Instructions to use Unbabel/Tower-Plus-2B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Unbabel/Tower-Plus-2B with Transformers:

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

pipe = pipeline("text-generation", model="Unbabel/Tower-Plus-2B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Unbabel/Tower-Plus-2B")
model = AutoModelForCausalLM.from_pretrained("Unbabel/Tower-Plus-2B", device_map="auto")
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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Unbabel/Tower-Plus-2B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Unbabel/Tower-Plus-2B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Unbabel/Tower-Plus-2B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Unbabel/Tower-Plus-2B

SGLang

How to use Unbabel/Tower-Plus-2B 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 "Unbabel/Tower-Plus-2B" \
    --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": "Unbabel/Tower-Plus-2B",
		"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 "Unbabel/Tower-Plus-2B" \
        --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": "Unbabel/Tower-Plus-2B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Unbabel/Tower-Plus-2B with Docker Model Runner:
```
docker model run hf.co/Unbabel/Tower-Plus-2B
```

jmprcp commited on Jun 9, 2025

Commit

3934aed

verified ·

1 Parent(s): c01c641

Update README.md

Browse files

Files changed (1) hide show

README.md +27 -30

README.md CHANGED Viewed

@@ -2,39 +2,36 @@
 base_model: google/gemma-2-2B
 license: cc-by-nc-sa-4.0
 language:
-  - de
-  - nl
-  - is
-  - es
-  - fr
-  - pt
-  - uk
-  - hi
-  - zh
-  - ru
-  - cs
-  - ko
-  - ja
-  - it
-  - en
-  - da
-  - pl
-  - hu
-  - sv
-  - 'no'
-  - ro
-  - fi
 library_name: transformers
-datasets:
-  - Widn/TowerBlocks-v4-250205
-  - Widn/TowerDPO-v4-anthill-250227
 ---
 # Model Description:
-**Tower-v4-Anthill** is build on top of Gemma 2 2B. The model goes through the Continuous Pretraining (CPT), Instruction Tuning (IT), Weighted Preference Optimization (WPO) and GRPO with verifiable rewards. During all stages we include parallel and multilingual data (covering 22 languages).
-This approach makes Tower v4 Anthill one of the best multilingual LLMs under 3B parameters.
 - **Developed by:** Widn
 - **Model type:** A 2B parameter model fine-tuned on a mix of _translation-related tasks_ as well as  _general instruction-following_ datasets that include reasoning, code instructions, etc.
@@ -80,7 +77,7 @@ sampling_params = SamplingParams(
   temperature=0,
   max_tokens=8192,
 )
-llm = LLM(model="Widn/Tower-4-Anthill", tensor_parallel_size=1)
 messages = [{"role": "user", "content": "Translate: Hello, world! into Portuguese."}]
 outputs = llm.chat(messages, sampling_params)
 # Make sure your prompt_token_ids look like this
@@ -97,10 +94,10 @@ print (outputs[0].outputs[0].text)
 import torch
 from transformers import pipeline
-pipe = pipeline("text-generation", model="Widn/Tower-4-Anthill", device_map="auto")
 # We use the tokenizer’s chat template to format each message - see https://huggingface.co/docs/transformers/main/en/chat_templating
 messages = [{"role": "user", "content": "Translate: Hello, world! into Portuguese."}]
 input_ids = pipe.tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=True)
 outputs = pipe(messages, max_new_tokens=256, do_sample=False)
 print(outputs[0]["generated_text"])
-```

 base_model: google/gemma-2-2B
 license: cc-by-nc-sa-4.0
 language:
+- de
+- nl
+- is
+- es
+- fr
+- pt
+- uk
+- hi
+- zh
+- ru
+- cs
+- ko
+- ja
+- it
+- en
+- da
+- pl
+- hu
+- sv
+- 'no'
+- ro
+- fi
 library_name: transformers
 ---
 # Model Description:
+**Tower+ 2B** is build on top of Gemma 2 2B. The model goes through the Continuous Pretraining (CPT), Instruction Tuning (IT), Weighted Preference Optimization (WPO) and GRPO with verifiable rewards. During all stages we include parallel and multilingual data (covering 22 languages).
+This approach makes Tower+ 2B one of the best multilingual LLMs under 3B parameters.
 - **Developed by:** Widn
 - **Model type:** A 2B parameter model fine-tuned on a mix of _translation-related tasks_ as well as  _general instruction-following_ datasets that include reasoning, code instructions, etc.
   temperature=0,
   max_tokens=8192,
 )
+llm = LLM(model="Unbabel/Tower-Plus-2B", tensor_parallel_size=1)
 messages = [{"role": "user", "content": "Translate: Hello, world! into Portuguese."}]
 outputs = llm.chat(messages, sampling_params)
 # Make sure your prompt_token_ids look like this
 import torch
 from transformers import pipeline
+pipe = pipeline("text-generation", model="Unbabel/Tower-Plus-2B", device_map="auto")
 # We use the tokenizer’s chat template to format each message - see https://huggingface.co/docs/transformers/main/en/chat_templating
 messages = [{"role": "user", "content": "Translate: Hello, world! into Portuguese."}]
 input_ids = pipe.tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=True)
 outputs = pipe(messages, max_new_tokens=256, do_sample=False)
 print(outputs[0]["generated_text"])
+```