---
title: Qwen3.8-Distilled-9B-NPU2
canonical_url: "https://www.modelscope.ai/models/Atomic-Germ/Qwen3.8-Distilled-9B-NPU2"
md_url: "https://www.modelscope.ai/models/Atomic-Germ/Qwen3.8-Distilled-9B-NPU2.md"
repository: Atomic-Germ/Qwen3.8-Distilled-9B-NPU2
last_updated: 2026-08-29
license: apache-2.0
pipeline_tag: image-text-to-text
tasks:
  - image-text-to-text
model_type:
  - qwen3_5
architectures:
  - Qwen3_5ForConditionalGeneration
base_model:
  - empero-ai/Qwen3.8-9B-Distill
base_model_relation: quantized
downloads: 0
stars: 0
tags:
  - empero-ai
  - qwen3.5
  - qwen3.8
  - distillation
  - reasoning
  - function-calling
  - vision
  - qwen3.8-froggeric-v22.4
  - npu2
  - fastflowlm
  - flm
  - q4nx
  - v1.0.1
---

# Qwen3.8-Distilled-9B-NPU2

> Qwen3.8-Distilled-9B-NPU2 - An open-source model by Atomic-Germ on ModelScope. qwen3.8-distilled-9b

Atomic-Germ/Qwen3.8-Distilled-9B-NPU2 is a image-text-to-text model on ModelScope. licensed under apache-2.0. derived from empero-ai/Qwen3.8-9B-Distill.

- **Repository**: Atomic-Germ/Qwen3.8-Distilled-9B-NPU2
- **License**: apache-2.0
- **Tasks**: image-text-to-text
- **Base model**: empero-ai/Qwen3.8-9B-Distill
- **Tags**: empero-ai, qwen3.5, qwen3.8, distillation, reasoning, function-calling, vision, qwen3.8-froggeric-v22.4, npu2, fastflowlm, flm, q4nx, v1.0.1
- **Downloads**: 0
- **Stars**: 0
- **Last updated**: 2026-08-29

Source: https://www.modelscope.ai/models/Atomic-Germ/Qwen3.8-Distilled-9B-NPU2

---

# Qwen3.8-Distilled-9B-NPU2

**FastFlowLM Q4NX conversion of `Qwen3.8-9B-Distill/`** for AMD XDNA NPU inference.

This repository contains a quantized **Q4NX** port of the model, compiled for the FastFlowLM (FLM) runtime. It is **not** a GGUF file.

| Item | Value |
|------|-------|
| Source model | `empero-ai/Qwen3.8-9B-Distill` |
| Weights | `model.q4nx` (7.11 GB) |
| Modality | language / vision |
| FLM version | `1.0.1` |
| Converted | 2026-08-26 |

## Install and run

This repository works with `flm-add`, a small installer that copies the model
into the FastFlowLM user directory and registers the tag. It never
modifies the system FastFlowLM install.

`pip install flm-add` or `uv tool install flm-add`

```bash
uv tool install flm-add
flm-add Atomic-Germ/Qwen3.8-Distilled-9B-NPU2 --family qwen3.5
FLM_CONFIG_PATH="$HOME/.config/flm/model_list.json" FLM_XCLBIN_PATH="$HOME/.config/flm" flm run qwen3.8-distilled:9b
```

## Files

| File | Description |
|------|-------------|
| `model.q4nx` | Quantized weights (Q8_0 / Q4_1 / BF16) |
| `config.json` | FLM runtime configuration |
| `tokenizer.json` | Tokenizer vocabulary |
| `tokenizer_config.json` | Tokenizer configuration |
| `chat_template.jinja` | Chat template |
| `vision_weight.q4nx` | Vision model |
| `results/` | FLM-Test results |


---

## Source model card

See the original model card: `Qwen3.8-9B-Distill/` on Hugging Face