Quantization made by Richard Erkhov.

Qwen2.5-0.5B-SFT-2e-4-3ep - GGUF

Model creator: https://huggingface.co/JayHyeon/
Original model: https://huggingface.co/JayHyeon/Qwen2.5-0.5B-SFT-2e-4-3ep/

Name	Quant method	Size
Qwen2.5-0.5B-SFT-2e-4-3ep.Q2_K.gguf	Q2_K	0.39GB
Qwen2.5-0.5B-SFT-2e-4-3ep.IQ3_XS.gguf	IQ3_XS	0.39GB
Qwen2.5-0.5B-SFT-2e-4-3ep.IQ3_S.gguf	IQ3_S	0.39GB
Qwen2.5-0.5B-SFT-2e-4-3ep.Q3_K_S.gguf	Q3_K_S	0.39GB
Qwen2.5-0.5B-SFT-2e-4-3ep.IQ3_M.gguf	IQ3_M	0.39GB
Qwen2.5-0.5B-SFT-2e-4-3ep.Q3_K.gguf	Q3_K	0.4GB
Qwen2.5-0.5B-SFT-2e-4-3ep.Q3_K_M.gguf	Q3_K_M	0.4GB
Qwen2.5-0.5B-SFT-2e-4-3ep.Q3_K_L.gguf	Q3_K_L	0.42GB
Qwen2.5-0.5B-SFT-2e-4-3ep.IQ4_XS.gguf	IQ4_XS	0.4GB
Qwen2.5-0.5B-SFT-2e-4-3ep.Q4_0.gguf	Q4_0	0.4GB
Qwen2.5-0.5B-SFT-2e-4-3ep.IQ4_NL.gguf	IQ4_NL	0.4GB
Qwen2.5-0.5B-SFT-2e-4-3ep.Q4_K_S.gguf	Q4_K_S	0.45GB
Qwen2.5-0.5B-SFT-2e-4-3ep.Q4_K.gguf	Q4_K	0.46GB
Qwen2.5-0.5B-SFT-2e-4-3ep.Q4_K_M.gguf	Q4_K_M	0.46GB
Qwen2.5-0.5B-SFT-2e-4-3ep.Q4_1.gguf	Q4_1	0.43GB
Qwen2.5-0.5B-SFT-2e-4-3ep.Q5_0.gguf	Q5_0	0.46GB
Qwen2.5-0.5B-SFT-2e-4-3ep.Q5_K_S.gguf	Q5_K_S	0.48GB
Qwen2.5-0.5B-SFT-2e-4-3ep.Q5_K.gguf	Q5_K	0.49GB
Qwen2.5-0.5B-SFT-2e-4-3ep.Q5_K_M.gguf	Q5_K_M	0.49GB
Qwen2.5-0.5B-SFT-2e-4-3ep.Q5_1.gguf	Q5_1	0.49GB
Qwen2.5-0.5B-SFT-2e-4-3ep.Q6_K.gguf	Q6_K	0.61GB
Qwen2.5-0.5B-SFT-2e-4-3ep.Q8_0.gguf	Q8_0	0.63GB

Original model description:

base_model: Qwen/Qwen2.5-0.5B datasets: HuggingFaceH4/ultrafeedback_binarized library_name: transformers model_name: Qwen2.5-0.5B-SFT-2e-4-3ep tags: - generated_from_trainer - trl - sft licence: license license: mit

Model Card for Qwen2.5-0.5B-SFT-2e-4-3ep

This model is a fine-tuned version of Qwen/Qwen2.5-0.5B on the HuggingFaceH4/ultrafeedback_binarized dataset. It has been trained using TRL.

Quick start

from transformers import pipeline

question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
generator = pipeline("text-generation", model="JayHyeon/Qwen2.5-0.5B-SFT-2e-4-3ep", device="cuda")
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
print(output["generated_text"])

Training procedure

This model was trained with SFT.

Framework versions

TRL: 0.13.0.dev0
Transformers: 4.47.0.dev0
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}}
}