ThomasBaruzier/Qwen2.5-Coder-1.5B-Instruct-GGUF

Llama.cpp imatrix quantizations of Qwen/Qwen2.5-Coder-1.5B-Instruct

Using llama.cpp commit 3ad5451 for quantization.

All quants were made using the imatrix option and Bartowski's calibration file.

---

Perplexity table (the lower the better)

Quant	Size (MB)	PPL	Size (%)	Accuracy (%)	PPL error rate
IQ1_S	416	65.5925	14.10	18.53	1.27
IQ1_M	442	37.5490	14.98	32.37	0.67
IQ2_XXS	487	28.2797	16.51	42.98	0.52
IQ2_XS	524	19.7773	17.76	61.45	0.34
IQ2_S	537	18.3488	18.20	66.24	0.31
IQ2_M	573	16.4023	19.42	74.10	0.28
Q2_K_S	610	16.9464	20.68	71.72	0.28
IQ3_XXS	637	14.0319	21.59	86.61	0.23
Q2_K	644	15.4455	21.83	78.69	0.25
IQ3_XS	697	13.2295	23.63	91.87	0.21
Q3_K_S	725	13.6908	24.58	88.77	0.22
IQ3_S	727	13.0527	24.64	93.11	0.21
IQ3_M	740	12.9347	25.08	93.96	0.21
Q3_K_M	785	13.1079	26.61	92.72	0.21
Q3_K_L	839	12.9958	28.44	93.52	0.21
IQ4_XS	854	12.5238	28.95	97.04	0.20
IQ4_NL	892	12.5165	30.24	97.10	0.20
Q4_0	894	12.6002	30.31	96.46	0.20
Q4_K_S	896	12.4550	30.37	97.58	0.20
Q4_K_M	940	12.4048	31.86	97.97	0.20
Q4_1	969	12.4660	32.85	97.49	0.20
Q5_K_S	1047	12.2279	35.49	99.39	0.20
Q5_0	1050	12.2580	35.59	99.15	0.20
Q5_K_M	1072	12.2216	36.34	99.44	0.20
Q5_1	1125	12.2391	38.13	99.30	0.20
Q6_K	1213	12.1951	41.12	99.66	0.20
Q8_0	1570	12.1583	53.22	99.96	0.20
F16	2950	12.1537	100	100	0.20

---

Qwen2.5-Coder-1.5B-Instruct

Introduction

Qwen2.5-Coder is the latest series of Code-Specific Qwen large language models (formerly known as CodeQwen). As of now, Qwen2.5-Coder has covered six mainstream model sizes, 0.5, 1.5, 3, 7, 14, 32 billion parameters, to meet the needs of different developers. Qwen2.5-Coder brings the following improvements upon CodeQwen1.5:

• Significantly improvements in code generation, code reasoning and code fixing. Base on the strong Qwen2.5, we scale up the training tokens into 5.5 trillion including source code, text-code grounding, Synthetic data, etc. Qwen2.5-Coder-32B has become the current state-of-the-art open-source codeLLM, with its coding abilities matching those of GPT-4o.
• A more comprehensive foundation for real-world applications such as Code Agents. Not only enhancing coding capabilities but also maintaining its strengths in mathematics and general competencies.

This repo contains the instruction-tuned 1.5B Qwen2.5-Coder model, which has the following features:

• Type: Causal Language Models
• Training Stage: Pretraining & Post-training
• Architecture: transformers with RoPE, SwiGLU, RMSNorm, Attention QKV bias and tied word embeddings
• Number of Parameters: 1.54B
• Number of Paramaters (Non-Embedding): 1.31B
• Number of Layers: 28
• Number of Attention Heads (GQA): 12 for Q and 2 for KV
• Context Length: Full 32,768 tokens

For more details, please refer to our blog, GitHub, Documentation, Arxiv.

Requirements

The code of Qwen2.5-Coder has been in the latest Hugging face transformers and we advise you to use the latest version of transformers.

With transformers<4.37.0, you will encounter the following error:

KeyError: 'qwen2'

Quickstart

Here provides a code snippet with apply_chat_template to show you how to load the tokenizer and model and how to generate contents.

from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "Qwen/Qwen2.5-Coder-1.5B-Instruct"

model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype="auto",
    device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained(model_name)

prompt = "write a quick sort algorithm."
messages = [
    {"role": "system", "content": "You are Qwen, created by Alibaba Cloud. You are a helpful assistant."},
    {"role": "user", "content": prompt}
]
text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True
)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)

generated_ids = model.generate(
    **model_inputs,
    max_new_tokens=512
)
generated_ids = [
    output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
]

response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]

Evaluation & Performance

Detailed evaluation results are reported in this 📑 blog.

For requirements on GPU memory and the respective throughput, see results here.

Citation

If you find our work helpful, feel free to give us a cite.

@article{hui2024qwen2,
      title={Qwen2. 5-Coder Technical Report},
      author={Hui, Binyuan and Yang, Jian and Cui, Zeyu and Yang, Jiaxi and Liu, Dayiheng and Zhang, Lei and Liu, Tianyu and Zhang, Jiajun and Yu, Bowen and Dang, Kai and others},
      journal={arXiv preprint arXiv:2409.12186},
      year={2024}
}
@article{qwen2,
      title={Qwen2 Technical Report}, 
      author={An Yang and Baosong Yang and Binyuan Hui and Bo Zheng and Bowen Yu and Chang Zhou and Chengpeng Li and Chengyuan Li and Dayiheng Liu and Fei Huang and Guanting Dong and Haoran Wei and Huan Lin and Jialong Tang and Jialin Wang and Jian Yang and Jianhong Tu and Jianwei Zhang and Jianxin Ma and Jin Xu and Jingren Zhou and Jinze Bai and Jinzheng He and Junyang Lin and Kai Dang and Keming Lu and Keqin Chen and Kexin Yang and Mei Li and Mingfeng Xue and Na Ni and Pei Zhang and Peng Wang and Ru Peng and Rui Men and Ruize Gao and Runji Lin and Shijie Wang and Shuai Bai and Sinan Tan and Tianhang Zhu and Tianhao Li and Tianyu Liu and Wenbin Ge and Xiaodong Deng and Xiaohuan Zhou and Xingzhang Ren and Xinyu Zhang and Xipin Wei and Xuancheng Ren and Yang Fan and Yang Yao and Yichang Zhang and Yu Wan and Yunfei Chu and Yuqiong Liu and Zeyu Cui and Zhenru Zhang and Zhihao Fan},
      journal={arXiv preprint arXiv:2407.10671},
      year={2024}
}