---
license: apache-2.0
base_model: Qwen/Qwen3-14B
pipeline_tag: text-generation
tags:
  - qwen3
  - trillim
  - quantized
  - custom-inference
  - text-generation
---
# Qwen3-14B-TRNQ

`Qwen3-14B-TRNQ` is a Trillim-packaged, requantized build of [`Qwen/Qwen3-14B`](https://huggingface.co/Qwen/Qwen3-14B), intended for efficient text generation with the Trillim inference engine.

## Overview

- Based on Qwen’s Qwen3-14B release
- Requantized and packaged for Trillim runtime
- Includes tokenizer, chat template, and Trillim-specific runtime artifacts
- Distributed under the upstream Apache 2.0 license with attribution preserved

## Model Details

| Item | Value |
|---|---|
| Architecture | Qwen3-14B dense |
| Parameters | 14B class |
| Source model | [`Qwen/Qwen3-14B`](https://huggingface.co/Qwen/Qwen3-14B) |
| Packaging | Trillim requantized bundle |
| License | Apache 2.0 |

## Usage

```bash
pip install trillim
trillim pull Trillim/Qwen3-14B-TRNQ
trillim chat Trillim/Qwen3-14B-TRNQ
```

This launches an interactive CLI chat session.

## Repository Contents

| File | Description |
|---|---|
| `qmodel.tensors` | Quantized weights in Trillim format |
| `rope.cache` | Precomputed RoPE cache for runtime |
| `config.json` | Model configuration |
| `generation_config.json` | Generation defaults |
| `trillim_config.json` | Trillim runtime metadata |
| `tokenizer.json` | Tokenizer data |
| `tokenizer_config.json` | Tokenizer configuration |
| `vocab.json`, `merges.txt` | Tokenizer assets |

## Provenance

This repository is derived from:
- [`Qwen/Qwen3-14B` by Alibaba Cloud](https://huggingface.co/Qwen/Qwen3-14B)

Changes made by Trillim:
- Requantized from the upstream release
- Repackaged for the Trillim inference engine

This repository is not affiliated with or endorsed by Alibaba Cloud.

## License

Released under the Apache 2.0 License, consistent with the upstream model license.

See:
- `LICENSE`