metadata

license: apache-2.0
datasets:
  - BAAI/COIG-PC
language:
  - zh
library_name: transformers
pipeline_tag: question-answering

Model Card for Model ID

This is an experimental product that can be used to create new LLM bassed on Chinese language. It has been created based on Chinese-LLaMA-Alpaca

Model Details

Model Description

Developed by: yjf9966
Model type: LLaMA with enhanced tokenizer-size-49964
Language(s) (NLP): Chinese
License: Apache 2.0
Finetuned from model: Chinese-LLaMA-Alpaca

Model Sources [optional]

Repository: https://huggingface.co/BlueWhaleX/bwx-13B-HF

Uses

You can use the raw model for next sentence prediction, but it's mostly intended to be fine-tuned on a downstream task. Note that this model is primarily aimed at being fine-tuned on tasks that use the whole sentence (potentially masked) to make decisions, such as sequence classification, token classification or question answering.

Bias, Risks, and Limitations

Even if the training data used for this model could be characterized as fairly neutral, this model can have biased predictions. It also inherits some of the bias of its dataset model.

Recommendations

Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.

How to Get Started with the Model

Use the code below to get started with the model.

import torch
import transformers
from transformers import LlamaTokenizer, LlamaForCausalLM

def generate_prompt(text):
    return f"""Below is an instruction that describes a task. Write a response that appropriately completes the request.\n\n" +
               ### Instruction:\n\n{text}\n\n### Response:\n\n"""

tokenizer = LlamaTokenizer.from_pretrained('BlueWhaleX/bwx-13B-HF')
model = LlamaForCausalLM.from_pretrained('BlueWhaleX/bwx-13B-HF').half().cuda()
model.eval()

text = '王国维说：“自周之衰，文王、周公势力之瓦解也，国民之智力成熟于内，政治之纷乱乘之于外，上无统一之制度，下迫于社会之要求，于是诸于九流各创其学说。” 他意在说明 A. 分封制的崩溃 B. 商鞅变法的作用 C. 兼并战争的后果 D. 百家争鸣的原因'
prompt = generate_prompt(text)
input_ids = tokenizer.encode(prompt, return_tensors='pt').to('cuda')

with torch.no_grad():
    output_ids = model.generate(
        input_ids=input_ids,
        max_new_tokens=400,
        temperature=0.2,
        top_k=40,
        top_p=0.9,
        repetition_penalty=1.3
    ).cuda()
output = tokenizer.decode(output_ids[0], skip_special_tokens=True)
response = output.split("### Response:")[1].strip() 
print("Response: ", response, '\n')

Training Details

Training Data

BAAI/COIG-PC

Training Procedure

Preprocessing [optional]

80% for train dataset and 20% for test dataset

Training Hyperparameters

Training regime: fp16 mixed precision, lr=1e-4, lora_rank=8, lora_alpha=32

Evaluation

Testing Data

20% of the BAAI/COIG-PC dataset.

Citation

@software{bwx-13B-HF,
      title={An Enchanced Chinese Language Model based on the Chinese-Alpaca}, 
      url={https://huggingface.co/BlueWhaleX/bwx-13B-HF},
      year={2023}
}