Spaces:
Runtime error
Runtime error
File size: 3,791 Bytes
cd7e83a 7cc87ca 82ea2ae cd7e83a 7cc87ca cd7e83a ce16a38 cd7e83a 7cc87ca ce16a38 cd7e83a ce16a38 a672d1b ce16a38 82ea2ae ce16a38 82ea2ae ce16a38 82ea2ae ce16a38 82ea2ae ce16a38 82ea2ae ce16a38 82ea2ae ce16a38 82ea2ae ce16a38 7cc87ca ce16a38 82ea2ae 5a2fb73 cd7e83a |
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 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 |
'''import os
import gradio as gr
from transformers import pipeline
from huggingface_hub import InferenceClient
hf_token = os.getenv("gpt2_token")
# Initialize the text generation pipeline
client =
generator = pipeline("text-generation", )
# Define the response function with additional options for customization
def text_generation(
prompt: str,
details: bool = False,
stream: bool = False,
model: str = None,
best_of: int = None,
decoder_input_details: bool = None,
do_sample: bool = False,
frequency_penalty: float = None,
grammar: None = None,
max_new_tokens: int = None,
repetition_penalty: float = None
):
# Setup the configuration for the model generation
gen_params = {
"max_length": 518, # Default, you can tweak it or set from parameters
"num_return_sequences": 1,
"do_sample": do_sample,
"temperature": 0.7, # Controls randomness
"top_k": 50, # You can adjust for more control over sampling
"top_p": 0.9, # Same as above, for sampling
}
if max_new_tokens:
gen_params["max_length"] = max_new_tokens + len(prompt.split())
if frequency_penalty:
gen_params["frequency_penalty"] = frequency_penalty
if repetition_penalty:
gen_params["repetition_penalty"] = repetition_penalty
# Generate the text based on the input prompt and parameters
generated_text = generator(prompt, **gen_params)[0]["generated_text"]
if details:
# Return additional details for debugging if needed
return {
"generated_text": generated_text,
"params_used": gen_params
}
else:
return generated_text
# Create Gradio interface
iface = gr.Interface(
fn=text_generation, # The function we defined
inputs=[
gr.Textbox(label="Input Prompt"), # User input prompt
gr.Checkbox(label="Show Details", default=False), # Option for additional details
gr.Checkbox(label="Stream Mode", default=False), # Streaming checkbox (not used in this example)
gr.Textbox(label="Model (optional)", default=None), # Optional model name
gr.Slider(minimum=1, maximum=5, label="Best of (Optional)", default=None),
gr.Slider(minimum=0.0, maximum=2.0, label="Frequency Penalty (Optional)", default=None),
gr.Slider(minimum=0.0, maximum=2.0, label="Repetition Penalty (Optional)", default=None),
],
outputs="text" # Output is plain text
)
# Launch the interface
iface.launch()
'''
import gradio as gr
from transformers import pipeline
# Load a text generation model (e.g., GPT-2 or a model of your choice)
generator = pipeline("text-generation", model="isitcoding/gpt2_120_finetuned", tokenizer="isitcoding/gpt2_120_finetuned")
# Function to generate assistant's response based on user input
def generate_response(messages):
# Extract the user message from the input format
user_message = messages[-1]["content"]
# Generate a response based on the user's input
response = generator(user_message, max_length=100, num_return_sequences=1)
# Get the assistant's message from the generated output
assistant_message = response[0]["generated_text"]
# Return the updated conversation with user and assistant messages
messages.append({"role": "assistant", "content": assistant_message})
return messages
# Set up the Gradio interface
iface = gr.Interface(
fn=generate_response,
inputs=gr.Textbox(placeholder="Type your message..."), # User input
outputs=gr.JSON(), # JSON format output to display the conversation
live=True, # Ensure real-time updates
title="Text Generation Pipeline",
description="Enter a message to get a response from the assistant."
)
iface.launch()
|