLMs for deployment on resource-constrained devices or when you want to accelerate inference without retraining the model.

When not to use it

This skill may not be suitable for scenarios requiring extensive retraining or fine-tuning of models, as it focuses on one-shot pruning techniques.

What you can build with it

Optimizing LLMs for Mobile Deployment

Use Model Pruning to compress large language models for deployment on mobile devices, ensuring they run efficiently without consuming excessive resources.

Accelerating Inference Times

Implement pruning techniques to significantly speed up inference times in applications where response time is critical.

Reducing Memory Footprint

Apply this skill to reduce the memory footprint of models, making them suitable for environments with limited computational power.

How to install Model Pruning

View source

1. Install with the skills CLI

npx skills add davila7/claude-code-templates/emerging-techniques-model-pruning --agent claude-code

2. Or install it manually

Download the skill folder and drop it into ~/.claude/skills/ for all projects, or .claude/skills/ to scope it to one repo. Restart Claude Code so it picks up the new skill.

Anthropic's agentic coding CLI, and the reference implementation of Agent Skills. Drop a skill folder into ~/.claude/skills and Claude Code loads it automatically whenever a task matches the skill's description. Claude Code docs

Inside SKILL.md

Written by davila7

Model Pruning: Compressing LLMs

When to Use This Skill

Use Model Pruning when you need to:

  • Reduce model size by 40-60% with <1% accuracy loss
  • Accelerate inference using hardware-friendly sparsity (2-4× speedup)
  • Deploy on constrained hardware (mobile, edge devices)
  • Compress without retraining using one-shot methods
  • Enable efficient serving with reduced memory footprint

Key Techniques: Wanda (weights × activations), SparseGPT (second-order), structured pruning, N:M sparsity

Papers: Wanda ICLR 2024 (arXiv 2306.11695), SparseGPT (arXiv 2301.00774)

Installation

# Wanda implementation
git clone https://github.com/locuslab/wanda
cd wanda
pip install -r requirements.txt

# Optional: SparseGPT
git clone https://github.com/IST-DASLab/sparsegpt
cd sparsegpt
pip install -e .

# Dependencies
pip install torch transformers accelerate

Quick Start

Wanda Pruning (One-Shot, No Retraining)

Source: ICLR 2024 (arXiv 2306.11695)

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

# Load model
model = AutoModelForCausalLM.from_pretrained(
    "meta-llama/Llama-2-7b-hf",
    torch_dtype=torch.float16,
    device_map="cuda"
)
tokenizer = AutoTokenizer.from_pretrained("meta-llama/Llama-2-7b-hf")

# Calibration data (small dataset for activation statistics)
calib_data = [
    "The quick brown fox jumps over the lazy dog.",
    "Machine learning is transforming the world.",
    "Artificial intelligence powers modern applications.",
]

# Wanda pruning function
def wanda_prune(model, calib_data, sparsity=0.5):
    """
    Wanda: Prune by weight magnitude × input activation.

    Args:
        sparsity: Fraction of weights to prune (0.5 = 50%)
    """
    # 1. Collect activation statistics
    activations = {}

    def hook_fn(name):
        def hook(module, input, output):
            # Store input activation norms
            activations[name] = input[0].detach().abs().mean(dim=0)
        return hook

    # Register hooks for all linear layers
    hooks = []
    for name, module in model.named_modules():
        if isinstance(module, torch.nn.Linear):
            hooks.append(module.register_forward_hook(hook_fn(name)))

    # Run calibration data
    model.eval()
    with torch.no_grad():
        for text in calib_data:
            inputs = tokenizer(text, return_tensors="pt").to(model.device)
            model(**inputs)

    # Remove hooks
    for hook in hooks:
        hook.remove()

    # 2. Prune weights based on |weight| × activation
    for name, module in model.named_modules():
        if isinstance(module, torch.nn.Linear) and name in activations:
            W = module.weight.data
            act = activations[name]

            # Compute importance: |weight| × activation
            importance = W.abs() * act.unsqueeze(0)

            # Flatten and find threshold
            threshold = torch.quantile(importance.flatten(), sparsity)

            # Create mask
            mask = importance >= threshold

            # Apply mask (prune)
            W *= mask.float()

    return model

# Apply Wanda pruning (50% sparsity, one-shot, no retraining)
pruned_model = wanda_prune(model, calib_data, sparsity=0.5)

# Save
pruned_model.save_pretrained("./llama-2-7b-wanda-50")

SparseGPT (Second-Order Pruning)

Source: arXiv 2301.00774

from sparsegpt import SparseGPT

# Load model
model = AutoModelForCausalLM.from_pretrained("meta-llama/Llama-2-7b-hf")

# Initialize SparseGPT
pruner = SparseGPT(model)

# Calibration data
calib_data = load_calibration_data()  # ~128 samples

# Prune (one-shot, layer-wise reconstruction)
pruned_model = pruner.prune(
    calib_data=calib_data,
    sparsity=0.5,           # 50% sparsity
    prunen=0,               # Unstructured (0) or N:M structured
    prunem=0,
    percdamp=0.01,          # Damping for Hessian inverse
)

# Results: Near-lossless pruning at 50% sparsity

N:M Structured Pruning (Hardware Accelerator)

def nm_prune(weight, n=2, m=4):
    """
    N:M pruning: Keep N weights per M consecutive weights.
    Example: 2:4 = keep 2 out of every 4 weights.

    Compatible with NVIDIA sparse tensor cores (2:4, 4:8).
    """
    # Reshape weight into groups of M
    shape = weight.shape
    weight_flat = weight.flatten()

    # Pad to multiple of M
    pad_size = (m - weight_flat.numel() % m) % m
    weight_padded = F.pad(weight_flat, (0, pad_size))

    # Reshape into (num_groups, m)
    weight_grouped = weight_padded.reshape(-1, m)

    # Find top-N in each group
    _, indices = torch.topk(weight_grouped.abs(), n, dim=-1)

    # Create mask
    mask = torch.zeros_like(weight_grouped)
    mask.scatter_(1, indices, 1.0)

    # Apply mask
    weight_pruned = weight_grouped * mask

    # Reshape back
    weight_pruned = weight_pruned.flatten()[:weight_flat.numel()]
    return weight_pruned.reshape(shape)

# Apply 2:4 sparsity (NVIDIA hardware)
for name, module in model.named_modules():
    if isinstance(module, torch.nn.Linear):
        module.weight.data = nm_prune(module.weight.data, n=2, m=4)

# 50% sparsity, 2× speedup on A100 with sparse tensor cores

Core Concepts

1. Pruning Criteria

Magnitude Pruning (baseline):

# Prune weights with smallest absolute values
importance = weight.abs()
threshold = torch.quantile(importance, sparsity)
mask = importance >= threshold

Wanda (weights × activations):

# Importance = |weight| × input_activation
importance = weight.abs() * activation
# Better than magnitude alone (considers usage)

SparseGPT (second-order):

# Uses Hessian (second derivative) for importance
# More accurate but computationally expensive
importance = weight^2 / diag(Hessian)

2. Structured vs Unstructured

Unstructured (fine-grained):

  • Prune individual weights
  • Higher quality (better accuracy)
  • No hardware speedup (irregular sparsity)

Structured (coarse-grained):

  • Prune entire neurons, heads, or layers
  • Lower quality (more accuracy loss)
  • Hardware speedup (regular sparsity)

Semi-structured (N:M):

  • Best of both worlds
  • 50% sparsity (2:4) → 2× speedup on NVIDIA GPUs
  • Minimal accuracy loss

3. Sparsity Patterns

# Unstructured (random)
# [1, 0, 1, 0, 1, 1, 0, 0]
# Pros: Flexible, high quality
# Cons: No speedup

# Structured (block)
# [1, 1, 0, 0, 1, 1, 0, 0]
# Pros: Hardware friendly
# Cons: More accuracy loss

# N:M (semi-structured)
# [1, 0, 1, 0] [1, 1, 0, 0]  (2:4 pattern)
# Pros: Hardware speedup + good quality
# Cons: Requires specific hardware (NVIDIA)

Pruning Strategies

Strategy 1: Gradual Magnitude Pruning

def gradual_prune(model, initial_sparsity=0.0, final_sparsity=0.5, num_steps=100):
    """Gradually increase sparsity during training."""
    for step in range(num_steps):
        # Current sparsity
        current_sparsity = initial_sparsity + (final_sparsity - initial_sparsity) * (step / num_steps)

        # Prune at current sparsity
        for module in model.modules():
            if isinstance(module, torch.nn.Linear):
                weight = module.weight.data
                threshold = torch.quantile(weight.abs().flatten(), current_sparsity)
                mask = weight.abs() >= threshold
                weight *= mask.float()

        # Train one step
        train_step(model)

    return model

Strategy 2: Layer-wise Pruning

def layer_wise_prune(model, sparsity_per_layer):
    """Different sparsity for different layers."""
    # Early layers: Less pruning (more important)
    # Late layers: More pruning (less critical)

    sparsity_schedule = {
        "layer.0": 0.3,   # 30% sparsity
        "layer.1": 0.4,
        "layer.2": 0.5,
        "layer.3": 0.6,   # 60% sparsity
    }

    for name, module in model.named_modules():
        if isinstance(module, torch.nn.Linear):
            # Find layer index
            for layer_name, sparsity in sparsity_schedule.items():
                if layer_name in name:
                    # Prune at layer-specific sparsity
                    prune_layer(module, sparsity)
                    break

    return model

Strategy 3: Iterative Pruning + Fine-tuning

def iterative_prune_finetune(model, target_sparsity=0.5, iterations=5):
    """Prune gradually with fine-tuning between iterations."""
    current_sparsity = 0.0
    sparsity_increment = target_sparsity / iterations

    for i in range(iterations):
        # Increase sparsity
        current_sparsity += sparsity_increment

        # Prune
        prune_model(model, sparsity=current_sparsity)

        # Fine-tune (recover accuracy)
        fine_tune(model, epochs=2, lr=1e-5)

    return model

# Results: Better accuracy than one-shot at high sparsity

Production Deployment

Complete Pruning Pipeline

from transformers import Trainer, TrainingArguments

def production_pruning_pipeline(
    model_name="meta-llama/Llama-2-7b-hf",
    target_sparsity=0.5,
    method="wanda",  # or "sparsegpt"
):
    # 1. Load model
    model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.float16)
    tokenizer = AutoTokenizer.from_pretrained(model_name)

    # 2. Load calibration data
    calib_dataset = load_dataset("wikitext", "wikitext-2-raw-v1", split="train[:1000]")

    # 3. Apply pruning
    if method == "wanda":
        pruned_model = wanda_prune(model, calib_dataset, sparsity=target_sparsity)
    elif method == "sparsegpt":
        pruner = SparseGPT(model)
        pruned_model = pruner.prune(calib_dataset, sparsity=target_sparsity)

    # 4. (Optional) Fine-tune to recover accuracy
    training_args = TrainingArguments(
        output_dir="./pruned-model",
        num_train_epochs=1,
        per_device_train_batch_size=4,
        learning_rate=1e-5,
        bf16=True,
    )

    trainer = Trainer(
        model=pruned_model,
        args=training_args,
        train_dataset=finetune_dataset,
    )

    trainer.train()

    # 5. Save
    pruned_model.save_pretrained("./pruned-llama-7b-50")
    tokenizer.save_pretrained("./pruned-llama-7b-50")

    return pruned_model

# Usage
pruned_model = production_pruning_pipeline(
    model_name="meta-llama/Llama-2-7b-hf",
    target_sparsity=0.5,
    method="wanda"
)

Evaluation

from lm_eval import evaluator

# Evaluate pruned vs original model
original_results = evaluator.simple_evaluate(
    model="hf",
    model_args="pretrained=meta-llama/Llama-2-7b-hf",
    tasks=["arc_easy", "hellaswag", "winogrande"],
)

pruned_results = evaluator.simple_evaluate(
    model="hf",
    model_args="pretrained=./pruned-llama-7b-50",
    tasks=["arc_easy", "hellaswag", "winogrande"],
)

# Compare
print(f"Original: {original_results['results']['arc_easy']['acc']:.3f}")
print(f"Pruned:   {pruned_results['results']['arc_easy']['acc']:.3f}")
print(f"Degradation: {(original_results - pruned_results):.3f}")

# Typical results at 50% sparsity:
# - Wanda: <1% accuracy loss
# - SparseGPT: <0.5% accuracy loss
# - Magnitude: 2-3% accuracy loss

Best Practices

1. Sparsity Selection

# Conservative (safe)
sparsity = 0.3  # 30%, <0.5% loss

# Balanced (recommended)
sparsity = 0.5  # 50%, ~1% loss

# Aggressive (risky)
sparsity = 0.7  # 70%, 2-5% loss

# Extreme (model-dependent)
sparsity = 0.9  # 90%, significant degradation

2. Method Selection

# One-shot, no retraining → Wanda or SparseGPT
if no_retraining_budget:
    use_method = "wanda"  # Faster

# Best quality → SparseGPT
if need_best_quality:
    use_method = "sparsegpt"  # More accurate

# Hardware speedup → N:M structured
if need_speedup:
    use_method = "nm_prune"  # 2:4 or 4:8

3. Avoid Common Pitfalls

# ❌ Bad: Pruning without calibration data
prune_random(model)  # No activation statistics

# ✅ Good: Use calibration data
prune_wanda(model, calib_data)

# ❌ Bad: Too high sparsity in one shot
prune(model, sparsity=0.9)  # Massive accuracy loss

# ✅ Good: Gradual or iterative
iterative_prune(model, target=0.9, steps=10)

Performance Comparison

Pruning methods at 50% sparsity (LLaMA-7B):

MethodAccuracy LossSpeedMemoryRetraining Needed
Magnitude-2.5%1.0×-50%No
Wanda-0.8%1.0×-50%No
SparseGPT-0.4%1.0×-50%No
N:M (2:4)-1.0%2.0×-50%No
Structured-3.0%2.0×-50%No

Source: Wanda paper (ICLR 2024), SparseGPT paper

Resources

Frequently asked questions about Model Pruning

Similar skills