Instructions to use tendrivalentin/witcher3-qwen35-08b-sft-ptbr with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use tendrivalentin/witcher3-qwen35-08b-sft-ptbr with Transformers:

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

pipe = pipeline("text-generation", model="tendrivalentin/witcher3-qwen35-08b-sft-ptbr")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("tendrivalentin/witcher3-qwen35-08b-sft-ptbr")
model = AutoModelForCausalLM.from_pretrained("tendrivalentin/witcher3-qwen35-08b-sft-ptbr", 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=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use tendrivalentin/witcher3-qwen35-08b-sft-ptbr with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "tendrivalentin/witcher3-qwen35-08b-sft-ptbr"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tendrivalentin/witcher3-qwen35-08b-sft-ptbr",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/tendrivalentin/witcher3-qwen35-08b-sft-ptbr

SGLang

How to use tendrivalentin/witcher3-qwen35-08b-sft-ptbr 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 "tendrivalentin/witcher3-qwen35-08b-sft-ptbr" \
    --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": "tendrivalentin/witcher3-qwen35-08b-sft-ptbr",
		"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 "tendrivalentin/witcher3-qwen35-08b-sft-ptbr" \
        --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": "tendrivalentin/witcher3-qwen35-08b-sft-ptbr",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use tendrivalentin/witcher3-qwen35-08b-sft-ptbr with Docker Model Runner:
```
docker model run hf.co/tendrivalentin/witcher3-qwen35-08b-sft-ptbr
```

Witcher 3 Qwen3.5-0.8B SFT (Portuguese)

Fine-tuned version of Qwen/Qwen3.5-0.8B on the witcher3-dataset-ptbr dataset for Witcher 3 themed conversations in Brazilian Portuguese.

Hugging Face Hub: tendrivalentin/witcher3-qwen35-08b-sft-ptbr

Quick Start

From Hugging Face Hub

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("tendrivalentin/witcher3-qwen35-08b-sft-ptbr")
tokenizer = AutoTokenizer.from_pretrained("tendrivalentin/witcher3-qwen35-08b-sft-ptbr")

messages = [{"role": "user", "content": "Quem é Ciri e qual é a sua importância?"}]
text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(text, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=150, temperature=0.7, do_sample=True)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

From Local Path

from transformers import AutoModelForCausalLM, AutoTokenizer

model_path = "runs/sft-witcher3-ptbr-qwen35-08b-full"
model = AutoModelForCausalLM.from_pretrained(model_path)
tokenizer = AutoTokenizer.from_pretrained(model_path)

messages = [{"role": "user", "content": "Fale sobre Yennefer."}]
text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(text, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=150, temperature=0.7, do_sample=True)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Training Results

Metric	Value
Train Loss	0.660
Train Accuracy	85.07%
Eval Loss (best)	2.391
Eval Accuracy	53.93%
Training Time	~35 min
Hardware	Apple Silicon (MPS)

Training Procedure

This model was trained with SFT.

Framework versions

TRL: 1.7.0
Transformers: 5.13.0
Pytorch: 2.12.1
Datasets: 5.0.0
Tokenizers: 0.22.2

Citations

Cite TRL as:

@software{vonwerra2020trl,
  title   = {{TRL: Transformers Reinforcement Learning}},
  author  = {von Werra, Leandro and Belkada, Younes and Tunstall, Lewis and Beeching, Edward and Thrush, Tristan and Lambert, Nathan and Huang, Shengyi and Rasul, Kashif and Gallouédec, Quentin},
  license = {Apache-2.0},
  url     = {https://github.com/huggingface/trl},
  year    = {2020}
}

Downloads last month: 9

Safetensors

Model size

0.8B params

Tensor type

F32

Model tree for tendrivalentin/witcher3-qwen35-08b-sft-ptbr

Base model

Qwen/Qwen3.5-0.8B-Base

Finetuned

Qwen/Qwen3.5-0.8B