hgdgng commited on
Commit
4efc6f2
1 Parent(s): 16f142e

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +5 -9
app.py CHANGED
@@ -1,6 +1,5 @@
1
- from transformers import MllamaForConditionalGeneration, AutoProcessor, AutoTokenizer, TextIteratorStreamer
2
  from PIL import Image
3
- import os
4
  import requests
5
  import torch
6
  from threading import Thread
@@ -8,16 +7,15 @@ import gradio as gr
8
  from gradio import FileData
9
  import time
10
  import spaces
11
- from src.search import perform_search
12
 
13
  hf_token = os.environ.get("HF_KEY")
14
 
15
  ckpt = "meta-llama/Llama-3.2-11B-Vision-Instruct"
16
 
17
  model = MllamaForConditionalGeneration.from_pretrained(ckpt,
18
- torch_dtype=torch.bfloat16).to("cuda")
19
- processor = AutoProcessor.from_pretrained(ckpt)
20
- tokenizer = AutoTokenizer.from_pretrained( "meta-llama/Llama-3.2-11B-Vision-Instruct", token=hf_token )
21
 
22
 
23
  @spaces.GPU
@@ -80,9 +78,7 @@ def bot_streaming(message, history, max_new_tokens=250):
80
  demo = gr.ChatInterface(fn=bot_streaming, title="Multimodal Llama", examples=[
81
  [{"text": "Which era does this piece belong to? Give details about the era.", "files":["./examples/rococo.jpg"]},
82
  200],
83
- [{"text": "How long does it take from invoice date to due date? Be short and concise.", "files":["./examples/invoice.png"]},
84
- 250],
85
- [{"text": "Where to find this monument? Can you give me other recommendations around the area?", "files":["./examples/wat_arun.jpg"]},
86
  250],
87
  ],
88
  textbox=gr.MultimodalTextbox(),
 
1
+ from transformers import MllamaForConditionalGeneration, AutoProcessor, TextIteratorStreamer
2
  from PIL import Image
 
3
  import requests
4
  import torch
5
  from threading import Thread
 
7
  from gradio import FileData
8
  import time
9
  import spaces
 
10
 
11
  hf_token = os.environ.get("HF_KEY")
12
 
13
  ckpt = "meta-llama/Llama-3.2-11B-Vision-Instruct"
14
 
15
  model = MllamaForConditionalGeneration.from_pretrained(ckpt,
16
+ torch_dtype=torch.bfloat16,
17
+ token=hf_token).to("cuda")
18
+ processor = AutoProcessor.from_pretrained(ckpt, token=hf_token)
19
 
20
 
21
  @spaces.GPU
 
78
  demo = gr.ChatInterface(fn=bot_streaming, title="Multimodal Llama", examples=[
79
  [{"text": "Which era does this piece belong to? Give details about the era.", "files":["./examples/rococo.jpg"]},
80
  200],
81
+ [{"text": "Where do the droughts happen according to this diagram?", "files":["./examples/weather_events.png"]},
 
 
82
  250],
83
  ],
84
  textbox=gr.MultimodalTextbox(),