first commit

Dockerfile

@@ -0,0 +1,27 @@
+# Utilizza l'immagine base ufficiale di Hugging Face per PyTorch con supporto CUDA
+FROM huggingface/transformers-pytorch-cuda:latest
+
+# Installa le dipendenze di sistema
+RUN apt-get update && apt-get install -y \
+    git \
+    wget \
+    && rm -rf /var/lib/apt/lists/*
+
+# Imposta le variabili d'ambiente per il token e l'utente
+ENV HUGGINGFACE_TOKEN=hf_1234567890abcdef1234567890abcdef12345678
+ENV HUGGINGFACE_USER=Rathalos
+ENV HUGGINGFACE_REPO=training_incite
+
+# Clona il repository privato
+RUN git clone https://${HUGGINGFACE_TOKEN}@huggingface.co/spaces/${HUGGINGFACE_USER}/${HUGGINGFACE_REPO}.git
+
+# Copia il file requirements.txt e installa le dipendenze Python
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+
+# Copia tutto il contenuto della directory corrente nella directory /app del container
+COPY . /app
+WORKDIR /app
+
+# Comando per eseguire lo script di training
+CMD ["python", "train_model.py"]

requirements.txt

@@ -0,0 +1,3 @@
+torch
+transformers
+datasets

train_model.py

@@ -0,0 +1,65 @@
+import torch
+from torch.utils.data import Dataset, DataLoader
+from transformers import AutoTokenizer, AutoModelForCausalLM, Trainer, TrainingArguments
+
+class TextDataset(Dataset):
+    def __init__(self, text, tokenizer, max_length):
+        self.tokenizer = tokenizer
+        self.input_ids = self.tokenizer(text, return_tensors='pt', max_length=max_length, truncation=True, padding="max_length").input_ids
+
+    def __len__(self):
+        return self.input_ids.size(1)
+
+    def __getitem__(self, idx):
+        return self.input_ids[:, idx]
+
+def main():
+    # Hyperparameters
+    max_length = 512
+    batch_size = 32
+    epochs = 3
+    learning_rate = 5e-5
+
+    # File path
+    text_file_path = 'path/to/your/text/file.txt'  # Modifica questo percorso
+
+    # Load text data
+    with open(text_file_path, 'r', encoding='utf-8') as file:
+        text = file.read()
+
+    # Load tokenizer and model
+    model_name = "togethercomputer/RedPajama-INCITE-Chat-3B-v1"
+    tokenizer = AutoTokenizer.from_pretrained(model_name)
+    model = AutoModelForCausalLM.from_pretrained(model_name)
+
+    # Preprocess data
+    dataset = TextDataset(text, tokenizer, max_length)
+    dataloader = DataLoader(dataset, batch_size=batch_size, shuffle=True)
+
+    # Setup device
+    device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+    model.to(device)
+
+    # Setup optimizer
+    optimizer = torch.optim.AdamW(model.parameters(), lr=learning_rate)
+
+    # Training loop
+    for epoch in range(epochs):
+        print(f"Epoch {epoch + 1}/{epochs}")
+        model.train()
+        for batch in dataloader:
+            inputs = batch.to(device)
+            outputs = model(inputs, labels=inputs)
+            loss = outputs.loss
+            optimizer.zero_grad()
+            loss.backward()
+            optimizer.step()
+            print(f"Loss: {loss.item()}")
+
+    # Save the model
+    model_save_path = 'model'  # Modifica questo percorso
+    model.save_pretrained(model_save_path)
+    tokenizer.save_pretrained(model_save_path)
+
+if __name__ == '__main__':
+    main()