许可证：bsd-3-clause

Nijkamp等人基础ProGen2-base模型的镜像（包含略微修改的配置和前向传递）。

另请参阅我的GitHub仓库，了解如何微调此模型的示例。

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-base", trust_remote_code=True)
tokenizer = Tokenizer.from_pretrained("hugohrban/progen2-base")
tokenizer.no_padding()

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

# 前向传递
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} %")