You can deploy this model and start earning money today!

sajjadiba/
Qwen3-0.6B-Chat-SFT-ultrachat3k-DPO-argilla6k

Name: Qwen3-0.6B-Chat-SFT-ultrachat3k-DPO-argilla6k
Availability: InStock

Views

No views yet

Qwen3-0.6B-Chat SFT + DPO Model

Fine-tuned from Qwen/Qwen3-0.6B

WARNING

This model is a research artifact and lacks safety guardrails. It may produce harmful, dangerous, or inappropriate content. Do not use in production or with users without implementing proper safety measures.

Safety Recommendations

Add content filtering
Use with supervised moderation
Not suitable for public deployment

Training

SFT: 3000 samples from ultrachat_200k
DPO: 6000 samples from argilla/distilabel-intel-orca-dpo-pairs

Usage

python

from transformers import AutoModelForCausalLM, AutoTokenizer

# Load model directly (no base model needed)
model = AutoModelForCausalLM.from_pretrained(
    'sajjadiba/Qwen3-0.6B-Chat-SFT-ultrachat3k-DPO-argilla6k',
    device_map="auto",
    dtype=torch.float16)
tokenizer = AutoTokenizer.from_pretrained('Qwen/Qwen3-0.6B')
# Set pad_token to eos_token to fix the warning
tokenizer.pad_token = tokenizer.eos_token

messages = [{"role": "user", "content": "What is AI?"}]

text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    enable_thinking=False,  # This disables the <think> tags!
    add_generation_prompt=True
)

# Tokenize
inputs = tokenizer(text, return_tensors="pt").to(model.device)

# Generate
outputs = model.generate(
    input_ids=inputs.input_ids,
    attention_mask=inputs.attention_mask,
    max_new_tokens=512,
    do_sample=True,
    temperature=0.7
)

response = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(response)