Spaces:
Sleeping
Sleeping
File size: 1,750 Bytes
9e69980 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 |
import torch
import pandas as pd
from torch.utils.data import Dataset, DataLoader
from transformers import AutoTokenizer
from evo_model import EvoTransformer
import torch.nn as nn
import torch.optim as optim
class FeedbackDataset(Dataset):
def __init__(self, csv_file):
self.data = pd.read_csv(csv_file).dropna()
self.tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
def __len__(self):
return len(self.data)
def __getitem__(self, idx):
row = self.data.iloc[idx]
prompt = row['prompt']
context = row['context']
label = int(row['label'])
text = f"{prompt} {context}"
encoded = self.tokenizer(text, truncation=True, padding='max_length', max_length=128, return_tensors="pt")
return encoded['input_ids'].squeeze(0), torch.tensor(label)
def fine_tune_on_feedback():
csv_file = "feedback_log.csv"
dataset = FeedbackDataset(csv_file)
dataloader = DataLoader(dataset, batch_size=8, shuffle=True)
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model = EvoTransformer().to(device)
model.load_state_dict(torch.load("evo_hellaswag.pt", map_location=device))
model.train()
criterion = nn.CrossEntropyLoss()
optimizer = optim.Adam(model.parameters(), lr=2e-5)
for epoch in range(2):
for input_ids, labels in dataloader:
input_ids = input_ids.to(device)
labels = labels.to(device)
outputs = model(input_ids)
loss = criterion(outputs, labels)
optimizer.zero_grad()
loss.backward()
optimizer.step()
torch.save(model.state_dict(), "evo_hellaswag.pt")
print("✅ Evo retrained and saved.")
|