Home

🚀 MindNLP

Run HuggingFace Models on MindSpore with Zero Code Changes

The easiest way to use 200,000+ HuggingFace models on Ascend NPU, GPU, and CPU

🎯 What is MindNLP?¶

MindNLP bridges the gap between HuggingFace's massive model ecosystem and MindSpore's hardware acceleration. With just import mindnlp, you can run any HuggingFace model on Ascend NPU, NVIDIA GPU, or CPU - no code changes required.

import mindnlp  # That's it! HuggingFace now runs on MindSpore
from transformers import pipeline

pipe = pipeline("text-generation", model="Qwen/Qwen2-0.5B")
print(pipe("Hello, I am")[0]["generated_text"])

⚡ Quick Start¶

Text Generation with LLMs¶

import mindspore
import mindnlp
from transformers import pipeline

pipe = pipeline(
    "text-generation",
    model="Qwen/Qwen3-8B",
    ms_dtype=mindspore.bfloat16,
    device_map="auto"
)

messages = [{"role": "user", "content": "Write a haiku about coding"}]
print(pipe(messages, max_new_tokens=100)[0]["generated_text"][-1]["content"])

Image Generation with Stable Diffusion¶

import mindspore
import mindnlp
from diffusers import DiffusionPipeline

pipe = DiffusionPipeline.from_pretrained(
    "stable-diffusion-v1-5/stable-diffusion-v1-5",
    ms_dtype=mindspore.float16
)
image = pipe("A sunset over mountains, oil painting style").images[0]
image.save("sunset.png")

✨ Features¶

🤗 HuggingFace Compatibility🚀 Hardware Acceleration🔧 Advanced Capabilities📦 Easy Integration

200,000+ models from HuggingFace Hub
Transformers - All model architectures
Diffusers - Stable Diffusion, SDXL, ControlNet
Zero code changes - Just import mindnlp

Ascend NPU - Full support for Huawei AI chips
NVIDIA GPU - CUDA acceleration
CPU - Optimized CPU execution
Multi-device - Automatic device placement

Mixed precision - FP16/BF16 training & inference
Quantization - INT8/INT4 with BitsAndBytes
Distributed - Multi-GPU/NPU training
PEFT/LoRA - Parameter-efficient fine-tuning

PyTorch-compatible API via mindtorch
Safetensors support for fast loading
Model Hub mirrors for faster downloads
Comprehensive documentation

📦 Installation¶

# From PyPI (recommended)
pip install mindnlp

# From source (latest features)
pip install git+https://github.com/mindspore-lab/mindnlp.git

Version Compatibility¶

MindNLP	MindSpore	Python
0.6.x	≥2.7.1	3.10-3.11
0.5.x	2.5.0-2.7.0	3.10-3.11
0.4.x	2.2.x-2.5.0	3.9-3.11
0.3.x	2.1.0-2.3.1	3.8-3.9

💡 Why MindNLP?¶

Feature	MindNLP	PyTorch + HF	TensorFlow + HF
HuggingFace Models	✅ 200K+	✅ 200K+	⚠️ Limited
Ascend NPU Support	✅ Native	❌	❌
Zero Code Migration	✅	-	❌
Chinese Model Support	✅ Excellent	✅ Good	⚠️ Limited

Key Advantages

Instant Migration: Your existing HuggingFace code works immediately
Ascend Optimization: Native support for Huawei NPU hardware
Production Ready: Battle-tested in enterprise deployments
Active Community: Regular updates and responsive support

🗺️ Supported Models¶

MindNLP supports all models from HuggingFace Transformers and Diffusers:

Category	Models
LLMs	Qwen, Llama, ChatGLM, Mistral, Phi, Gemma, BLOOM, Falcon
Vision	ViT, CLIP, Swin, ConvNeXt, SAM, BLIP
Audio	Whisper, Wav2Vec2, HuBERT, MusicGen
Diffusion	Stable Diffusion, SDXL, ControlNet
Multimodal	LLaVA, Qwen-VL, ALIGN

👉 View all supported models

📚 Next Steps¶

Tutorials

Quick Start - Fine-tune BERT for sentiment analysis
Use Trainer - Configure training with Trainer API
PEFT/LoRA - Parameter-efficient fine-tuning
Data Processing - Dataset handling

Resources

API Reference - API documentation
FAQ - Frequently asked questions
Contributing - How to contribute
Changelog - Version history

🤝 Community¶

Join the MindSpore NLP SIG for discussions and collaboration:

QQ Group

📄 License¶

MindNLP is released under the Apache 2.0 License.

📖 Citation¶

@misc{mindnlp2022,
    title={MindNLP: Easy-to-use and High-performance NLP and LLM Framework Based on MindSpore},
    author={MindNLP Contributors},
    howpublished={\url{https://github.com/mindspore-lab/mindnlp}},
    year={2022}
}

Made with ❤️ by the MindSpore Lab team