Text Generation
Transformers
Safetensors
qwen2
multilingual
sea
sailor
conversational
text-generation-inference
Instructions to use sail/Sailor2-1B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use sail/Sailor2-1B with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="sail/Sailor2-1B") messages = [ {"role": "user", "content": "Who are you?"}, ] pipe(messages)# Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("sail/Sailor2-1B") model = AutoModelForCausalLM.from_pretrained("sail/Sailor2-1B") 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
- vLLM
How to use sail/Sailor2-1B with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "sail/Sailor2-1B" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "sail/Sailor2-1B", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/sail/Sailor2-1B
- SGLang
How to use sail/Sailor2-1B 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 "sail/Sailor2-1B" \ --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": "sail/Sailor2-1B", "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 "sail/Sailor2-1B" \ --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": "sail/Sailor2-1B", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use sail/Sailor2-1B with Docker Model Runner:
docker model run hf.co/sail/Sailor2-1B
Add library_name and pipeline_tag metadata
#3
by nielsr HF Staff - opened
README.md
CHANGED
|
@@ -1,4 +1,6 @@
|
|
| 1 |
---
|
|
|
|
|
|
|
| 2 |
language:
|
| 3 |
- en
|
| 4 |
- zh
|
|
@@ -12,6 +14,8 @@ language:
|
|
| 12 |
- km
|
| 13 |
- su
|
| 14 |
- tl
|
|
|
|
|
|
|
| 15 |
tags:
|
| 16 |
- multilingual
|
| 17 |
- sea
|
|
@@ -29,18 +33,16 @@ widget:
|
|
| 29 |
example_title: Indonesian
|
| 30 |
- text: Làm thế nào để nướng cá?
|
| 31 |
example_title: Vietnamese
|
| 32 |
-
|
| 33 |
-
base_model:
|
| 34 |
-
- Qwen/Qwen2.5-0.5B
|
| 35 |
---
|
| 36 |
|
|
|
|
| 37 |
<div align="center">
|
| 38 |
<img src="sailor2_banner.jpg" width="700"/>
|
| 39 |
</div>
|
| 40 |
|
| 41 |
> The logo was generated by MidJourney
|
| 42 |
|
| 43 |
-
|
| 44 |
Sailor2 is a community-driven initiative that brings cutting-edge multilingual language models to South-East Asia (SEA).
|
| 45 |
Our research highlights a strong demand for models in the **8B and 20B parameter** range for production use, alongside **1B models** for specialized applications,
|
| 46 |
such as speculative decoding and research purposes.
|
|
@@ -58,7 +60,6 @@ The Sailor2 model comes in three sizes, 1B, 8B, and 20B, which are **expanded fr
|
|
| 58 |
- **Codebase:** [github.com/sail-sg/sailor2](https://github.com/sail-sg/sailor2)
|
| 59 |
- **Technical Report:** [Sailor2 Report](https://arxiv.org/pdf/2502.12982)
|
| 60 |
|
| 61 |
-
|
| 62 |
## Training details
|
| 63 |
|
| 64 |
During development, we employ a range of advanced technologies to ensure top-tier performance and efficiency:
|
|
@@ -70,7 +71,6 @@ During development, we employ a range of advanced technologies to ensure top-tie
|
|
| 70 |
|
| 71 |
Please refer to [Sailor2 Blog](https://sea-sailor.github.io/blog/sailor2/) for more training details.
|
| 72 |
|
| 73 |
-
|
| 74 |
## Requirements
|
| 75 |
The code of Sailor2 has been in the latest Hugging face transformers and we advise you to install `transformers==4.46.3`.
|
| 76 |
|
|
@@ -145,4 +145,5 @@ If you find Sailor2 useful, please cite our work as follows:
|
|
| 145 |
|
| 146 |
# Contact Us
|
| 147 |
|
| 148 |
-
If you have any questions, please raise an issue or contact us at [doulx@sea.com](mailto:doulx@sea.com) or [liuqian.sea@gmail.com](mailto:liuqian.sea@gmail.com).
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
base_model:
|
| 3 |
+
- Qwen/Qwen2.5-0.5B
|
| 4 |
language:
|
| 5 |
- en
|
| 6 |
- zh
|
|
|
|
| 14 |
- km
|
| 15 |
- su
|
| 16 |
- tl
|
| 17 |
+
license: apache-2.0
|
| 18 |
+
library_name: transformers
|
| 19 |
tags:
|
| 20 |
- multilingual
|
| 21 |
- sea
|
|
|
|
| 33 |
example_title: Indonesian
|
| 34 |
- text: Làm thế nào để nướng cá?
|
| 35 |
example_title: Vietnamese
|
| 36 |
+
pipeline_tag: text-generation
|
|
|
|
|
|
|
| 37 |
---
|
| 38 |
|
| 39 |
+
```markdown
|
| 40 |
<div align="center">
|
| 41 |
<img src="sailor2_banner.jpg" width="700"/>
|
| 42 |
</div>
|
| 43 |
|
| 44 |
> The logo was generated by MidJourney
|
| 45 |
|
|
|
|
| 46 |
Sailor2 is a community-driven initiative that brings cutting-edge multilingual language models to South-East Asia (SEA).
|
| 47 |
Our research highlights a strong demand for models in the **8B and 20B parameter** range for production use, alongside **1B models** for specialized applications,
|
| 48 |
such as speculative decoding and research purposes.
|
|
|
|
| 60 |
- **Codebase:** [github.com/sail-sg/sailor2](https://github.com/sail-sg/sailor2)
|
| 61 |
- **Technical Report:** [Sailor2 Report](https://arxiv.org/pdf/2502.12982)
|
| 62 |
|
|
|
|
| 63 |
## Training details
|
| 64 |
|
| 65 |
During development, we employ a range of advanced technologies to ensure top-tier performance and efficiency:
|
|
|
|
| 71 |
|
| 72 |
Please refer to [Sailor2 Blog](https://sea-sailor.github.io/blog/sailor2/) for more training details.
|
| 73 |
|
|
|
|
| 74 |
## Requirements
|
| 75 |
The code of Sailor2 has been in the latest Hugging face transformers and we advise you to install `transformers==4.46.3`.
|
| 76 |
|
|
|
|
| 145 |
|
| 146 |
# Contact Us
|
| 147 |
|
| 148 |
+
If you have any questions, please raise an issue or contact us at [doulx@sea.com](mailto:doulx@sea.com) or [liuqian.sea@gmail.com](mailto:liuqian.sea@gmail.com).
|
| 149 |
+
```
|