LLM Inference — llama.cpp + vLLM

This skill covers running LLM inference: local/edge with llama.cpp (GGUF) and production serving with vLLM (PagedAttention). Includes Hugging Face Hub model discovery.

Which tool to use?

Rule of thumb: llama.cpp for single-user/edge/CPU; vLLM for multi-user/GPU/production. Both expose OpenAI-compatible endpoints.

llama.cpp + GGUF

Use this section for local GGUF inference, quant selection, or Hugging Face repo discovery for llama.cpp.

When to use llama.cpp

• Run local models on CPU, Apple Silicon, CUDA, ROCm, or Intel GPUs
• Find the right GGUF for a specific Hugging Face repo
• Build a llama-server or llama-cli command from the Hub
• Search the Hub for models that already support llama.cpp
• Enumerate available .gguf files and sizes for a repo
• Decide between Q4/Q5/Q6/IQ variants for the user’s RAM or VRAM

Model Discovery workflow

Prefer URL workflows before asking for hf, Python, or custom scripts.

1. Search for candidate repos on the Hub:
   • Base: https://huggingface.co/models?apps=llama.cpp&sort=trending
   • Add search=<term> for a model family
   • Add num_parameters=min:0,max:24B or similar when the user has size constraints
2. Open the repo with the llama.cpp local-app view:
   • https://huggingface.co/<repo>?local-app=llama.cpp
3. Treat the local-app snippet as the source of truth when it is visible:
   • copy the exact llama-server or llama-cli command
   • report the recommended quant exactly as HF shows it
4. Read the same ?local-app=llama.cpp URL as page text or HTML and extract the section under Hardware compatibility:
   • prefer its exact quant labels and sizes over generic tables
   • keep repo-specific labels such as UD-Q4_K_M or IQ4_NL_XL
   • if that section is not visible in the fetched page source, say so and fall back to the tree API plus generic quant guidance
5. Query the tree API to confirm what actually exists:
   • https://huggingface.co/api/models/<repo>/tree/main?recursive=true
   • keep entries where type is file and path ends with .gguf
   • use path and size as the source of truth for filenames and byte sizes
   • separate quantized checkpoints from mmproj-*.gguf projector files and BF16/ shard files
   • use https://huggingface.co/<repo>/tree/main only as a human fallback
6. If the local-app snippet is not text-visible, reconstruct the command from the repo plus the chosen quant:
   • shorthand quant selection: llama-server -hf <repo>:<QUANT>
   • exact-file fallback: llama-server --hf-repo <repo> --hf-file <filename.gguf>
7. Only suggest conversion from Transformers weights if the repo does not already expose GGUF files.

Quick start

Install llama.cpp

# macOS / Linux (simplest)
brew install llama.cpp

winget install llama.cpp

git clone https://github.com/ggml-org/llama.cpp
cd llama.cpp
cmake -B build
cmake --build build --config Release

Run directly from the Hugging Face Hub

llama-cli -hf bartowski/Llama-3.2-3B-Instruct-GGUF:Q8_0

llama-server -hf bartowski/Llama-3.2-3B-Instruct-GGUF:Q8_0

Run an exact GGUF file from the Hub

Use this when the tree API shows custom file naming or the exact HF snippet is missing.

llama-server \
    --hf-repo microsoft/Phi-3-mini-4k-instruct-gguf \
    --hf-file Phi-3-mini-4k-instruct-q4.gguf \
    -c 4096

OpenAI-compatible server check

curl http://localhost:8080/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "messages": [
      {"role": "user", "content": "Write a limerick about Python exceptions"}
    ]
  }'

Python bindings (llama-cpp-python)

pip install llama-cpp-python (CUDA: CMAKE_ARGS="-DGGML_CUDA=on" pip install llama-cpp-python --force-reinstall --no-cache-dir; Metal: CMAKE_ARGS="-DGGML_METAL=on" ...).

Basic generation

from llama_cpp import Llama

llm = Llama(
    model_path="./model-q4_k_m.gguf",
    n_ctx=4096,
    n_gpu_layers=35,     # 0 for CPU, 99 to offload everything
    n_threads=8,
)

out = llm("What is machine learning?", max_tokens=256, temperature=0.7)
print(out["choices"][0]["text"])

Chat + streaming

llm = Llama(
    model_path="./model-q4_k_m.gguf",
    n_ctx=4096,
    n_gpu_layers=35,
    chat_format="llama-3",   # or "chatml", "mistral", etc.
)

resp = llm.create_chat_completion(
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "What is Python?"},
    ],
    max_tokens=256,
)
print(resp["choices"][0]["message"]["content"])

# Streaming
for chunk in llm("Explain quantum computing:", max_tokens=256, stream=True):
    print(chunk["choices"][0]["text"], end="", flush=True)

Embeddings

llm = Llama(model_path="./model-q4_k_m.gguf", embedding=True, n_gpu_layers=35)
vec = llm.embed("This is a test sentence.")
print(f"Embedding dimension: {len(vec)}")

You can also load a GGUF straight from the Hub:

llm = Llama.from_pretrained(
    repo_id="bartowski/Llama-3.2-3B-Instruct-GGUF",
    filename="*Q4_K_M.gguf",
    n_gpu_layers=35,
)

Choosing a quant

Use the Hub page first, generic heuristics second.

• Prefer the exact quant that HF marks as compatible for the user’s hardware profile.
• For general chat, start with Q4_K_M.
• For code or technical work, prefer Q5_K_M or Q6_K if memory allows.
• For very tight RAM budgets, consider Q3_K_M, IQ variants, or Q2 variants only if the user explicitly prioritizes fit over quality.
• For multimodal repos, mention mmproj-*.gguf separately. The projector is not the main model file.
• Do not normalize repo-native labels. If the page says UD-Q4_K_M, report UD-Q4_K_M.

Extracting available GGUFs from a repo

When the user asks what GGUFs exist, return:

• filename
• file size
• quant label
• whether it is a main model or an auxiliary projector

Ignore unless requested:

• README
• BF16 shard files
• imatrix blobs or calibration artifacts

Use the tree API for this step:

• https://huggingface.co/api/models/<repo>/tree/main?recursive=true

For a repo like unsloth/Qwen3.6-35B-A3B-GGUF, the local-app page can show quant chips such as UD-Q4_K_M, UD-Q5_K_M, UD-Q6_K, and Q8_0, while the tree API exposes exact file paths such as Qwen3.6-35B-A3B-UD-Q4_K_M.gguf and Qwen3.6-35B-A3B-Q8_0.gguf with byte sizes. Use the tree API to turn a quant label into an exact filename.

Search patterns

Use these URL shapes directly:

https://huggingface.co/models?apps=llama.cpp&sort=trending
https://huggingface.co/models?search=<term>&apps=llama.cpp&sort=trending
https://huggingface.co/models?search=<term>&apps=llama.cpp&num_parameters=min:0,max:24B&sort=trending
https://huggingface.co/<repo>?local-app=llama.cpp
https://huggingface.co/api/models/<repo>/tree/main?recursive=true
https://huggingface.co/<repo>/tree/main

Output format

When answering discovery requests, prefer a compact structured result like:

Repo: <repo>
Recommended quant from HF: <label> (<size>)
llama-server: <command>
Other GGUFs:
- <filename> - <size>
- <filename> - <size>
Source URLs:
- <local-app URL>
- <tree API URL>

References

• hub-discovery.md - URL-only Hugging Face workflows, search patterns, GGUF extraction, and command reconstruction
• advanced-usage.md — speculative decoding, batched inference, grammar-constrained generation, LoRA, multi-GPU, custom builds, benchmark scripts
• quantization.md — quant quality tradeoffs, when to use Q4/Q5/Q6/IQ, model size scaling, imatrix
• server.md — direct-from-Hub server launch, OpenAI API endpoints, Docker deployment, NGINX load balancing, monitoring
• optimization.md — CPU threading, BLAS, GPU offload heuristics, batch tuning, benchmarks
• troubleshooting.md — install/convert/quantize/inference/server issues, Apple Silicon, debugging

Resources

• GitHub: https://github.com/ggml-org/llama.cpp
• Hugging Face GGUF + llama.cpp docs: https://huggingface.co/docs/hub/gguf-llamacpp
• Hugging Face Local Apps docs: https://huggingface.co/docs/hub/main/local-apps
• Hugging Face Local Agents docs: https://huggingface.co/docs/hub/agents-local
• Example local-app page: https://huggingface.co/unsloth/Qwen3.6-35B-A3B-GGUF?local-app=llama.cpp
• Example tree API: https://huggingface.co/api/models/unsloth/Qwen3.6-35B-A3B-GGUF/tree/main?recursive=true
• Example llama.cpp search: https://huggingface.co/models?num_parameters=min:0,max:24B&apps=llama.cpp&sort=trending
• License: MIT

vLLM — High-Throughput Serving

Absorbed from serving-llms-vllm.

Use vLLM when deploying production LLM APIs that need high throughput (100+ req/sec), low latency, multi-user serving, or serving large models with limited GPU memory. vLLM achieves 24x higher throughput than standard transformers through PagedAttention (block-based KV cache) and continuous batching.

Quick start

pip install vllm

Offline inference:

from vllm import LLM, SamplingParams
llm = LLM(model="meta-llama/Llama-3-8B-Instruct")
sampling = SamplingParams(temperature=0.7, max_tokens=256)
outputs = llm.generate(["Explain quantum computing"], sampling)
print(outputs[0].outputs[0].text)

OpenAI-compatible server:

vllm serve meta-llama/Llama-3-8B-Instruct

Production deployment

# Single GPU (7B-13B models)
vllm serve meta-llama/Llama-3-8B-Instruct \
  --gpu-memory-utilization 0.9 --max-model-len 8192 --port 8000

# Multi-GPU tensor parallelism (30B-70B models)
vllm serve meta-llama/Llama-2-70b-hf \
  --tensor-parallel-size 4 --quantization awq --port 8000

# With caching and Prometheus monitoring
vllm serve meta-llama/Llama-3-8B-Instruct \
  --gpu-memory-utilization 0.9 --enable-prefix-caching \
  --enable-metrics --metrics-port 9090 --host 0.0.0.0

Docker:

docker run --gpus all -p 8000:8000 vllm/vllm-openai:latest \
  --model meta-llama/Llama-3-8B-Instruct --gpu-memory-utilization 0.9

Quantized model serving

vllm serve TheBloke/Llama-2-70B-AWQ --quantization awq --gpu-memory-utilization 0.95

Hardware requirements

Supported: NVIDIA (primary), AMD ROCm, Intel GPUs, TPUs

Common issues

vLLM references

• vllm-server-deployment.md — Docker, Kubernetes, load balancing
• vllm-optimization.md — PagedAttention tuning, batching, benchmarks
• vllm-quantization.md — AWQ/GPTQ/FP8 setup, accuracy comparisons
• vllm-troubleshooting.md — Error diagnosis, performance debugging
• Official docs: https://docs.vllm.ai
• GitHub: https://github.com/vllm-project/vllm

Hugging Face Hub CLI

Absorbed from huggingface-hub.

The hf command is the CLI for Hugging Face Hub — search, download, upload models and datasets.

Full reference: huggingface-hub-cli.md

Quick reference:

hf download REPO_ID                    # Download model/dataset
hf upload REPO_ID LOCAL_PATH           # Upload files
hf datasets list                       # List datasets
hf models list                         # List models

Install: curl -LsSf https://hf.co/cli/install.sh | bash -s