Hugging Face Hub — Modell-Discovery und Deployment

/models          → Durchsuche 500k+ Modelle
/datasets        → Datasets für Training/Fine-Tuning
/spaces          → Gradio/Streamlit Apps
/papers          → arxiv-Preprints
/organizations   → Company/Lab Seiten

├── Model Description (Was ist das?)
├── Intended Use (Wofür soll es genutzt werden?)
├── Training Data (Was wurde zum Trainieren genutzt?)
├── Benchmarks (MMLU, HellaSwag, etc.)
├── Limitations (Bekannte Schwächen)
├── Ethical Considerations
└── How to Use (Code-Beispiele)

pip install transformers torch accelerate bitsandbytes

from transformers import pipeline

# Text Generation
generator = pipeline("text-generation",
                     model="Qwen/Qwen2.5-7B")
result = generator("Explain quantum computing",
                   max_length=200)
print(result[0]["generated_text"])

# Klassifikation
classifier = pipeline("text-classification",
                      model="distilbert-base-uncased-finetuned-sst-2-english")
result = classifier("This movie is great!")
print(result)  # [{'label': 'POSITIVE', 'score': 0.99}]

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

model_name = "meta-llama/Llama-2-7b"
tokenizer = AutoTokenizer.from_pretrained(model_name)

# 4-Bit Quantization
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    load_in_4bit=True,
    bnb_4bit_compute_dtype=torch.float16
)

inputs = tokenizer("Hello, how are you?", return_tensors="pt")
outputs = model.generate(**inputs, max_length=50)
print(tokenizer.decode(outputs[0]))

pip install huggingface-hub

# Login
huggingface-cli login
# Eingabe: https://huggingface.co/settings/tokens kopieren

# Modell downloaden
huggingface-cli download Qwen/Qwen2.5-7B

# Repo auf Hub pushen
huggingface-cli upload username/my-model ./local_folder

# Repo clonen
git clone https://huggingface.co/Qwen/Qwen2.5-7B
cd Qwen2.5-7B

from huggingface_hub import hf_hub_download, snapshot_download

# Single file
model_file = hf_hub_download(
    repo_id="Qwen/Qwen2.5-7B",
    filename="pytorch_model.bin"
)

# Gesamtes Repo
repo_path = snapshot_download(
    repo_id="Qwen/Qwen2.5-7B",
    cache_dir="/models"
)

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained(
    "meta-llama/Llama-2-7b"  # Funktioniert jetzt (wenn Zugriff approved)
)

# app.py
import gradio as gr
from transformers import pipeline

pipe = pipeline("text-generation", model="Qwen/Qwen2.5-7B")

def chat(message):
    response = pipe(message, max_length=200)
    return response[0]["generated_text"]

iface = gr.Interface(
    fn=chat,
    inputs="text",
    outputs="text",
    title="Local LLM Chat",
    live=False
)

iface.launch(share=False)

git add app.py requirements.txt
git commit -m "Add chatbot"
git push
# App wird automatisch deployed

from datasets import load_dataset

# CSV-Dataset
dataset = load_dataset("csv", data_files="file.csv")

# Hugging Face gehostetes Dataset
dataset = load_dataset("wikitext", "wikitext-2")
print(dataset["train"][0])
# {'text': 'A long article...'}

pip install peft

from peft import get_peft_model, LoraConfig
from transformers import AutoModelForCausalLM

model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen2.5-7B")

lora_config = LoraConfig(
    r=8,                          # Rank
    lora_alpha=16,
    target_modules=["q_proj", "v_proj"],
    lora_dropout=0.1,
    bias="none"
)

model = get_peft_model(model, lora_config)
print(model.print_trainable_parameters())
# Output: trainable params: 1048576 || all params: 7342506880

# Jetzt: Train mit Trainer API
from transformers import Trainer, TrainingArguments

training_args = TrainingArguments(
    output_dir="./lora-qwen",
    per_device_train_batch_size=4,
    num_train_epochs=3,
    learning_rate=1e-4,
    save_strategy="steps",
    save_steps=100,
)

trainer = Trainer(
    model=model,
    args=training_args,
    train_dataset=train_dataset,
)

trainer.train()

model.save_pretrained("./my-lora-model")

from huggingface_hub import upload_folder
upload_folder(
    repo_id="username/my-lora-model",
    folder_path="./my-lora-model"
)

# Fork & Clone
git clone https://huggingface.co/Qwen/Qwen2.5-7B
cd Qwen2.5-7B
git checkout -b improve-description

# Edit README.md
# Add gptq quantization info
git add README.md
git commit -m "Add GPTQ quantization details"
git push
# Öffne PR auf Hub

{
  "nodes": [
    {
      "type": "http",
      "url": "https://api-inference.huggingface.co/models/Qwen/Qwen2.5-7B",
      "method": "POST",
      "headers": {
        "Authorization": "Bearer {{ $secret.HF_API_KEY }}"
      },
      "body": {
        "inputs": "{{ $json.prompt }}"
      }
    }
  ]
}