language:
- en
- zh
- id
- th
- vi
- ms
- lo
license: apache-2.0
tags:
- multilingual
- sea
- sailor
datasets:
- cerebras/SlimPajama-627B
- Skywork/SkyPile-150B
- allenai/MADLAD-400
- cc100
base_model: Qwen/Qwen1.5-4B
model-index:
- name: Sailor-4B
results:
- task:
type: text-generation
dataset:
name: XQuAD-Thai
type: XQuAD-Thai
metrics:
- type: EM (3-Shot)
value: 46.82
name: EM (3-Shot)
- type: F1 (3-Shot)
value: 63.34
name: F1 (3-Shot)
- task:
type: text-generation
dataset:
name: TyDiQA-Indonesian
type: TyDiQA-Indonesian
metrics:
- type: EM (3-Shot)
value: 53.98
name: EM (3-Shot)
- type: F1 (3-Shot)
value: 73.48
name: F1 (3-Shot)
- task:
type: text-generation
dataset:
name: XQuAD-Vietnamese
type: XQuAD-Vietnamese
metrics:
- type: EM (3-Shot)
value: 47.65
name: EM (3-Shot)
- type: F1 (3-Shot)
value: 67.09
name: F1 (3-Shot)
- task:
type: text-generation
dataset:
name: XCOPA-Thai
type: XCOPA-Thai
metrics:
- type: EM (3-Shot)
value: 53.4
name: EM (3-Shot)
- task:
type: text-generation
dataset:
name: XCOPA-Indonesian
type: XCOPA-Indonesian
metrics:
- type: EM (3-Shot)
value: 69.2
name: EM (3-Shot)
- task:
type: text-generation
dataset:
name: XCOPA-Vietnamese
type: XCOPA-Vietnamese
metrics:
- type: EM (3-Shot)
value: 68.2
name: EM (3-Shot)
- task:
type: text-generation
dataset:
name: M3Exam-Thai
type: M3Exam-Thai
metrics:
- type: EM (3-Shot)
value: 27.88
name: EM (3-Shot)
- task:
type: text-generation
dataset:
name: M3Exam-Indonesian
type: M3Exam-Indonesian
metrics:
- type: EM (3-Shot)
value: 31.27
name: EM (3-Shot)
- task:
type: text-generation
dataset:
name: M3Exam-Vietnamese
type: M3Exam-Vietnamese
metrics:
- type: EM (3-Shot)
value: 40.69
name: EM (3-Shot)
- task:
type: text-generation
dataset:
name: BELEBELE-Thai
type: BELEBELE-Thai
metrics:
- type: EM (3-Shot)
value: 36.11
name: EM (3-Shot)
- task:
type: text-generation
dataset:
name: BELEBELE-Indonesian
type: BELEBELE-Indonesian
metrics:
- type: EM (3-Shot)
value: 41.33
name: EM (3-Shot)
- task:
type: text-generation
dataset:
name: BELEBELE-Vietnamese
type: BELEBELE-Vietnamese
metrics:
- type: EM (3-Shot)
value: 38.89
name: EM (3-Shot)
- task:
type: text-generation
name: Text Generation
dataset:
name: AI2 Reasoning Challenge (25-Shot)
type: ai2_arc
config: ARC-Challenge
split: test
args:
num_few_shot: 25
metrics:
- type: acc_norm
value: 44.45
name: normalized accuracy
source:
url: >-
https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=sail/Sailor-4B
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: HellaSwag (10-Shot)
type: hellaswag
split: validation
args:
num_few_shot: 10
metrics:
- type: acc_norm
value: 69.53
name: normalized accuracy
source:
url: >-
https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=sail/Sailor-4B
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: MMLU (5-Shot)
type: cais/mmlu
config: all
split: test
args:
num_few_shot: 5
metrics:
- type: acc
value: 38.99
name: accuracy
source:
url: >-
https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=sail/Sailor-4B
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: TruthfulQA (0-shot)
type: truthful_qa
config: multiple_choice
split: validation
args:
num_few_shot: 0
metrics:
- type: mc2
value: 37.02
source:
url: >-
https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=sail/Sailor-4B
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: Winogrande (5-shot)
type: winogrande
config: winogrande_xl
split: validation
args:
num_few_shot: 5
metrics:
- type: acc
value: 66.06
name: accuracy
source:
url: >-
https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=sail/Sailor-4B
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: GSM8k (5-shot)
type: gsm8k
config: main
split: test
args:
num_few_shot: 5
metrics:
- type: acc
value: 9.1
name: accuracy
source:
url: >-
https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=sail/Sailor-4B
name: Open LLM Leaderboard
Sailor is a suite of Open Language Models tailored for South-East Asia (SEA), focusing on languages such as 🇮🇩Indonesian, 🇹🇭Thai, 🇻🇳Vietnamese, 🇲🇾Malay, and 🇱🇦Lao. Developed with careful data curation, Sailor models are designed to understand and generate text across diverse linguistic landscapes of SEA region. Built from Qwen 1.5 , Sailor encompasses models of varying sizes, spanning from 0.5B to 7B versions for different requirements. We further fine-tune the base model with open-source datasets to get instruction-tuned models, namedly Sailor-Chat. Benchmarking results demonstrate Sailor's proficiency in tasks such as question answering, commonsense reasoning, and other tasks in SEA languages.
The logo was generated by MidJourney
Model Summary
- Model Collections: Base Model & Chat Model
- Project Website: sailorllm.github.io
- Codebase: github.com/sail-sg/sailor-llm
- Technical Report: Coming Soon
Training details
Sailor is crafted by continually pre-training from language models like the remarkable Qwen 1.5 models, which already has a great performance on SEA languages. The pre-training corpus heavily leverages the publicly available corpus, including SlimPajama, SkyPile, CC100 and MADLAD-400.
By employing aggressive data deduplication and careful data cleaning on the collected corpus, we have attained a high-quality dataset spanning various languages. Through systematic experiments to determine the weights of different languages, Sailor models undergo training from 200B to 400B tokens, tailored to different model sizes. The approach boosts their performance on SEA languages while maintaining proficiency in English and Chinese without significant compromise. Finally, we continually pre-train the Qwen1.5-0.5B model with 400 Billion tokens, and other models with 200 Billion tokens to obtain the Sailor models.
Requirements
The code of Sailor has been in the latest Hugging face transformers and we advise you to install transformers>=4.37.0.
Quickstart
Here provides a code snippet to show you how to load the tokenizer and model and how to generate contents.
from transformers import AutoModelForCausalLM, AutoTokenizer
device = "cuda" # the device to load the model
model = AutoModelForCausalLM.from_pretrained("sail/Sailor-4B", device_map="auto")
tokenizer = AutoTokenizer.from_pretrained("sail/Sailor-4B")
input_message = "Model bahasa adalah model probabilistik"
### The given Indonesian input translates to 'A language model is a probabilistic model of.'
model_inputs = tokenizer([input_message], return_tensors="pt").to(device)
generated_ids = model.generate(
model_inputs.input_ids,
max_new_tokens=64
)
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]
print(response)
License
Sailor is distributed under the terms of the Apache License 2.0. No restrict on the research and the commercial use, but should comply with the Qwen License.
Contact Us
If you have any questions, please raise an issue or contact us at doulx@sea.com or liuqian@sea.com.
Open LLM Leaderboard Evaluation Results
Detailed results can be found here
| Metric | Value |
|---|---|
| Avg. | 44.19 |
| AI2 Reasoning Challenge (25-Shot) | 44.45 |
| HellaSwag (10-Shot) | 69.53 |
| MMLU (5-Shot) | 38.99 |
| TruthfulQA (0-shot) | 37.02 |
| Winogrande (5-shot) | 66.06 |
| GSM8k (5-shot) | 9.10 |