license: other
license_name: qwen
language:
- th
- en
library_name: transformers
pipeline_tag: text-generation
tags:
- openthaigpt
- qwen
🇹🇭 OpenThaiGPT 7b 1.5 Instruct
🇹🇭 OpenThaiGPT 7b Version 1.5 is an advanced 7-billion-parameter Thai language chat model based on Qwen v2.5 released on September 30, 2024. It has been specifically fine-tuned on over 2,000,000 Thai instruction pairs and is capable of answering Thai-specific domain questions.
Online Demo:
Example code for API Calling
https://github.com/OpenThaiGPT/openthaigpt1.5_api_examples
Highlights
- State-of-the-art Thai language LLM, achieving the highest average scores across various Thai language exams compared to other open-source Thai LLMs.
- Multi-turn conversation support for extended dialogues.
- Retrieval Augmented Generation (RAG) compatibility for enhanced response generation.
- Impressive context handling: Processes up to 131,072 tokens of input and generates up to 8,192 tokens, enabling detailed and complex interactions.
Benchmark on OpenThaiGPT Eval
** Please take a look at openthaigpt/openthaigpt1.5-7b-instruct
for this model's evaluation result.
Exam names | scb10x/llama-3-typhoon-v1.5x-8b-instruct | meta-llama/Llama-3.1-7B-Instruct | Qwen/Qwen2.5-7B-Instruct_stat | openthaigpt/openthaigpt1.5-7b |
---|---|---|---|---|
01_a_level | 46.67% | 47.50% | 58.33% | 60.00% |
02_tgat | 32.00% | 36.00% | 32.00% | 36.00% |
03_tpat1 | 52.50% | 55.00% | 57.50% | 57.50% |
04_investment_consult | 56.00% | 48.00% | 68.00% | 76.00% |
05_facebook_beleble_th_200 | 78.00% | 73.00% | 79.00% | 81.00% |
06_xcopa_th_200 | 79.50% | 69.00% | 80.50% | 81.00% |
07_xnli2.0_th_200 | 56.50% | 55.00% | 53.00% | 54.50% |
08_onet_m3_thai | 48.00% | 32.00% | 72.00% | 64.00% |
09_onet_m3_social | 75.00% | 50.00% | 90.00% | 80.00% |
10_onet_m3_math | 25.00% | 18.75% | 31.25% | 31.25% |
11_onet_m3_science | 46.15% | 42.31% | 46.15% | 46.15% |
12_onet_m3_english | 70.00% | 76.67% | 86.67% | 83.33% |
13_onet_m6_thai | 47.69% | 29.23% | 46.15% | 53.85% |
14_onet_m6_math | 29.41% | 17.65% | 29.41% | 29.41% |
15_onet_m6_social | 50.91% | 43.64% | 56.36% | 58.18% |
16_onet_m6_science | 42.86% | 32.14% | 57.14% | 57.14% |
17_onet_m6_english | 65.38% | 71.15% | 78.85% | 80.77% |
Micro Average | 60.65% | 55.60% | 64.41% | 65.78% |
Thai language multiple choice exams, Test on unseen test set, Zero-shot learning. Benchmark source code and exams information: https://github.com/OpenThaiGPT/openthaigpt_eval
(Updated on: 30 September 2024)
Benchmark on scb10x/thai_exam
Models | Thai Exam (Acc) |
---|---|
api/claude-3-5-sonnet-20240620 | 69.2 |
openthaigpt/openthaigpt1.5-72b-instruct* | 64.07 |
api/gpt-4o-2024-05-13 | 63.89 |
hugging-quants/Meta-Llama-3.1-405B-Instruct-AWQ-INT4 | 63.54 |
Qwen/Qwen2-72B-Instruct | 58.23 |
meta-llama/Meta-Llama-3.1-70B-Instruct | 58.23 |
scb10x/llama-3-typhoon-v1.5x-70b-instruct | 58.76 |
Qwen/Qwen2.5-14B-Instruct | 57.35 |
api/gpt-4o-mini-2024-07-18 | 54.51 |
openthaigpt/openthaigpt1.5-7b-instruct* | 52.04 |
SeaLLMs/SeaLLMs-v3-7B-Chat | 51.33 |
openthaigpt/openthaigpt-1.0.0-70b-chat | 50.09 |
* Evaluated by OpenThaiGPT team using scb10x/thai_exam.
Licenses
- Built with Qwen
- Qwen License: Allow Research and
Commercial uses but if your user base exceeds 100 million monthly active users, you need to negotiate a separate commercial license. Please see LICENSE file for more information.
Sponsors
Supports
- Official website: https://openthaigpt.aieat.or.th
- Facebook page: https://web.facebook.com/groups/openthaigpt
- A Discord server for discussion and support here
- E-mail: kobkrit@aieat.or.th
Prompt Format
Prompt format is based on ChatML.
<|im_start|>system\n{sytem_prompt}<|im_end|>\n<|im_start|>user\n{prompt}<|im_end|>\n<|im_start|>assistant\n
System prompt:
คุณคือผู้ช่วยตอบคำถามที่ฉลาดและซื่อสัตย์
Examples
Single Turn Conversation Example
<|im_start|>system\nคุณคือผู้ช่วยตอบคำถามที่ฉลาดและซื่อสัตย์<|im_end|>\n<|im_start|>user\nสวัสดีครับ<|im_end|>\n<|im_start|>assistant\n
Single Turn Conversation with Context (RAG) Example
<|im_start|>system\nคุณคือผู้ช่วยตอบคำถามที่ฉลาดและซื่อสัตย์<|im_end|>\n<|im_start|>user\nกรุงเทพมหานคร เป็นเมืองหลวง นครและมหานครที่มีประชากรมากที่สุดของประเทศไทย กรุงเทพมหานครมีพื้นที่ทั้งหมด 1,568.737 ตร.กม. มีประชากรตามทะเบียนราษฎรกว่า 8 ล้านคน\nกรุงเทพมหานครมีพื้นที่เท่าไร่<|im_end|>\n<|im_start|>assistant\n
Multi Turn Conversation Example
First turn
<|im_start|>system\nคุณคือผู้ช่วยตอบคำถามที่ฉลาดและซื่อสัตย์<|im_end|>\n<|im_start|>user\nสวัสดีครับ<|im_end|>\n<|im_start|>assistant\n
Second turn
<|im_start|>system\nคุณคือผู้ช่วยตอบคำถามที่ฉลาดและซื่อสัตย์<|im_end|>\n<|im_start|>user\nสวัสดีครับ<|im_end|>\n<|im_start|>assistant\nสวัสดีครับ ยินดีต้อนรับครับ คุณต้องการให้ฉันช่วยอะไรครับ?<|im_end|>\n<|im_start|>user\nกรุงเทพมหานคร ชื่อเต็มยาวๆคืออะไร<|im_end|>\n<|im_start|>assistant\n
Result
<|im_start|>system\nคุณคือผู้ช่วยตอบคำถามที่ฉลาดและซื่อสัตย์<|im_end|>\n<|im_start|>user\nสวัสดีครับ<|im_end|>\n<|im_start|>assistant\nสวัสดีครับ ยินดีต้อนรับครับ คุณต้องการให้ฉันช่วยอะไรครับ?<|im_end|>\n<|im_start|>user\nกรุงเทพมหานคร ชื่อเต็มยาวๆคืออะไร<|im_end|>\n<|im_start|>assistant\nชื่อเต็มของกรุงเทพมหานครคือ \"กรุงเทพมหานคร อมรรัตนโกสินทร์ มหินทรายุธยา มหาดิลกภพ นพรัตนราชธานีบูรีรมย์ อุดมราชนิเวศน์มหาสถาน อมรพิมานอวตารสถิต สักกะทัตติยวิษณุกรรมประสิทธิ์\"
How to use
Huggingface
from transformers import AutoModelForCausalLM, AutoTokenizer
model_name = "openthaigpt/openthaigpt1.5-72b-instruct"
model = AutoModelForCausalLM.from_pretrained(
model_name,
torch_dtype="auto",
device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained(model_name)
prompt = "ประเทศไทยคืออะไร"
messages = [
{"role": "system", "content": "คุณคือผู้ช่วยตอบคำถามที่ฉลาดและซื่อสัตย์"},
{"role": "user", "content": prompt}
]
text = tokenizer.apply_chat_template(
messages,
tokenize=False,
add_generation_prompt=True
)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
generated_ids = model.generate(
**model_inputs,
max_new_tokens=512
)
generated_ids = [
output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
]
response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
vLLM
Install VLLM (https://github.com/vllm-project/vllm)
Run server
vllm serve openthaigpt/openthaigpt1.5-72b-instruct --tensor-parallel-size 4
- Note, change
--tensor-parallel-size 4
to the amount of available GPU cards.
- Run inference (CURL example)
curl -X POST 'http://127.0.0.1:8000/v1/completions' \
-H 'Content-Type: application/json' \
-d '{
"model": ".",
"prompt": "<|im_start|>system\nคุณคือผู้ช่วยตอบคำถามที่ฉลาดและซื่อสัตย์<|im_end|>\n<|im_start|>user\nสวัสดีครับ<|im_end|>\n<|im_start|>assistant\n",
"max_tokens": 512,
"temperature": 0.7,
"top_p": 0.8,
"top_k": 40,
"stop": ["<|im_end|>"]
}'
Processing Long Texts
The current config.json
is set for context length up to 32,768 tokens.
To handle extensive inputs exceeding 32,768 tokens, we utilize YaRN, a technique for enhancing model length extrapolation, ensuring optimal performance on lengthy texts.
For supported frameworks, you could add the following to config.json
to enable YaRN:
{
...
"rope_scaling": {
"factor": 4.0,
"original_max_position_embeddings": 32768,
"type": "yarn"
}
}
GPU Memory Requirements
Number of Parameters | FP 16 bits | 8 bits (Quantized) | 4 bits (Quantized) | Example Graphic Card for 4 bits |
---|---|---|---|---|
7b | 24 GB | 12 GB | 6 GB | Nvidia RTX 4060 8GB |
13b | 48 GB | 24 GB | 12 GB | Nvidia RTX 4070 16GB |
72b | 192 GB | 96 GB | 48 GB | Nvidia RTX 4090 24GB x 2 cards |
Authors
- Sumeth Yuenyong (sumeth.yue@mahidol.edu)
- Kobkrit Viriyayudhakorn (kobkrit@aieat.or.th)
- Apivadee Piyatumrong (apivadee.piy@nectec.or.th)
- Jillaphat Jaroenkantasima (autsadang41@gmail.com)
- Thaweewat Rugsujarit (thaweewr@scg.com)
- Norapat Buppodom (new@norapat.com)
- Koravich Sangkaew (kwankoravich@gmail.com)
- Peerawat Rojratchadakorn (peerawat.roj@gmail.com)
- Surapon Nonesung (nonesungsurapon@gmail.com)
- Chanon Utupon (chanon.utupon@gmail.com)
- Sadhis Wongprayoon (sadhis.tae@gmail.com)
- Nucharee Thongthungwong (nuchhub@hotmail.com)
- Chawakorn Phiantham (mondcha1507@gmail.com)
- Patteera Triamamornwooth (patt.patteera@gmail.com)
- Nattarika Juntarapaoraya (natt.juntara@gmail.com)
- Kriangkrai Saetan (kraitan.ss21@gmail.com)
- Pitikorn Khlaisamniang (pitikorn32@gmail.com)
Disclaimer: Provided responses are not guaranteed.