EADST

Save Hugging Face Model with One Bin

max_shard_size (int or str, optional, defaults to "10GB") — Only applicable for models. The maximum size for a checkpoint before being sharded. Checkpoints shard will then be each of size lower than this size. If expressed as a string, needs to be digits followed by a unit (like "5MB").

Based on the introduction, one bin model can be saved by changing the "max_shard_size".

LlamaForCausalLM.save_pretrained(base_model, output_dir, max_shard_size="100GB") # save one bin if the model is less than 100GB

Reference

PreTrainedModel

About Me
XD
Goals determine what you are going to be.
Category
标签云
Crawler Proxy Quantize Claude CAM Ubuntu Card Jupyter 论文速读 XML Bert 域名 CLAP Domain Augmentation Data Random Nginx SQL Cloudreve scipy Github 论文 tar Qwen Use ONNX PyCharm Breakpoint Llama XGBoost 财报 Distillation InvalidArgumentError Tensor git UNIX Base64 transformers 关于博主 llama.cpp Paper PyTorch ChatGPT PIP Baidu 第一性原理 Google tqdm 净利润 Safetensors BeautifulSoup Food IndexTTS2 搞笑 TensorFlow YOLO News Algorithm Diagram CTC SPIE Math Transformers Django 算法题 HaggingFace Michelin Disk LoRA OCR uWSGI API COCO Git CC Website Magnet Harness SAM 飞书 Windows Streamlit 腾讯云 图标 音频 OpenAI PDF Rebuttal DeepSeek Hilton Qwen2 Statistics AI Pillow Image2Text TensorRT Pandas Vmess JSON Clash 云服务器 证件照 Video RGB 多线程 Qwen2.5 Vim ResNet-50 MD5 LeetCode Docker BTC WAN LaTeX EXCEL SVR Mixtral FP64 报税 Tracking Logo hf 图形思考法 Bitcoin 多进程 继承 LLAMA logger PDB NLTK WebCrawler torchinfo Input v2ray Ptyhon Anaconda Color Zip VPN FP16 Paddle CUDA CV ms-swift 阿里云 Jev Hungarian uwsgi Bin Freesound GoogLeNet Datetime Sklearn Shortcut Linux SQLite Quantization 公式 FP32 强化学习 LLM Password Pytorch Conda Firewall Translation VSCode API网关 递归学习法 GPTQ OpenCV Agent NameSilo Pickle NLP Review Attention 签证 Web icon Land UI Permission git-lfs 顶会 mmap diffusers HuggingFace Tiktoken RAR Interview v0.dev Bipartite Animate Excel FastAPI RL Dataset C++ GPT4 Jetson 版权 Template TSV TTS QWEN printf Miniforge Plate Markdown ModelScope Python Gemma Numpy DeepStream Heatmap Knowledge FlashAttention GGML Plotly FP8 CSV GIT Search CEIR VGG-16 Hotel BF16
站点统计

本站现有博文337篇,共被浏览961513次

本站已经建立2673天!

热门文章
文章归档
回到顶部