File size: 1,450 Bytes
5b60a87
5527a29
9e3433c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
93d9450
 
 
9e3433c
6b10d1a
9e3433c
 
 
14ed4a6
9e3433c
14ed4a6
9e3433c
 
 
 
14ed4a6
 
 
9e3433c
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
import gradio as gr
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM

# Define the BLOOM model name
model_name = "CreitinGameplays/bloom-3b-conversational"

# Load tokenizer and model
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)

def generate_text(prompt):
  """Generates text using the BLOOM model from Hugging Face Transformers."""
  # Encode the prompt into tokens
  input_ids = tokenizer(prompt, return_tensors="pt").input_ids

  # Generate text with the prompt and limit the maximum length to 256 tokens
  output = model.generate(
      input_ids=input_ids,
      max_length=256,
      num_return_sequences=1,  # Generate only 1 sequence
      do_sample=True,  # Enable sampling for creativity
      top_k=50,  # Sample from the top 50 most likely tokens at each step
      top_p=0.15,  # Filter out highly probable unlikely continuations
      temperature=0.1  # Control the randomness of the generated text (1.0 for default)
      repetition_penalty=1.165
  )

  # Decode the generated token sequence back to text
  generated_text = tokenizer.decode(output[0], skip_special_tokens=True)
  return generated_text

# Define the Gradio interface
interface = gr.Interface(
  fn=generate_text,
  inputs="text",
  outputs="text",
  description="Interact with BLOOM (Loaded with Hugging Face Transformers)",
)

# Launch the Gradio interface
interface.launch()