Instructions to use sengtha/Qwen3-0.6B-khm-ft3 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use sengtha/Qwen3-0.6B-khm-ft3 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="sengtha/Qwen3-0.6B-khm-ft3")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("sengtha/Qwen3-0.6B-khm-ft3")
model = AutoModelForCausalLM.from_pretrained("sengtha/Qwen3-0.6B-khm-ft3", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use sengtha/Qwen3-0.6B-khm-ft3 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "sengtha/Qwen3-0.6B-khm-ft3"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "sengtha/Qwen3-0.6B-khm-ft3",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/sengtha/Qwen3-0.6B-khm-ft3

SGLang

How to use sengtha/Qwen3-0.6B-khm-ft3 with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "sengtha/Qwen3-0.6B-khm-ft3" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "sengtha/Qwen3-0.6B-khm-ft3",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "sengtha/Qwen3-0.6B-khm-ft3" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "sengtha/Qwen3-0.6B-khm-ft3",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use sengtha/Qwen3-0.6B-khm-ft3 with Docker Model Runner:
```
docker model run hf.co/sengtha/Qwen3-0.6B-khm-ft3
```

Qwen3-0.6B-khm-ft3 — a small Khmer model for on-device RAG

A 0.6B Khmer-focused language model fine-tuned from Qwen3-0.6B, built to run on a phone, offline — it's the answering model behind iAny and runs on a 2019 Galaxy S10.

The key trick: the base Qwen3 vocabulary (~150k tokens) is trimmed to ~32k Khmer-focused tokens, which shrinks the output logits buffer enough to fit weak devices while keeping Khmer coverage. It is then continued-pretrained and instruction-tuned on Khmer.

Lineage

Stage	What	Data
Base	Qwen3-0.6B	—
Vocab trim	~32k Khmer-focused vocabulary (smaller logits → fits low-RAM phones)	via alphaedge-ai
CPT	Continued pre-training for Khmer fluency	FineWeb-2 (Khmer) + ParaCrawl
SFT (ft3)	Instruction / Q&A tuning for correct, fuller answers	`sengtha/khmer-qa`

ft3 is the current release — retrained on a richer khmer-qa for fuller answers than the earlier ft/ft2 checkpoints.

Intended use

Grounded Khmer question-answering / RAG on-device: the app retrieves context from the user's own documents and this model writes the answer in Khmer. Small and Khmer-first — not a general-purpose assistant, and it can be wrong or hallucinate, especially without retrieved context.

Usage (transformers)

from transformers import AutoModelForCausalLM, AutoTokenizer

m = "sengtha/Qwen3-0.6B-khm-ft3"
tok = AutoTokenizer.from_pretrained(m)
model = AutoModelForCausalLM.from_pretrained(m)

msgs = [{"role": "user", "content": "តើភ្នំពេញជាអ្វី?"}]
inputs = tok.apply_chat_template(msgs, add_generation_prompt=True, return_tensors="pt")
out = model.generate(inputs, max_new_tokens=256)
print(tok.decode(out[0][inputs.shape[1]:], skip_special_tokens=True))

Uses the standard Qwen3 chat template (ChatML). Append /no_think to a message to skip the reasoning block for faster, direct answers.

On-device (GGUF)

Quantized GGUF builds (Q4_K_M + Q8_0) for llama.cpp / llama.rn live in sengtha/Qwen3-0.6B-khm-ft3-Q8_0-GGUF. Q4 (~0.4 GB) fits weak phones; Q8 is higher quality.

License & attribution

Apache-2.0, inherited from the base Qwen3-0.6B (© Alibaba, Apache-2.0). 32k-vocab base via alphaedge-ai. Training corpora: FineWeb-2, ParaCrawl, and sengtha/khmer-qa. Please keep this attribution in derivatives.

Downloads last month: 330

Safetensors

Model size

0.5B params

Tensor type

BF16

Model tree for sengtha/Qwen3-0.6B-khm-ft3

Base model

Qwen/Qwen3-0.6B-Base

Finetuned

Qwen/Qwen3-0.6B

Finetuned

(1070)

this model

sengtha
/

Qwen3-0.6B-khm-ft3