ingloryon/ Ling-3.0-tiny:Q8_0

44 Downloads Updated 3 days ago

Hybrid model with 7.9B total parameters, only 1.3B activated per token

tools thinking

ollama run ingloryon/Ling-3.0-tiny:Q8_0

curl http://localhost:11434/api/chat \
  -d '{
    "model": "ingloryon/Ling-3.0-tiny:Q8_0",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='ingloryon/Ling-3.0-tiny:Q8_0',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'ingloryon/Ling-3.0-tiny:Q8_0',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)

Details

Updated 3 days ago

3 days ago

40519ff4ef37 · 8.4GB

model

archbailingmoe3

·

parameters7.89B

·

quantizationQ8_0

8.4GB

template

{{ .Prompt }}

13B

Readme

Model Overview

Ling-3.0-tiny inherits the hybrid linear attention architecture of Ling-3.0 series, while being specifically optimized for lightweight and accessible deployment. The model has 7.9B total parameters, with only 1.3B parameters activated per token.

The architecture of Ling-3.0-tiny is designed to make computational efficiency serve real-world agentic performance.

A 3:1 KDA–MLA architecture (3 KDA layers and 1 MLA layer per 4-layer block) provides more efficient long-context processing;
A sparse MoE FFN with 128 experts activates 8 routed experts and 1 shared expert per token, enabling broad model capabilities with only 1.3B activated parameters per token.
Native hybrid reasoning enables fast responses for routine tasks and multi-step reasoning for complex tasks within a single model.

Overall, these designs deliver the inference efficiency needed to deploy lightweight models in real-world agentic workflows.

Model developed by inclusionAI. For the full model card, see inclusionAI/Ling-3.0-tiny.

To pull the original weights directly, run:

ollama pull hf.co/inclusionAI/Ling-3.0-tiny-GGUF:Q8_0

# Model Overview
Ling-3.0-tiny inherits the hybrid linear attention architecture of Ling-3.0 series, while being specifically optimized for lightweight and accessible deployment. The model has 7.9B total parameters, with only 1.3B parameters activated per token.

The architecture of Ling-3.0-tiny is designed to make computational efficiency serve real-world agentic performance.

+ A 3:1 KDA–MLA architecture (3 KDA layers and 1 MLA layer per 4-layer block) provides more efficient long-context processing;
+ A sparse MoE FFN with 128 experts activates 8 routed experts and 1 shared expert per token, enabling broad model capabilities with only 1.3B activated parameters per token.
+ Native hybrid reasoning enables fast responses for routine tasks and multi-step reasoning for complex tasks within a single model.

Overall, these designs deliver the inference efficiency needed to deploy lightweight models in real-world agentic workflows.

![](https://cdn-uploads.huggingface.co/production/uploads/675fc648a63fff7b5be452f6/mifa16S1bZajl5mX90s_O.png)

---

Model developed by **inclusionAI**. For the full model card, see [inclusionAI/Ling-3.0-tiny](https://huggingface.co/inclusionAI/Ling-3.0-tiny).

To pull the original weights directly, run:

```bash
ollama pull hf.co/inclusionAI/Ling-3.0-tiny-GGUF:Q8_0
```

Paste, drop or click to upload images (.png, .jpeg, .jpg, .svg, .gif)