Spaces:
Sleeping
Sleeping
File size: 2,518 Bytes
c5fb076 c6b669b e2f7280 c6b669b c5fb076 ed678a4 c6b669b ed678a4 c6b669b ed678a4 c6b669b |
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 54 55 56 57 58 59 60 61 62 |
import gradio as gr
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
from transformers import StoppingCriteria, StoppingCriteriaList, TextIteratorStreamer
from threading import Thread
# Loading the tokenizer and model from Hugging Face's model hub.
tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-v0.1")
model = AutoModelForCausalLM.from_pretrained("mistralai/Mistral-7B-v0.1")
# using CUDA for an optimal experience
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
model = model.to(device)
# Defining a custom stopping criteria class for the model's text generation.
class StopOnTokens(StoppingCriteria):
def __call__(self, input_ids: torch.LongTensor, scores: torch.FloatTensor, **kwargs) -> bool:
stop_ids = [2] # IDs of tokens where the generation should stop.
for stop_id in stop_ids:
if input_ids[0][-1] == stop_id: # Checking if the last generated token is a stop token.
return True
return False
# Function to generate model predictions.
def predict(message, history):
history_transformer_format = history + [[message, ""]]
stop = StopOnTokens()
# Formatting the input for the model.
messages = "</s>".join(["</s>".join(["\n<|user|>:" + item[0], "\n<|assistant|>:" + item[1]])
for item in history_transformer_format])
model_inputs = tokenizer([messages], return_tensors="pt").to(device)
streamer = TextIteratorStreamer(tokenizer, timeout=10., skip_prompt=True, skip_special_tokens=True)
generate_kwargs = dict(
model_inputs,
streamer=streamer,
max_new_tokens=1024,
do_sample=True,
top_p=0.95,
top_k=50,
temperature=0.7,
num_beams=1,
stopping_criteria=StoppingCriteriaList([stop])
)
t = Thread(target=model.generate, kwargs=generate_kwargs)
t.start() # Starting the generation in a separate thread.
partial_message = ""
for new_token in streamer:
partial_message += new_token
if '</s>' in partial_message: # Breaking the loop if the stop token is generated.
break
yield partial_message
# Setting up the Gradio chat interface.
gr.ChatInterface(predict,
title="Tinyllama_chatBot",
description="Ask Tiny llama any questions",
examples=['How to cook a fish?', 'Who is the president of US now?']
).launch() # Launching the web interface. |