Ko-Qwen2-7B-Instruct

Model Description

This model is a Supervised fine-tuned version of Qwen2-7B -Instruct with DeepSpeed and trl for korean.

Trained Data

• Trained with public data and private data and Generated data (about 50k)

Usage

from transformers import TextStreamer, pipeline, AutoTokenizer, AutoModelForCausalLM

model_id = 'spow12/Ko-Qwen2-7B-Instruct'
tokenizer = AutoTokenizer.from_pretrained(model_id)
# %%
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype=torch.bfloat16,
    attn_implementation="flash_attention_2", 
    device_map='auto',
)
model.eval()

pipe = pipeline("conversational", model=model, tokenizer=tokenizer, device_map='auto')

streamer = TextStreamer(tokenizer)

generation_configs = dict(
    max_new_tokens=2048,
    num_return_sequences=1, 
    temperature=0.1,
    # early_stopping=True,
    repetition_penalty=1.2,
    num_beams=1,
    do_sample=True,
    top_k=20,
    top_p=0.9,
    eos_token_id=tokenizer.eos_token_id,
    pad_token_id=tokenizer.eos_token_id,
    streamer=streamer
)

sys_message = """당신은 친절한 챗봇으로서 상대방의 요청에 최대한 자세하고 친절하게 답해야합니다. 
사용자가 제공하는 정보를 세심하게 분석하여 사용자의 의도를 신속하게 파악하고 그에 따라 답변을 생성해야합니다.  

항상 매우 자연스러운 한국어로 응답하세요."""

message = [
    {
        'role': "system",
        'content': sys_message
    },
    {
        'role': 'user',
        'content': "현재의 경제상황에 대해 어떻게 생각해?."
    }
]
conversation = pipe(message, **generation_configs)
conversation[-1]

<|im_start|>system
당신은 친절한 챗봇으로서 상대방의 요청에 최대한 자세하고 친절하게 답해야합니다. 
사용자가 제공하는 정보를 세심하게 분석하여 사용자의 의도를 신속하게 파악하고 그에 따라 답변을 생성해야합니다.  

항상 매우 자연스러운 한국어로 응답하세요.<|im_end|>
<|im_start|>user
현재의 경제상황에 대해 어떻게 생각해?<|im_end|>
<|im_start|>assistant
저는 인공지능이기 때문에 현재의 경제 상황이나 개인적인 의견을 가지고 있지 않습니다. 하지만, 최근에는 전 세계적으로 경기가 어려워지고 있는 추세입니다. 많은 나라들이 대처하기 위한 다양한 정책과 조치들을 시행하고 있습니다. 이러한 상황에서 각자 자신의 역량을 발휘하며 살아가시길 바랍니다.<|im_end|>

License

This model is licensed under the cc-by-nc-4.0. which allows others to share and adapt the model for non-commercial purposes.

Here is Original Readme.md

Qwen2-7B-Instruct

Introduction

Qwen2 is the new series of Qwen large language models. For Qwen2, we release a number of base language models and instruction-tuned language models ranging from 0.5 to 72 billion parameters, including a Mixture-of-Experts model. This repo contains the instruction-tuned 7B Qwen2 model.

Compared with the state-of-the-art opensource language models, including the previous released Qwen1.5, Qwen2 has generally surpassed most opensource models and demonstrated competitiveness against proprietary models across a series of benchmarks targeting for language understanding, language generation, multilingual capability, coding, mathematics, reasoning, etc.

Qwen2-7B-Instruct supports a context length of up to 131,072 tokens, enabling the processing of extensive inputs. Please refer to this section for detailed instructions on how to deploy Qwen2 for handling long texts.

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

Model Details

Qwen2 is a language model series including decoder language models of different model sizes. For each size, we release the base language model and the aligned chat model. It is based on the Transformer architecture with SwiGLU activation, attention QKV bias, group query attention, etc. Additionally, we have an improved tokenizer adaptive to multiple natural languages and codes.

Training details

We pretrained the models with a large amount of data, and we post-trained the models with both supervised finetuning and direct preference optimization.

Requirements

The code of Qwen2 has been in the latest Hugging face transformers and we advise you to install transformers>=4.37.0, or you might 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
device = "cuda" # the device to load the model onto

model = AutoModelForCausalLM.from_pretrained(
    "Qwen/Qwen2-7B-Instruct",
    torch_dtype="auto",
    device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2-7B-Instruct")

prompt = "Give me a short introduction to large language model."
messages = [
    {"role": "system", "content": "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(device)

generated_ids = model.generate(
    model_inputs.input_ids,
    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]

Processing Long Texts

To handle extensive inputs exceeding 32,768 tokens, we utilize YARN, a technique for enhancing model length extrapolation, ensuring optimal performance on lengthy texts.

For deployment, we recommend using vLLM. You can enable the long-context capabilities by following these steps:

1. Install vLLM: You can install vLLM by running the following command.

pip install "vllm>=0.4.3"

Or you can install vLLM from source.

2. Configure Model Settings: After downloading the model weights, modify the config.json file by including the below snippet:

       {
           "architectures": [
               "Qwen2ForCausalLM"
           ],
           // ...
           "vocab_size": 152064,
   
           // adding the following snippets
           "rope_scaling": {
               "factor": 4.0,
               "original_max_position_embeddings": 32768,
               "type": "yarn"
           }
       }
   

   This snippet enable YARN to support longer contexts.

3. Model Deployment: Utilize vLLM to deploy your model. For instance, you can set up an openAI-like server using the command:

   python -m vllm.entrypoints.openai.api_server --served-model-name Qwen2-7B-Instruct --model path/to/weights
   

   Then you can access the Chat API by:

   curl http://localhost:8000/v1/chat/completions \
       -H "Content-Type: application/json" \
       -d '{
       "model": "Qwen2-7B-Instruct",
       "messages": [
           {"role": "system", "content": "You are a helpful assistant."},
           {"role": "user", "content": "Your Long Input Here."}
       ]
       }'
   

   For further usage instructions of vLLM, please refer to our Github.

Note: Presently, vLLM only supports static YARN, which means the scaling factor remains constant regardless of input length, potentially impacting performance on shorter texts. We advise adding the rope_scaling configuration only when processing long contexts is required.

Evaluation

We briefly compare Qwen2-7B-Instruct with similar-sized instruction-tuned LLMs, including Qwen1.5-7B-Chat. The results are shown below:

Datasets	Llama-3-8B-Instruct	Yi-1.5-9B-Chat	GLM-4-9B-Chat	Qwen1.5-7B-Chat	Qwen2-7B-Instruct
English
MMLU	68.4	69.5	72.4	59.5	70.5
MMLU-Pro	41.0	-	-	29.1	44.1
GPQA	34.2	-	-	27.8	25.3
TheroemQA	23.0	-	-	14.1	25.3
MT-Bench	8.05	8.20	8.35	7.60	8.41
Coding
Humaneval	62.2	66.5	71.8	46.3	79.9
MBPP	67.9	-	-	48.9	67.2
MultiPL-E	48.5	-	-	27.2	59.1
Evalplus	60.9	-	-	44.8	70.3
LiveCodeBench	17.3	-	-	6.0	26.6
Mathematics
GSM8K	79.6	84.8	79.6	60.3	82.3
MATH	30.0	47.7	50.6	23.2	49.6
Chinese
C-Eval	45.9	-	75.6	67.3	77.2
AlignBench	6.20	6.90	7.01	6.20	7.21

Citation

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

@article{qwen2,
  title={Qwen2 Technical Report},
  year={2024}
}