File size: 1,000 Bytes
8194866 3bf71d2 8194866 d037b1c 3bf71d2 1116122 fcdc41c 1116122 3bf71d2 d037b1c 3bf71d2 fcdc41c 3bf71d2 d08a677 3bf71d2 1fdbfe6 3bf71d2 d037b1c 1fdbfe6 3bf71d2 d08a677 3bf71d2 |
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 |
import gradio as gr
from peft import PeftModel, PeftConfig
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
# Load the model and config when the script starts
config = PeftConfig.from_pretrained("phearion/bigbrain-v0.0.1")
model = AutoModelForCausalLM.from_pretrained("mistralai/Mistral-7B-v0.1")
model = PeftModel.from_pretrained(model, "phearion/bigbrain-v0.0.1")
# Move the model to the device
model = model.to(device)
# Load the tokenizer
tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-v0.1")
def greet(text):
batch = tokenizer(f"'{text}' ->: ", return_tensors='pt')
# Use torch.no_grad to disable gradient calculation
with torch.no_grad():
output_tokens = model.generate(**batch, max_new_tokens=20)
return tokenizer.decode(output_tokens[0], skip_special_tokens=True)
iface = gr.Interface(fn=greet, inputs="text", outputs="text")
iface.launch() |