Model Card for 522H0134-NguyenNhatHuy/Sailor-1.8B-Chat-SFT

This is a supervised fine-tuned version of sail/Sailor-1.8B-Chat using LoRA and PEFT, tailored for Vietnamese open-domain conversation.

Model Details

Model Description

• Model type: Causal Language Model (Chat-style)
• Language(s): Vietnamese
• License: Apache 2.0
• Fine-tuned from: sail/Sailor-1.8B-Chat

This model is designed for natural and fluent Vietnamese conversation, with improved instruction-following capabilities through supervised fine-tuning.

Uses

Direct Use

• Open-domain chat in Vietnamese
• Educational and general knowledge Q&A
• Vietnamese instruction-following tasks

Out-of-Scope Use

• Use in high-risk applications (e.g. medical, legal, or financial advice)
• English-centric tasks (this model is tuned for Vietnamese)

Bias, Risks, and Limitations

This model may reflect biases present in its training data and base model. Outputs may be inaccurate or inappropriate in some contexts.

Recommendations

Always review model outputs before deployment in user-facing applications. Do not rely on the model for critical decisions.

How to Get Started with the Model

from transformers import AutoTokenizer, AutoModelForCausalLM
from peft import PeftModel

tokenizer = AutoTokenizer.from_pretrained("sail/Sailor-1.8B-Chat")
base_model = AutoModelForCausalLM.from_pretrained("sail/Sailor-1.8B-Chat")
model = PeftModel.from_pretrained(base_model, "522H0134-NguyenNhatHuy/Sailor-1.8B-Chat-SFT")

inputs = tokenizer("Xin chào, bạn khỏe không?", return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=100)
print(tokenizer.decode(outputs[0]))