--- title: "Using Noema Overfit" description: "Run compatible mixture-of-experts models with a resident core, locally paged expert weights, and device-specific performance guidance." version: "Noema 3.6+" platforms: ["iPhone", "iPad", "Mac", "Vision Pro"] reviewed: "August 24, 2026" canonical: "https://noemaai.com/docs/noema-overfit" --- # Using Noema Overfit Run compatible Qwen, Gemma, Laguna, and DeepSeek mixture-of-experts models using a resident core and locally paged expert weights, then measure whether the model is practical on your device. ## How Overfit works Noema Overfit lets compatible mixture-of-experts models run even when their complete weights cannot fit in device memory. Overfit is not model training or “overfitting.” It reorganizes a compatible GGUF model into a complete `.noema-paged` package containing: - `resident.gguf`, with the weights required throughout inference. - One or more `experts-*.bin` files containing routed expert weights. - `manifest.json`, describing the package structure, model geometry, file hashes, expert records, and integrity information. During inference, Noema keeps the resident core and a device-sized expert bank in unified memory. When the model selects an expert that is not already present, Noema reads it from local storage, verifies it, and loads it into the bank. > Overfit provides capacity, not guaranteed speed. Storage is not RAM, and a smaller model that fits normally will usually start faster, generate faster, and consume less energy. ## Download and run an Overfit model 1. Open **Explore** and select the **GGUF** catalog. 2. Open the **Noema Overfit** paged-model collection. 3. Select a complete paged package and review its storage size, resident memory, expert-bank estimate, context limit, and estimated working set. 4. Download the package. Noema treats the manifest, resident GGUF, and expert files as one model. 5. Open the model in **Stored**, then open its settings. 6. Leave **Overfit (Paged Experts)** set to **Automatic**. 7. Run the **Canary Test** to measure the model on this device. 8. Load the model and begin a chat. > Overfit packages are models, not ordinary quantization choices. Never download only `resident.gguf`; it cannot represent the complete model without the manifest and expert files. ## Create a package on Mac Noema for macOS can convert an existing compatible MoE GGUF into a complete paged package. 1. Download and register the source GGUF. 2. In **Stored**, right-click the model. 3. Choose **Create Paged Package…** 4. Wait while Noema extracts and verifies the expert data. 5. Choose **Add to Stored** when conversion finishes. The result is a sibling `.noema-paged` directory. Conversion does not retrain, prune, or replace the model’s weights. > The original GGUF remains available, so enough storage is required for both copies during conversion. ## The Canary Test The Canary Test performs a short, repeatable local run for this package and device. - Validates a sample of the package. - Measures cached and uncached storage reads. - Loads the paged runtime. - Runs a fixed 64-token completion. - Records time to first token and generation speed. - Measures latency percentiles and long stalls. - Records expert-bank hit rate, misses per token, peak memory, and thermal state. The result belongs to the exact package fingerprint, device, storage volume, native runtime contract, and app build. Available memory is evaluated separately at launch. ## Understand the Canary result - **Paged — interactive** - **Paged — slow** - **Too slow on this device** - **Use Constellation** - **Not supported for this model** > The Canary Test is recommended rather than mandatory. Performance can change with free memory, thermal conditions, storage activity, prompt length, and model settings. ## Current compatible models As of August 24, 2026, the public [Noema Overfit repository](https://huggingface.co/NoemaAI-labs/Noema-Overfit) contains these complete packages. | Model package | Approximate download | Resident core | Expert payloads | | --- | --- | --- | --- | | DeepSeek V4 Flash 0731, UD-IQ4_NL | 136.66 GB | 7.81 GB | 128.85 GB | | Gemma 4 26B-A4B QAT, UD-Q4_K_XL | 14.36 GB | 1.40 GB | 12.96 GB | | Qwen 3.6 35B-A3B, UD-Q4_K_M | 22.15 GB | 2.57 GB | 19.57 GB | | Qwen 3.5 122B-A10B, Q4_K_M | 74.27 GB | 4.00 GB | 70.26 GB | The runtime whitelist currently covers `qwen3moe`, `qwen35moe`, `gemma4`, `laguna`, and `deepseek4`. This includes the published DeepSeek V4 Flash, Qwen 3.5, Qwen 3.6, and Gemma 4 packages. Poolside's [Laguna S 2.1 GGUF](https://huggingface.co/poolside/Laguna-S-2.1-GGUF) and [Laguna XS 2.1 GGUF](https://huggingface.co/poolside/Laguna-XS-2.1-GGUF) can be converted into complete paged packages on macOS; the public Noema collection does not currently list pre-built Laguna packages. DeepSeek V4 Flash uses the dedicated `deepseek4` runtime and native contract v4. Treat it as a model-specific package, not as a conventional Qwen or Gemma package. Dense models and arbitrary GGUF architectures are not supported. ## Privacy and safety After the model is downloaded, Overfit inference can remain entirely local. - Model weights remain on local storage. - Prompts and responses remain on the device. - No cloud inference provider is required. - Expert payloads are excluded from device backups. - Matching conversation-prefix states can be saved locally for faster follow-up launches. - Using a model from an awake Mac through Constellation remains a separate, optional choice. Packages are checked for valid geometry, safe paths, file sizes, record bounds, coverage, fingerprints, and checksums. The native runtime validates the package independently at launch. Under memory pressure, Noema can reduce prefetching, discard optional queued work, or stop generation before an unsafe memory condition. ## Current limitations - Overfit is restricted to compatible MoE architectures. - Package creation is currently macOS-only. - Paged launches currently cap context at 8,192 tokens on Mac and 4,096 tokens on iPhone, iPad, and Vision Pro. - Multimodal projectors and some utility workflows do not use the paged path yet. - Low Power Mode disables expert prefetching to favor energy use. - Prompt-state restoration only works when the package and relevant runtime configuration still match. - **Force (Experimental)** may bypass performance recommendations, but it never bypasses package-integrity validation. ## Related documentation - [Downloading Models](https://noemaai.com/docs/downloading-models) - [Running Models Locally](https://noemaai.com/docs/running-llms) - [Model Settings](https://noemaai.com/docs/model-settings)