Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -11,7 +11,6 @@ multimodal_phi2 = MultiModalPhi2(
|
|
| 11 |
device="cpu",
|
| 12 |
)
|
| 13 |
|
| 14 |
-
|
| 15 |
def add_content(chatbot, input_data, input_type) -> gr.Chatbot:
|
| 16 |
textflag, imageflag, audioflag = False, False, False
|
| 17 |
if input_type == "text":
|
|
@@ -88,155 +87,5 @@ with gr.Blocks() as demo:
|
|
| 88 |
# chatbot.like(print_like_dislike, None, None)
|
| 89 |
|
| 90 |
|
| 91 |
-
|
| 92 |
-
add_content,
|
| 93 |
-
inputs=[chatbot, prompt, image, audio_upload, audio_mic],
|
| 94 |
-
outputs=[chatbot],
|
| 95 |
-
).success(
|
| 96 |
-
run,
|
| 97 |
-
inputs=[chatbot, prompt, image, audio_upload, audio_mic],
|
| 98 |
-
outputs=[chatbot, prompt, image, audio_upload, audio_mic],
|
| 99 |
-
)
|
| 100 |
-
|
| 101 |
-
clear.click(
|
| 102 |
-
clear_data,
|
| 103 |
-
outputs=[prompt, image, audio_upload, audio_mic, chatbot],
|
| 104 |
-
)
|
| 105 |
-
|
| 106 |
-
demo.launch()
|
| 107 |
-
|
| 108 |
-
|
| 109 |
-
import gradio as gr
|
| 110 |
-
from PIL import Image
|
| 111 |
-
from inference.main import MultiModalPhi2
|
| 112 |
-
import os
|
| 113 |
-
|
| 114 |
-
messages = []
|
| 115 |
-
|
| 116 |
-
multimodal_phi2 = MultiModalPhi2(
|
| 117 |
-
modelname_or_path="Navyabhat/Llava-Phi2",
|
| 118 |
-
temperature=0.2,
|
| 119 |
-
max_new_tokens=1024,
|
| 120 |
-
device="cpu",
|
| 121 |
-
)
|
| 122 |
-
|
| 123 |
-
|
| 124 |
-
def add_content(chatbot, text, image, audio_upload, audio_mic) -> gr.Chatbot:
|
| 125 |
-
textflag, imageflag, audioflag = False, False, False
|
| 126 |
-
if text not in ["", None]:
|
| 127 |
-
chatbot.append((text, None))
|
| 128 |
-
textflag = True
|
| 129 |
-
if image is not None:
|
| 130 |
-
chatbot.append(((image,), None))
|
| 131 |
-
imageflag = True
|
| 132 |
-
if audio_mic is not None:
|
| 133 |
-
chatbot.append(((audio_mic,), None))
|
| 134 |
-
audioflag = True
|
| 135 |
-
else:
|
| 136 |
-
if audio_upload is not None:
|
| 137 |
-
chatbot.append(((audio_upload,), None))
|
| 138 |
-
audioflag = True
|
| 139 |
-
if not any([textflag, imageflag, audioflag]):
|
| 140 |
-
# Raise an error if neither text nor file is provided
|
| 141 |
-
raise gr.Error("Enter a valid text, image or audio")
|
| 142 |
-
return chatbot
|
| 143 |
-
|
| 144 |
-
|
| 145 |
-
def clear_data():
|
| 146 |
-
return {"text": None, "image": None, "audio_upload": None, "audio_mic": None, "chatbot": []}
|
| 147 |
-
|
| 148 |
-
|
| 149 |
-
def run(history, text, image, audio_upload, audio_mic):
|
| 150 |
-
if text in [None, ""]:
|
| 151 |
-
text = None
|
| 152 |
-
|
| 153 |
-
if audio_upload is not None:
|
| 154 |
-
audio = audio_upload
|
| 155 |
-
elif audio_mic is not None:
|
| 156 |
-
audio = audio_mic
|
| 157 |
-
else:
|
| 158 |
-
audio = None
|
| 159 |
-
|
| 160 |
-
print("text", text)
|
| 161 |
-
print("image", image)
|
| 162 |
-
print("audio", audio)
|
| 163 |
-
|
| 164 |
-
if image is not None:
|
| 165 |
-
image = Image.open(image)
|
| 166 |
-
outputs = multimodal_phi2(text, audio, image)
|
| 167 |
-
|
| 168 |
-
history.append((None, outputs.title()))
|
| 169 |
-
return history, None, None, None, None
|
| 170 |
-
|
| 171 |
-
|
| 172 |
-
# def print_like_dislike(x: gr.LikeData):
|
| 173 |
-
# print(x.index, x.value, x.liked)
|
| 174 |
-
|
| 175 |
-
|
| 176 |
-
def add_text(history, text):
|
| 177 |
-
history = history + [(text, None)]
|
| 178 |
-
return history, gr.Textbox(value="", interactive=False)
|
| 179 |
-
|
| 180 |
-
|
| 181 |
-
def add_file(history, file):
|
| 182 |
-
history = history + [((file.name,), None)]
|
| 183 |
-
return history
|
| 184 |
-
|
| 185 |
-
|
| 186 |
-
def bot(history):
|
| 187 |
-
response = "**That's cool!**"
|
| 188 |
-
history[-1][1] = ""
|
| 189 |
-
for character in response:
|
| 190 |
-
history[-1][1] += character
|
| 191 |
-
time.sleep(0.05)
|
| 192 |
-
yield history
|
| 193 |
-
|
| 194 |
-
|
| 195 |
-
with gr.Blocks() as demo:
|
| 196 |
-
chatbot = gr.Chatbot(
|
| 197 |
-
[],
|
| 198 |
-
elem_id="chatbot",
|
| 199 |
-
bubble_full_width=False,
|
| 200 |
-
avatar_images=(None, (os.path.join(os.path.dirname(__file__), "avatar.png"))),
|
| 201 |
-
)
|
| 202 |
-
|
| 203 |
-
with gr.Row():
|
| 204 |
-
txt = gr.Textbox(
|
| 205 |
-
scale=4,
|
| 206 |
-
show_label=False,
|
| 207 |
-
placeholder="Enter text and press enter",
|
| 208 |
-
container=False,
|
| 209 |
-
)
|
| 210 |
-
img_audio = gr.UploadButton("📁", file_types=["image", "audio"], label="Upload Image or Audio")
|
| 211 |
-
|
| 212 |
-
with gr.Row():
|
| 213 |
-
# Adding a Button
|
| 214 |
-
submit = gr.Button()
|
| 215 |
-
clear = gr.Button(value="Clear")
|
| 216 |
-
|
| 217 |
-
txt_msg = txt.submit(add_input, [chatbot, txt], [chatbot, txt, "text"], queue=False).then(
|
| 218 |
-
bot, chatbot, chatbot, api_name="bot_response"
|
| 219 |
-
)
|
| 220 |
-
img_audio_msg = img_audio.upload(add_input, [chatbot, img_audio], [chatbot, "image"], queue=False).then(
|
| 221 |
-
bot, chatbot, chatbot
|
| 222 |
-
)
|
| 223 |
-
|
| 224 |
-
# submit.click(
|
| 225 |
-
# add_content,
|
| 226 |
-
# inputs=[chatbot, txt, image, audio_upload, audio_mic],
|
| 227 |
-
# outputs=[chatbot],
|
| 228 |
-
# ).success(
|
| 229 |
-
# run,
|
| 230 |
-
# inputs=[chatbot, txt, image, audio_upload, audio_mic],
|
| 231 |
-
# outputs=[chatbot, txt, image, audio_upload, audio_mic],
|
| 232 |
-
# )
|
| 233 |
-
|
| 234 |
-
clear.click(
|
| 235 |
-
clear_data,
|
| 236 |
-
outputs=[prompt, image, audio_upload, audio_mic, chatbot],
|
| 237 |
-
)
|
| 238 |
-
|
| 239 |
-
# chatbot.like(print_like_dislike, None, None)
|
| 240 |
-
|
| 241 |
-
# demo.queue()
|
| 242 |
demo.launch()
|
|
|
|
| 11 |
device="cpu",
|
| 12 |
)
|
| 13 |
|
|
|
|
| 14 |
def add_content(chatbot, input_data, input_type) -> gr.Chatbot:
|
| 15 |
textflag, imageflag, audioflag = False, False, False
|
| 16 |
if input_type == "text":
|
|
|
|
| 87 |
# chatbot.like(print_like_dislike, None, None)
|
| 88 |
|
| 89 |
|
| 90 |
+
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 91 |
demo.launch()
|