Instructions to use Polygl0t/Tucano2-qwen-0.5B-Base with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Polygl0t/Tucano2-qwen-0.5B-Base with Transformers:

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

pipe = pipeline("text-generation", model="Polygl0t/Tucano2-qwen-0.5B-Base")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Polygl0t/Tucano2-qwen-0.5B-Base")
model = AutoModelForCausalLM.from_pretrained("Polygl0t/Tucano2-qwen-0.5B-Base")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use Polygl0t/Tucano2-qwen-0.5B-Base with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Polygl0t/Tucano2-qwen-0.5B-Base"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Polygl0t/Tucano2-qwen-0.5B-Base",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Polygl0t/Tucano2-qwen-0.5B-Base

SGLang

How to use Polygl0t/Tucano2-qwen-0.5B-Base 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 "Polygl0t/Tucano2-qwen-0.5B-Base" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Polygl0t/Tucano2-qwen-0.5B-Base",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "Polygl0t/Tucano2-qwen-0.5B-Base" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Polygl0t/Tucano2-qwen-0.5B-Base",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Polygl0t/Tucano2-qwen-0.5B-Base with Docker Model Runner:
```
docker model run hf.co/Polygl0t/Tucano2-qwen-0.5B-Base
```

Tucano2-qwen-0.5B-Base / training_config.yaml

nicholasKluge

Update training_config.yaml

ddeb77c verified 5 months ago

raw

history blame contribute delete

2.6 kB

	# Directory settings
	checkpoint_dir: "/lustre/scratch/data/polyglot_datasets/portuguese/checkpoints/models/Tucano2-qwen-0.5B-Base"
	train_dataset_dir:
	# Total: ~48B
	# Web Text (~28B, 58%)
	- /lustre/scratch/data/polyglot_datasets/portuguese/tokenized/gigaverbo_v2/5 # 0.1B (PT)
	- /lustre/scratch/data/polyglot_datasets/portuguese/tokenized/gigaverbo_v2/4 # 28B (PT)
	# Synthetic Text (~20B, 42%)
	- /lustre/scratch/data/polyglot_datasets/portuguese/tokenized/gigaverbo_v2_synth # 10B (PT)
	- /lustre/scratch/data/polyglot_datasets/portuguese/tokenized/gigaverbo_v2_synth # 10B (PT)
	val_dataset_dir: "/lustre/scratch/data/polyglot_datasets/portuguese/tokenized/validation"
	dataset_type: "parquet"
	cache_dir: "/lustre/mlnvme/data/nanotronics/.cache"

	# Data loading settings
	pin_memory: true
	num_workers_for_dataloader: 16
	shuffle_dataset: true
	mask_eos_token: false
	mask_pad_token: false

	# Model architecture settings
	vocab_size: 49152
	num_hidden_layers: 28
	num_attention_heads: 16
	num_key_value_heads: 8
	head_dim: 128
	hidden_size: 1024
	intermediate_size: 3072
	max_position_embeddings: 4096
	tie_word_embeddings: true
	hidden_act: "silu"
	output_hidden_states: false
	attn_implementation: "flash_attention_2"
	use_cache: false
	no_rope_layer_interval: null
	rope_theta: 1000000.0
	rope_scale_factor: null
	rms_norm_eps: 0.000001

	# Training settings
	total_batch_size: 1048576
	micro_batch_size: 8
	eval_micro_batch_size: 8
	num_train_epochs: 1
	warmup_steps: 100
	max_learning_rate: 0.00025
	min_learning_rate: 0.0
	muon_learning_rate: null
	weight_decay: 0.1
	beta1: 0.9
	beta2: 0.95
	eps: 0.00000001
	lr_decay_type: "cosine"
	use_sqrt: true
	lr_decay_iters_coef: 1.
	seed: 42
	max_steps: 50000
	max_grad_norm: 1.0

	# Precision and optimization settings
	torch_compile: false
	mat_mul_precision: "highest"
	tf32: true
	bf16: true
	gradient_checkpointing: false
	use_liger_kernel: true
	static_graph: false

	# Hub settings
	push_to_hub: false
	hub_token: null
	hub_model_id: null

	# Tokenizer and Reference model
	tokenizer_name_or_path: "/lustre/scratch/data/polyglot_datasets/portuguese/checkpoints/models/tucano2-qwen-0.5B"
	chat_template_path: null
	reference_model: "/lustre/scratch/data/polyglot_datasets/portuguese/checkpoints/models/tucano2-qwen-0.5B"
	continual_pretraining: true

	# Checkpoint settings
	resume_from_checkpoint: null
	checkpointing_steps: 2500
	begin_new_stage: true
	stage_name: "single_cosine"

	# Miscellaneous settings
	sanity_check: false
	sanity_check_num_samples: 100000
	wandb_token: null
	wandb_id: "tucano2-qwen-0.5b"
	wandb_project: "Polyglot"
	wandb_desc: "Developing LLMs for low-resource languages"