Instructions to use thealper2/t5-small-commitbench with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use thealper2/t5-small-commitbench with Transformers:

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

pipe = pipeline("text-generation", model="thealper2/t5-small-commitbench")

# Load model directly
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM

tokenizer = AutoTokenizer.from_pretrained("thealper2/t5-small-commitbench")
model = AutoModelForSeq2SeqLM.from_pretrained("thealper2/t5-small-commitbench", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use thealper2/t5-small-commitbench with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "thealper2/t5-small-commitbench"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "thealper2/t5-small-commitbench",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/thealper2/t5-small-commitbench

SGLang

How to use thealper2/t5-small-commitbench 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 "thealper2/t5-small-commitbench" \
    --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": "thealper2/t5-small-commitbench",
		"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 "thealper2/t5-small-commitbench" \
        --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": "thealper2/t5-small-commitbench",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use thealper2/t5-small-commitbench with Docker Model Runner:
```
docker model run hf.co/thealper2/t5-small-commitbench
```

t5-small-commitbench

File size: 1,589 Bytes

1eb78b1

{
  "data": {
    "dataset_name": "Maxscha/commitbench",
    "cache_dir": null,
    "prefix": "generate commit message: ",
    "target_mode": "full",
    "max_source_length": 512,
    "max_target_length": 64,
    "min_diff_chars": 1,
    "min_message_chars": 1,
    "num_proc": 8,
    "max_train_samples": 500000,
    "max_eval_samples": 2000,
    "max_predict_samples": null
  },
  "train": {
    "model_name": "google-t5/t5-small",
    "output_dir": "outputs/t5-small-commitbench",
    "seed": 42,
    "deterministic": false,
    "learning_rate": 3e-05,
    "num_train_epochs": 2.0,
    "per_device_train_batch_size": 32,
    "per_device_eval_batch_size": 64,
    "gradient_accumulation_steps": 1,
    "weight_decay": 0.01,
    "warmup_ratio": 0.05,
    "label_smoothing_factor": 0.0,
    "max_grad_norm": 1.0,
    "lr_scheduler_type": "linear",
    "eval_strategy": "steps",
    "save_strategy": "steps",
    "eval_steps": 3000,
    "save_steps": 3000,
    "logging_steps": 250,
    "save_total_limit": 2,
    "metric_for_best_model": "eval_loss",
    "greater_is_better": false,
    "load_best_model_at_end": true,
    "gradient_checkpointing": false,
    "torch_compile": true,
    "torch_compile_mode": "",
    "group_by_length": true,
    "dataloader_num_workers": 4,
    "precision": "auto",
    "resume_from_checkpoint": null,
    "max_steps": -1
  },
  "generation": {
    "num_beams": 4,
    "max_new_tokens": 64,
    "min_new_tokens": 0,
    "length_penalty": 1.0,
    "no_repeat_ngram_size": 3,
    "early_stopping": true,
    "do_sample": false,
    "eval_num_beams": 1
  }
}