Instructions to use PaxiAI/Vietnamese-Tokenizer with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use PaxiAI/Vietnamese-Tokenizer with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="PaxiAI/Vietnamese-Tokenizer")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("PaxiAI/Vietnamese-Tokenizer", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use PaxiAI/Vietnamese-Tokenizer with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "PaxiAI/Vietnamese-Tokenizer"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "PaxiAI/Vietnamese-Tokenizer",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/PaxiAI/Vietnamese-Tokenizer

SGLang

How to use PaxiAI/Vietnamese-Tokenizer 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 "PaxiAI/Vietnamese-Tokenizer" \
    --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": "PaxiAI/Vietnamese-Tokenizer",
		"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 "PaxiAI/Vietnamese-Tokenizer" \
        --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": "PaxiAI/Vietnamese-Tokenizer",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use PaxiAI/Vietnamese-Tokenizer with Docker Model Runner:
```
docker model run hf.co/PaxiAI/Vietnamese-Tokenizer
```

PaxiAI commited on 19 days ago

Commit

6444ecf

verified ·

1 Parent(s): 05813f4

Update README.md

Browse files

Files changed (1) hide show

README.md +0 -54

README.md CHANGED Viewed

@@ -105,16 +105,6 @@ print(ids)
 print(tokenizer.decode(ids))
 ```
-Example behavior:
-```text
-Input:
-Trí tuệ nhân tạo đang thay đổi cách con người làm việc và học tập.
-Decoded:
-Trí tuệ nhân tạo đang thay đổi cách con người làm việc và học tập.
-```
 ## Chat Template
 The tokenizer includes a simple ChatML-style template:
@@ -195,36 +185,6 @@ was tokenized into approximately one token per common Vietnamese syllable or wor
 The tokenizer is intentionally optimized more strongly for Vietnamese than for code identifiers or rare English terms.
-## Architecture Independence
-This tokenizer is not tied to a specific model implementation.
-It can be used with a model initialized from scratch, for example:
-```python
-from transformers import AutoTokenizer, Qwen2Config, Qwen2ForCausalLM
-tokenizer = AutoTokenizer.from_pretrained(
-    "PaxiAI/VietToken"
-)
-config = Qwen2Config(
-    vocab_size=len(tokenizer),
-    hidden_size=1024,
-    intermediate_size=2816,
-    num_hidden_layers=20,
-    num_attention_heads=16,
-    num_key_value_heads=4,
-    pad_token_id=tokenizer.pad_token_id,
-    bos_token_id=tokenizer.bos_token_id,
-    eos_token_id=tokenizer.eos_token_id,
-)
-model = Qwen2ForCausalLM(config)
-```
-The model above is randomly initialized. No Qwen weights are required.
 ## Compatibility Warning
 Once a model has been pretrained with this tokenizer, the following must remain unchanged:
@@ -236,20 +196,6 @@ Once a model has been pretrained with this tokenizer, the following must remain
 Changing any of these creates a different tokenizer and should be released under a new version.
-## Recommended Versioning
-This release should be treated as:
-```text
-VietnameseTokenizer-v1
-```
-Future incompatible tokenizer changes should use a new version such as:
-```text
-VietnameseTokenizer-v2
-```
 ## Intended Use
 This tokenizer is suitable for:

 print(tokenizer.decode(ids))
 ```
 ## Chat Template
 The tokenizer includes a simple ChatML-style template:
 The tokenizer is intentionally optimized more strongly for Vietnamese than for code identifiers or rare English terms.
 ## Compatibility Warning
 Once a model has been pretrained with this tokenizer, the following must remain unchanged:
 Changing any of these creates a different tokenizer and should be released under a new version.
 ## Intended Use
 This tokenizer is suitable for: