RichardErkhov/sky-2002_-_tiny-starcoder-ft-gguf

Quantization made by Richard Erkhov.

Github

Discord

Request more models

tiny-starcoder-ft - GGUF

• Model creator: https://huggingface.co/sky-2002/
• Original model: https://huggingface.co/sky-2002/tiny-starcoder-ft/

Name	Quant method	Size
tiny-starcoder-ft.Q2_K.gguf	Q2_K	0.1GB
tiny-starcoder-ft.IQ3_XS.gguf	IQ3_XS	0.1GB
tiny-starcoder-ft.IQ3_S.gguf	IQ3_S	0.1GB
tiny-starcoder-ft.Q3_K_S.gguf	Q3_K_S	0.1GB
tiny-starcoder-ft.IQ3_M.gguf	IQ3_M	0.11GB
tiny-starcoder-ft.Q3_K.gguf	Q3_K	0.11GB
tiny-starcoder-ft.Q3_K_M.gguf	Q3_K_M	0.11GB
tiny-starcoder-ft.Q3_K_L.gguf	Q3_K_L	0.12GB
tiny-starcoder-ft.IQ4_XS.gguf	IQ4_XS	0.11GB
tiny-starcoder-ft.Q4_0.gguf	Q4_0	0.12GB
tiny-starcoder-ft.IQ4_NL.gguf	IQ4_NL	0.12GB
tiny-starcoder-ft.Q4_K_S.gguf	Q4_K_S	0.12GB
tiny-starcoder-ft.Q4_K.gguf	Q4_K	0.12GB
tiny-starcoder-ft.Q4_K_M.gguf	Q4_K_M	0.12GB
tiny-starcoder-ft.Q4_1.gguf	Q4_1	0.12GB
tiny-starcoder-ft.Q5_0.gguf	Q5_0	0.13GB
tiny-starcoder-ft.Q5_K_S.gguf	Q5_K_S	0.13GB
tiny-starcoder-ft.Q5_K.gguf	Q5_K	0.14GB
tiny-starcoder-ft.Q5_K_M.gguf	Q5_K_M	0.14GB
tiny-starcoder-ft.Q5_1.gguf	Q5_1	0.14GB
tiny-starcoder-ft.Q6_K.gguf	Q6_K	0.15GB
tiny-starcoder-ft.Q8_0.gguf	Q8_0	0.18GB

Original model description:

base_model: bigcode/tiny_starcoder_py library_name: transformers model_name: tiny-starcoder-ft tags: - generated_from_trainer - smol-course - module_1 - code_generation - trl - sft licence: license

Model Card for tiny-starcoder-ft

This model is a fine-tuned version of bigcode/tiny_starcoder_py using a samples from iamtarun/python_code_instructions_18k_alpaca dataset. It has been trained using TRL.

Quick start

model_name = "sky-2002/tiny-starcoder-ft"
model = AutoModelForCausalLM.from_pretrained(
    pretrained_model_name_or_path=model_name
).to(device)
tokenizer = AutoTokenizer.from_pretrained(pretrained_model_name_or_path=model_name)

prompt = "Write python code to calculate sum of a list"

# Format with template
messages = [{"role": "user", "content": prompt}]
formatted_prompt = tokenizer.apply_chat_template(messages, tokenize=False)

inputs = tokenizer(formatted_prompt, return_tensors="pt").to(device)

outputs = model.generate(**inputs, max_new_tokens=100)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Training procedure

This model was trained with SFT.

Framework versions

• TRL: 0.12.1
• Transformers: 4.46.3
• Pytorch: 2.5.1
• Datasets: 3.1.0
• Tokenizers: 0.20.3

Citations

Cite TRL as:

@misc{vonwerra2022trl,
    title        = {{TRL: Transformer Reinforcement Learning}},
    author       = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
    year         = 2020,
    journal      = {GitHub repository},
    publisher    = {GitHub},
    howpublished = {\url{https://github.com/huggingface/trl}}
}