Instructions to use FORMAS/Carapicu-Qwen3-0.6B-CPT-SFT with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use FORMAS/Carapicu-Qwen3-0.6B-CPT-SFT with Transformers:

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

pipe = pipeline("text-generation", model="FORMAS/Carapicu-Qwen3-0.6B-CPT-SFT")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("FORMAS/Carapicu-Qwen3-0.6B-CPT-SFT")
model = AutoModelForCausalLM.from_pretrained("FORMAS/Carapicu-Qwen3-0.6B-CPT-SFT", 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 FORMAS/Carapicu-Qwen3-0.6B-CPT-SFT with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "FORMAS/Carapicu-Qwen3-0.6B-CPT-SFT"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "FORMAS/Carapicu-Qwen3-0.6B-CPT-SFT",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/FORMAS/Carapicu-Qwen3-0.6B-CPT-SFT

SGLang

How to use FORMAS/Carapicu-Qwen3-0.6B-CPT-SFT 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 "FORMAS/Carapicu-Qwen3-0.6B-CPT-SFT" \
    --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": "FORMAS/Carapicu-Qwen3-0.6B-CPT-SFT",
		"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 "FORMAS/Carapicu-Qwen3-0.6B-CPT-SFT" \
        --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": "FORMAS/Carapicu-Qwen3-0.6B-CPT-SFT",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use FORMAS/Carapicu-Qwen3-0.6B-CPT-SFT with Docker Model Runner:
```
docker model run hf.co/FORMAS/Carapicu-Qwen3-0.6B-CPT-SFT
```

See axolotl config

axolotl version: 0.16.1

base_model: Qwen/Qwen3-0.6B
trust_remote_code: true
strict: false

chat_template: qwen3

plugins:
#  - axolotl.integrations.cut_cross_entropy.CutCrossEntropyPlugin

datasets:

  - path: ./dataset_livro_cpt.jsonl
    type: completion
    field: text

  - path: ./dataset_bpln.jsonl
    type: chat_template
    field_messages: conversations
    message_property_mappings:
      role: from
      content: value
    roles:
      user: ["human"]
      assistant: ["gpt"]
      system: ["system"]

dataset_prepared_path: ./process
val_set_size: 0.05
output_dir: ./outputs_cpt_sft/out

sequence_len: 2048
sample_packing: true
eval_sample_packing: false
pad_to_sequence_len: true

wandb_project: Carapicu
wandb_entity:
wandb_watch:
wandb_name:
wandb_log_model:

load_in_8bit: false
load_in_4bit: false

gradient_accumulation_steps: 2
micro_batch_size: 4
num_epochs: 2

optimizer: adamw_torch
lr_scheduler: cosine
learning_rate: 2e-5
weight_decay: 0.01

warmup_ratio: 0.05

bf16: true
fp16: false
tf32: true

gradient_checkpointing: true
gradient_checkpointing_kwargs:
  use_reentrant: false

flash_attention: true

logging_steps: 1
evals_per_epoch: 2
saves_per_epoch: 1
save_total_limit: 2

special_tokens:
  eos_token: "<|im_end|>"

outputs_cpt_sft/out

This model is a fine-tuned version of Qwen/Qwen3-0.6B on the ./dataset_livro_cpt.jsonl and the ./dataset_bpln.jsonl datasets. It achieves the following results on the evaluation set:

Loss: 1.5797
Ppl: 4.8534
Memory/max Active (gib): 9.15
Memory/max Allocated (gib): 9.15
Memory/device Reserved (gib): 12.77

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 3e-05
train_batch_size: 2
eval_batch_size: 2
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 8
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 3
training_steps: 66

Training results

Training Loss	Epoch	Step	Validation Loss	Ppl	Active (gib)	Allocated (gib)	Reserved (gib)
No log	0	0	2.5275	12.5225	6.92	6.92	7.09
2.1827	0.4889	11	1.6815	5.3738	9.15	9.15	14.05
2.0736	0.9778	22	1.6125	5.0154	9.15	9.15	12.77
1.8840	1.4444	33	1.5932	4.9194	9.15	9.15	12.77
1.7737	1.9333	44	1.5815	4.8623	9.15	9.15	12.77
1.7784	2.4	55	1.5793	4.8517	9.15	9.15	12.77
1.9440	2.8889	66	1.5797	4.8534	9.15	9.15	12.77

Framework versions

Transformers 5.8.0
Pytorch 2.11.0+cu128
Datasets 4.5.0
Tokenizers 0.22.2

Curiosity

O carapicu (Eucinostomus gula) é uma espécie de peixe que habita o Oceano Atlântico desde a América do Norte até a Bahia. Chega a medir até 25 centímetros de comprimento. Em janeiro de cada ano, normalmente ocorre o Torneio de Pesca do Carapicu na Ilha de Itaparica na Bahia.

Downloads last month: 28

Safetensors

Model size

0.6B params

Tensor type

BF16

Model tree for FORMAS/Carapicu-Qwen3-0.6B-CPT-SFT

Base model

Qwen/Qwen3-0.6B-Base

Finetuned

Qwen/Qwen3-0.6B

Finetuned

(1149)

this model