airllm 0.9.2

AirLLM allows single 4GB GPU card to run 70B large language models without quantization, distillation or pruning.

AirLLM optimizes inference memory usage, allowing 70B large language models to run inference on a single 4GB GPU card. No quantization, distillation, pruning or other model compression techniques that would result in degraded model performance are needed.

AirLLM优化inference内存，4GB单卡GPU可以运行70B大语言模型推理。不需要任何损失模型性能的量化和蒸馏，剪枝等模型压缩。

Quickstart 如何使用？

install package 安装

First, install airllm pip package.

首先安装airllm包。

pip install airllm

如果找不到package，可能是因为默认的镜像问题。可以尝试制定原始镜像：

pip install -i https://pypi.org/simple/ airllm

Inference 推理

Then, initialize AirLLMLlama2, pass in the huggingface repo ID of the model being used, or the local path, and inference can be performed similar to a regular transformer model.

然后，初始化AirLLMLlama2，传入所使用模型的huggingface repo ID，或者本地路径即可类似于普通的transformer模型进行推理。

from airllm import AirLLMLlama2

MAX_LENGTH = 128
# could use hugging face model repo id:
model = AirLLMLlama2("garage-bAInd/Platypus2-70B-instruct")

# or use model's local path...
#model = AirLLMLlama2("/home/ubuntu/.cache/huggingface/hub/models--garage-bAInd--Platypus2-70B-instruct/snapshots/b585e74bcaae02e52665d9ac6d23f4d0dbc81a0f")

input_text = [
        'What is the capital of United States?',
        #'I like',
    ]

input_tokens = model.tokenizer(input_text,
    return_tensors="pt", 
    return_attention_mask=False, 
    truncation=True, 
    max_length=MAX_LENGTH, 
    padding=True)
           
generation_output = model.generate(
    input_tokens['input_ids'].cuda(), 
    max_new_tokens=2,
    use_cache=True,
    return_dict_in_generate=True)

output = model.tokenizer.decode(generation_output.sequences[0])

print(output)

Note: During inference, the original model will first be decomposed and saved layer-wise. Please ensure there is sufficient disk space in the huggingface cache directory.

注意：推理过程会首先将原始模型按层分拆，转存。请保证huggingface cache目录有足够的磁盘空间。