Spaces:
Runtime error
Runtime error
File size: 4,870 Bytes
efe75b3 fb4c1f6 efe75b3 4df4bed efe75b3 7167e27 efe75b3 63d617b efe75b3 9c6cb45 efe75b3 9c6cb45 efe75b3 8cd024c fd3fa6a 8cd024c efe75b3 9c6cb45 efe75b3 9c6cb45 efe75b3 9c6cb45 efe75b3 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 |
import gradio as gr
from PIL import Image
from inference.main import MultiModalPhi2
messages = []
multimodal_phi2 = MultiModalPhi2(
modelname_or_path="GunaKoppula/Llava-Phi2",
temperature=0.2,
max_new_tokens=1024,
device="cpu",
)
def add_content(chatbot, text, image, audio_upload, audio_mic) -> gr.Chatbot:
textflag, imageflag, audioflag = False, False, False
if text not in ["", None]:
chatbot.append((text, None))
textflag = True
if image is not None:
chatbot.append(((image,), None))
imageflag = True
if audio_mic is not None:
chatbot.append(((audio_mic,), None))
audioflag = True
else:
if audio_upload is not None:
chatbot.append(((audio_upload,), None))
audioflag = True
if not any([textflag, imageflag, audioflag]):
# Raise an error if neither text nor file is provided
raise gr.Error("Enter a valid text, image or audio")
return chatbot
def clear_data():
return {prompt: None, image: None, audio_upload: None, audio_mic: None, chatbot: []}
def run(history, text, image, audio_upload, audio_mic):
if text in [None, ""]:
text = None
if audio_upload is not None:
audio = audio_upload
elif audio_mic is not None:
audio = audio_mic
else:
audio = None
print("text", text)
print("image", image)
print("audio", audio)
if image is not None:
image = Image.open(image)
outputs = multimodal_phi2(text, audio, image)
# outputs = ""
history.append((None, outputs.title()))
return history, None, None, None, None
with gr.Blocks() as demo:
gr.Markdown("## MulitModal Phi2 Model Pretraining and Finetuning from Scratch")
gr.Markdown(
"""This is a multimodal implementation of [Phi2](https://huggingface.co/microsoft/phi-2) model.
Please find the source code and training details [here](https://github.com/RaviNaik/ERA-CAPSTONE/MultiModalPhi2).
### Details:
1. LLM Backbone: [Phi2](https://huggingface.co/microsoft/phi-2)
2. Vision Tower: [clip-vit-large-patch14-336](https://huggingface.co/openai/clip-vit-large-patch14-336)
3. Audio Model: [Whisper Tiny](https://huggingface.co/openai/whisper-tiny)
4. Pretraining Dataset: [LAION-CC-SBU dataset with BLIP captions(200k samples)](https://huggingface.co/datasets/liuhaotian/LLaVA-Pretrain)
5. Finetuning Dataset: [Instruct 150k dataset based on COCO](https://huggingface.co/datasets/liuhaotian/LLaVA-Instruct-150K)
6. Finetuned Model: [GunaKoppula/Llava-Phi2](https://huggingface.co/GunaKoppula/Llava-Phi2)
"""
)
with gr.Row():
with gr.Column(scale=4):
# Creating a column with a scale of 6
with gr.Box():
with gr.Row():
# Adding a Textbox with a placeholder "write prompt"
prompt = gr.Textbox(
placeholder="Enter Prompt", lines=2, label="Query", value=None
)
# Creating a column with a scale of 2
with gr.Row():
# Adding image
image = gr.Image(type="filepath", value=None)
# Creating a column with a scale of 2
with gr.Row():
# Add audio
audio_upload = gr.Audio(source="upload", type="filepath")
audio_mic = gr.Audio(
source="microphone", type="filepath", format="mp3"
)
with gr.Row():
# Adding a Button
submit = gr.Button()
clear = gr.Button(value="Clear")
with gr.Row(scale=8):
with gr.Box():
with gr.Row():
chatbot = gr.Chatbot(
avatar_images=("🧑", "🤖"),
height=550,
)
# with gr.Column(scale=8):
# with gr.Box():
# with gr.Row():
# chatbot = gr.Chatbot(
# avatar_images=("🧑", "🤖"),
# height=550,
# )
# with gr.Row():
# # Adding a Button
# submit = gr.Button()
# clear = gr.Button(value="Clear")
submit.click(
add_content,
inputs=[chatbot, prompt, image, audio_upload, audio_mic],
outputs=[chatbot],
).success(
run,
inputs=[chatbot, prompt, image, audio_upload, audio_mic],
outputs=[chatbot, prompt, image, audio_upload, audio_mic],
)
clear.click(
clear_data,
outputs=[prompt, image, audio_upload, audio_mic, chatbot],
)
demo.launch()
|