库名称：transformers
许可证：bsd-3-clause

Nijkamp等人提出的基础ProGen2-xlarge模型镜像（配置和前向传播略有修改）。

关于如何微调该模型的示例，请参阅我的GitHub仓库。

使用示例：

from transformers import AutoModelForCausalLM
from tokenizers import Tokenizer
import torch
import torch.nn.functional as F

# 加载模型和分词器
model = AutoModelForCausalLM.from_pretrained("hugohrban/progen2-xlarge", trust_remote_code=True)
tokenizer = Tokenizer.from_pretrained("hugohrban/progen2-xlarge")
tokenizer.no_padding()

# 准备输入
prompt = "1MEVVIVTGMSGAGK"
input_ids = torch.tensor(tokenizer.encode(prompt).ids).to(model.device)

# 前向传播
logits = model(input_ids).logits

# 打印输出概率
next_token_logits = logits[-1, :]
next_token_probs = F.softmax(next_token_logits, dim=-1)
for i in range(tokenizer.get_vocab_size(with_added_tokens=False)):
    print(f"{tokenizer.id_to_token(i)}: {100 * next_token_probs[i].item():.2f} %")