bwx-13B-hf

Model Card for Model ID

<!-- Provide a quick summary of what the model is/does. -->

This is an experimental product that can be used to create new LLM bassed on Chinese language.

Model Details

Model Description

<!-- Provide a longer summary of what this model is. -->

• 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]

<!-- Provide the basic links for the model. -->

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

Uses

<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->

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

<!-- This section is meant to convey both technical and sociotechnical 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

<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->

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.

from transformers import LlamaForCausalLM, LlamaTokenizer
import torch

base_model_name = "BlueWhaleX/bwx-13B-hf"
load_type = torch.float16
device = None

generation_config = dict(
    temperature=0.2,
    top_k=40,
    top_p=0.9,
    do_sample=True,
    num_beams=1,
    repetition_penalty=1.3,
    max_new_tokens=400
    )

prompt_input = (
    "Below is an instruction that describes a task. "
    "Write a response that appropriately completes the request.\n\n"
    "### Instruction:\n\n{instruction}\n\n### Response:\n\n"
)
if torch.cuda.is_available():
    device = torch.device(0)
else:
    device = torch.device('cpu')

def generate_prompt(instruction, input=None):
    if input:
        instruction = instruction + '\n' + input
    return prompt_input.format_map({'instruction': instruction})

tokenizer = LlamaTokenizer.from_pretrained(base_model_name)
model = LlamaForCausalLM.from_pretrained(
        base_model_name,
        load_in_8bit=False,
        torch_dtype=load_type,
        low_cpu_mem_usage=True,
        device_map='auto',
        )

model_vocab_size = model.get_input_embeddings().weight.size(0)
tokenzier_vocab_size = len(tokenizer)
if model_vocab_size != tokenzier_vocab_size:
    model.resize_token_embeddings(tokenzier_vocab_size)

raw_input_text = input("Input:")
input_text = generate_prompt(instruction=raw_input_text)
inputs = tokenizer(input_text, return_tensors="pt") 
generation_output = model.generate(
input_ids=inputs["input_ids"].to(device),
    attention_mask=inputs['attention_mask'].to(device),
    eos_token_id=tokenizer.eos_token_id,
    pad_token_id=tokenizer.pad_token_id,
    **generation_config
)
s = generation_output[0]
output = tokenizer.decode(s, skip_special_tokens=True)
response = output.split("### Response:")[1].strip()
print("Response: ", response)
print("\n")

Training Details

Training Data

<!-- This should link to a Data Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->

BAAI/COIG-PC

Training Procedure

<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the 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 <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->

Evaluation

Testing Data

<!-- This should link to a Data Card if possible. -->

20% of the BAAI/COIG-PC dataset.

Citation

<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->

@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}
}