Instructions to use NANI-Nithin/Ling-3.0-tiny-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use NANI-Nithin/Ling-3.0-tiny-GGUF with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf NANI-Nithin/Ling-3.0-tiny-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf NANI-Nithin/Ling-3.0-tiny-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf NANI-Nithin/Ling-3.0-tiny-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf NANI-Nithin/Ling-3.0-tiny-GGUF:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf NANI-Nithin/Ling-3.0-tiny-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf NANI-Nithin/Ling-3.0-tiny-GGUF:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf NANI-Nithin/Ling-3.0-tiny-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf NANI-Nithin/Ling-3.0-tiny-GGUF:Q4_K_M

Use Docker

docker model run hf.co/NANI-Nithin/Ling-3.0-tiny-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use NANI-Nithin/Ling-3.0-tiny-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "NANI-Nithin/Ling-3.0-tiny-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "NANI-Nithin/Ling-3.0-tiny-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/NANI-Nithin/Ling-3.0-tiny-GGUF:Q4_K_M

Ollama
How to use NANI-Nithin/Ling-3.0-tiny-GGUF with Ollama:
```
ollama run hf.co/NANI-Nithin/Ling-3.0-tiny-GGUF:Q4_K_M
```
Unsloth Desktop

How to use NANI-Nithin/Ling-3.0-tiny-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf NANI-Nithin/Ling-3.0-tiny-GGUF:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "NANI-Nithin/Ling-3.0-tiny-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use NANI-Nithin/Ling-3.0-tiny-GGUF with Docker Model Runner:
```
docker model run hf.co/NANI-Nithin/Ling-3.0-tiny-GGUF:Q4_K_M
```

Lemonade

How to use NANI-Nithin/Ling-3.0-tiny-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull NANI-Nithin/Ling-3.0-tiny-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Ling-3.0-tiny-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use NANI-Nithin/Ling-3.0-tiny-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf NANI-Nithin/Ling-3.0-tiny-GGUF:Q4_K_M

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default NANI-Nithin/Ling-3.0-tiny-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use NANI-Nithin/Ling-3.0-tiny-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf NANI-Nithin/Ling-3.0-tiny-GGUF:Q4_K_M

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "NANI-Nithin/Ling-3.0-tiny-GGUF:Q4_K_M" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

NANI-Nithin commited on 7 days ago

Commit

4b5a1b1

verified ·

1 Parent(s): 4b91c6e

Upload README.md with huggingface_hub

Browse files

Files changed (1) hide show

README.md +66 -116

README.md CHANGED Viewed

@@ -1,134 +1,84 @@
 ---
-language:
-- en
-license: apache-2.0
 base_model: inclusionAI/Ling-3.0-tiny
-tags:
-- gguf
-- llama.cpp
-- quantized
-- moe
-- bailingmoev3
-- hybrid-model
-- local-llm
-- text-generation
 pipeline_tag: text-generation
 ---
-# Ling-3.0-tiny-GGUF
-GGUF quantizations of [inclusionAI/Ling-3.0-tiny](https://huggingface.co/inclusionAI/Ling-3.0-tiny), converted for use with compatible `llama.cpp`-based runtimes.
-This repository includes a complete selection of standard K-quants and importance-matrix (IQ) quantizations, so you can choose the best balance of model size, speed, and output quality for your hardware.
-> **Runtime compatibility:** Ling-3.0-tiny uses the BailingMoeV3 / hybrid architecture. Use a runtime with explicit support for this architecture. Generic or older `llama.cpp` builds may not load these files correctly.
-## Available files
-| Quantization | Best for |
-|---|---|
-| `F16` | Highest-fidelity baseline; re-quantization and high-memory systems |
-| `Q8_0` | Near-F16 quality with substantially lower memory use |
-| `Q6_K` | High-quality local inference |
-| `Q5_K_M` | Strong quality-to-size balance |
-| `Q5_K_S` | Slightly smaller alternative to Q5_K_M |
-| `Q5_0` | Legacy-style 5-bit option |
-| `Q4_K_M` | Recommended default for most users |
-| `Q4_K_S` | Smaller Q4 K-quant alternative |
-| `Q4_0` | Compact legacy-style 4-bit option |
-| `IQ4_NL` | High-quality importance-matrix 4-bit option |
-| `IQ4_XS` | Compact importance-matrix 4-bit option |
-| `Q3_K_L` | Higher-quality 3-bit K-quant |
-| `Q3_K_M` | Balanced 3-bit K-quant |
-| `Q3_K_S` | Smaller 3-bit K-quant |
-| `IQ3_M` | Strong quality-per-GB option for constrained systems |
-| `IQ3_S` | Smaller 3-bit IQ option |
-| `IQ3_XS` | Very compact IQ 3-bit option |
-| `IQ3_XXS` | Extremely compact IQ 3-bit option |
-| `Q2_K` | Low-memory K-quant option |
-| `IQ2_M` | Compact IQ quant with better quality potential than very-low-bit options |
-| `IQ2_S` | Low-memory IQ option |
-| `IQ2_XS` | Very small IQ option |
-| `IQ2_XXS` | Extremely small IQ option |
-| `IQ1_M` | Experimental ultra-low-memory option |
-| `IQ1_S` | Smallest experimental option |
-## Recommended downloads
-| Your priority | Recommended file |
-|---|---|
-| Best quality | `Ling-3.0-tiny-F16.gguf` |
-| Near-original quality | `Ling-3.0-tiny-Q8_0.gguf` |
-| High quality with lower memory use | `Ling-3.0-tiny-Q6_K.gguf` |
-| Best general-purpose choice | `Ling-3.0-tiny-Q4_K_M.gguf` |
-| Small but capable | `Ling-3.0-tiny-IQ3_M.gguf` |
-| Tight VRAM / RAM budget | `Ling-3.0-tiny-IQ2_M.gguf` |
-| Experimental minimum size | `Ling-3.0-tiny-IQ1_S.gguf` |
-For most users, start with **Q4_K_M**. If you have more RAM or VRAM, try **Q5_K_M**, **Q6_K**, or **Q8_0**. IQ quants can offer attractive quality-to-size trade-offs, but results and compatibility may vary by runtime and hardware.
 ## Usage
-Download one `.gguf` file, then run it with a compatible build of `llama.cpp`.
 ```bash
-llama-cli \
-  -m Ling-3.0-tiny-Q4_K_M.gguf \
-  -ngl 99 \
-  -c 4096 \
-  -p "Write a concise explanation of retrieval-augmented generation."
 ```
-`-ngl 99` attempts to offload all supported layers to the GPU. Remove it or set `-ngl 0` for CPU-only inference.
-## Important notes
-- These files are quantized derivatives of the original model; output quality changes depending on the chosen quantization.
-- Very low-bit quants, especially IQ1 and IQ2 variants, are intended for memory-constrained or experimental use and may noticeably reduce output quality.
-- Use the original model’s license, terms, and usage requirements.
-- Validate the selected quantization on your own workload before production use.
-## Conversion details
-- Base model: [`inclusionAI/Ling-3.0-tiny`](https://huggingface.co/inclusionAI/Ling-3.0-tiny)
-- Format: GGUF
-- Conversion/runtime branch: BailingMoeV3-compatible `llama.cpp` fork
-- Standard K-quants: generated from the F16 GGUF
-- IQ quants: generated using an importance matrix calibrated on a text corpus
-## Credits
-- Original model by [inclusionAI](https://huggingface.co/inclusionAI)
-- GGUF conversion and quantization by [NANI-Nithin](https://huggingface.co/NANI-Nithin)
-- GGUF tooling by the [llama.cpp](https://github.com/ggml-org/llama.cpp) community
-## Disclaimer
-This is a community GGUF conversion and is not an official release by inclusionAI. Please report conversion, loading, or compatibility issues in this repository’s Discussions section.
-## Reproducibility
-This repository was generated with a BailingMoeV3-enabled llama.cpp fork.
-The exact source checkout checkpoint is recorded below:
-```json
-{
-  "stage": "01_checkout_bailing_llama",
-  "status": "complete",
-  "timestamp_utc": "2026-08-11T10:36:13.114546+00:00",
-  "model": "inclusionAI/Ling-3.0-tiny",
-  "llama_repo": "https://github.com/aetherbird/llama.cpp.git",
-  "llama_branch": "bailingmoe3-support",
-  "repo_dir": "/mnt/ling/src/llama.cpp",
-  "commit": "3a0124fa8c20356ed5e6bf0c0ebae1566d6f49c1"
-}
 ```
-## Files
-- `F16`: Conversion baseline.
-- `Q4_K_M`: General local-inference default.
-- `Q5_K_M`, `Q6_K`, `Q8_0`: Higher-fidelity variants.
-- `IQ*`: Importance-matrix variants, generated only when supported by the pinned quantizer.
-Use a Ling/BailingMoeV3-compatible runtime to load these files.

 ---
 base_model: inclusionAI/Ling-3.0-tiny
+library_name: gguf
 pipeline_tag: text-generation
+tags:
+  - gguf
+  - llama.cpp
+  - quantized
+  - imatrix
 ---
+# Ling-3.0-tiny GGUF
+GGUF quantizations of [inclusionAI/Ling-3.0-tiny](https://huggingface.co/inclusionAI/Ling-3.0-tiny), covering 32 files (139.3 GB total).
+## Files
+| File | Quant | Size | Notes |
+|---|---|---:|---|
+| [Ling-3.0-tiny-BF16.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-BF16.gguf) | `BF16` | 14.72 GB | Full precision source. Every quant below is cut from this file. |
+| [Ling-3.0-tiny-F16.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-F16.gguf) | `F16` | 14.72 GB | Full precision source. |
+| [Ling-3.0-tiny-Q8_0.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-Q8_0.gguf) | `Q8_0` | 7.83 GB | Effectively lossless. Use when disk and RAM are not the constraint. |
+| [Ling-3.0-tiny-Q6_K.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-Q6_K.gguf) | `Q6_K` | 6.05 GB | Near-lossless; the last stop before quality becomes measurable. |
+| [Ling-3.0-tiny-Q5_K_M.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-Q5_K_M.gguf) | `Q5_K_M` | 5.25 GB | Very good quality, noticeably smaller than Q6_K. |
+| [Ling-3.0-tiny-Q5_K_S.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-Q5_K_S.gguf) | `Q5_K_S` | 5.11 GB | Slightly smaller than Q5_K_M for a slight quality cost. |
+| [Ling-3.0-tiny-Q5_1.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-Q5_1.gguf) | `Q5_1` | 5.55 GB | Legacy. Prefer Q5_K_M. |
+| [Ling-3.0-tiny-Q5_0.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-Q5_0.gguf) | `Q5_0` | 5.11 GB | Legacy. Prefer Q5_K_M. |
+| [Ling-3.0-tiny-Q4_K_M.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-Q4_K_M.gguf) | `Q4_K_M` | 4.49 GB | The usual default. Best quality-per-byte for most people. |
+| [Ling-3.0-tiny-Q4_K_S.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-Q4_K_S.gguf) | `Q4_K_S` | 4.24 GB | A little smaller than Q4_K_M, a little worse. |
+| [Ling-3.0-tiny-IQ4_NL.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-IQ4_NL.gguf) | `IQ4_NL` | 4.22 GB | Non-linear 4-bit; good on hardware without fast K-quant kernels. |
+| [Ling-3.0-tiny-IQ4_XS.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-IQ4_XS.gguf) | `IQ4_XS` | 3.99 GB | Best sub-4.5bpw option; usually beats Q4_K_S at a smaller size. |
+| [Ling-3.0-tiny-Q4_1.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-Q4_1.gguf) | `Q4_1` | 4.66 GB | Legacy. Prefer Q4_K_M. |
+| [Ling-3.0-tiny-Q4_0.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-Q4_0.gguf) | `Q4_0` | 4.22 GB | Legacy round-to-nearest. Prefer Q4_K_M unless a runtime needs this. |
+| [Ling-3.0-tiny-MXFP4_MOE.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-MXFP4_MOE.gguf) | `MXFP4_MOE` | 4.39 GB | MoE-only 4-bit microscaling format for the expert tensors. |
+| [Ling-3.0-tiny-Q3_K_L.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-Q3_K_L.gguf) | `Q3_K_L` | 3.86 GB | Small, with real quality loss. Usable when RAM is tight. |
+| [Ling-3.0-tiny-Q3_K_M.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-Q3_K_M.gguf) | `Q3_K_M` | 3.58 GB | Smaller again; noticeable degradation. |
+| [Ling-3.0-tiny-IQ3_M.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-IQ3_M.gguf) | `IQ3_M` | 3.31 GB | Strong at ~3.7bpw, clearly better than Q3_K_M. |
+| [Ling-3.0-tiny-IQ3_S.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-IQ3_S.gguf) | `IQ3_S` | 3.27 GB | Slightly smaller than IQ3_M. |
+| [Ling-3.0-tiny-Q3_K_S.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-Q3_K_S.gguf) | `Q3_K_S` | 3.27 GB | Aggressive. Prefer IQ3_M at a similar size. |
+| [Ling-3.0-tiny-IQ3_XS.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-IQ3_XS.gguf) | `IQ3_XS` | 3.11 GB | Aggressive but coherent. |
+| [Ling-3.0-tiny-IQ3_XXS.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-IQ3_XXS.gguf) | `IQ3_XXS` | 2.91 GB | Very aggressive; imatrix carries it. |
+| [Ling-3.0-tiny-Q2_K.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-Q2_K.gguf) | `Q2_K` | 2.78 GB | Very small, heavily degraded. For experimentation. |
+| [Ling-3.0-tiny-IQ2_M.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-IQ2_M.gguf) | `IQ2_M` | 2.52 GB | The smallest size most people find usable. |
+| [Ling-3.0-tiny-Q2_K_S.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-Q2_K_S.gguf) | `Q2_K_S` | 2.59 GB | Smaller than Q2_K, requires the imatrix. |
+| [Ling-3.0-tiny-IQ2_S.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-IQ2_S.gguf) | `IQ2_S` | 2.31 GB | Below the usual usability line. |
+| [Ling-3.0-tiny-IQ2_XS.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-IQ2_XS.gguf) | `IQ2_XS` | 2.27 GB | Experimental. |
+| [Ling-3.0-tiny-IQ2_XXS.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-IQ2_XXS.gguf) | `IQ2_XXS` | 2.06 GB | Experimental. |
+| [Ling-3.0-tiny-Q2_0.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-Q2_0.gguf) | `Q2_0` | 2.28 GB | Extreme, group-64. Included for completeness. |
+| [Ling-3.0-tiny-IQ1_M.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-IQ1_M.gguf) | `IQ1_M` | 1.80 GB | Extreme. Expect substantial degradation. |
+| [Ling-3.0-tiny-IQ1_S.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-IQ1_S.gguf) | `IQ1_S` | 1.64 GB | Extreme. Expect substantial degradation. |
+| [Ling-3.0-tiny-Q1_0.gguf](https://huggingface.co/NANI-Nithin/Ling-3.0-tiny-GGUF/blob/main/Ling-3.0-tiny-Q1_0.gguf) | `Q1_0` | 1.21 GB | Extreme. Included for completeness. |
+## Which one should I download?
+Pick the largest file that leaves a couple of gigabytes of headroom on the device you will run it on — the model has to fit in RAM (or VRAM, if you are offloading) alongside the KV cache and the OS.
+- Plenty of memory: **Q6_K** or **Q8_0**.
+- The usual choice: **Q4_K_M**.
+- Tight on memory: **IQ4_XS**, then **IQ3_M**, then **IQ2_M**.
+- The `IQ*` files are imatrix-guided and generally beat a `Q*` file of similar size, at the cost of slightly slower inference on some hardware.
+## Quantization details
+- Importance matrix computed with `llama-imatrix` over 500 rows of [Salesforce/wikitext](https://huggingface.co/datasets/Salesforce/wikitext) (`wikitext-2-raw-v1`).
+- The matrix was computed on the **BF16** weights.
+- K-quants below 6 bit and the whole `IQ` set are imatrix-guided. `Q4_0`/`Q4_1`/`Q5_0`/`Q5_1` are legacy round-to-nearest and ignore it; `Q6_K`/`Q8_0` are near-lossless and do not need it.
+- All files are cut from the same BF16 GGUF, so differences between them are quantization only.
 ## Usage
 ```bash
+llama-cli -hf NANI-Nithin/Ling-3.0-tiny-GGUF:Q4_K_M -p "Hello"
 ```
+Or download one file and point at it directly:
+```bash
+huggingface-cli download NANI-Nithin/Ling-3.0-tiny-GGUF Ling-3.0-tiny-Q4_K_M.gguf --local-dir .
+llama-cli -m Ling-3.0-tiny-Q4_K_M.gguf -p "Hello"
 ```
+---
+Quantized with [llama.cpp](https://github.com/ggml-org/llama.cpp) by AgentQuantix on 2026-09-05. Licensing follows the base model.