feat: style updates, added license and data protection

app.py

@@ -1,7 +1,17 @@
 import gradio as gr
-import chatmodel as chat
+import chatmodel as model
 import interpret as shap
 import visualize as viz
+import markdown
+
+def load_md(filename):
+    path = "./public/"+str(filename)
+
+    # credit: official python-markdown documentation (https://python-markdown.github.io/reference/)
+    with open(path, "r") as file:
+        text = file.read()
+
+    return markdown.markdown(text)
 
 with gr.Blocks() as ui:
     with gr.Row():
@@ -17,47 +27,33 @@ with gr.Blocks() as ui:
                 ### ChatBot Demo
                 Mitral AI 7B Model fine-tuned for instruction and fully open source (see at [HGF](https://huggingface.co/mistralai/Mistral-7B-v0.1))
                 """)
+
+        with gr.Row():
+            chatbot = gr.Chatbot(layout="panel", show_copy_button=True,avatar_images=("./public/human.jpg","./public/bot.jpg"))
+        with gr.Row():
+            gr.Markdown(
+                """
+                ##### ⚠️ All Conversations are recorded for qa assurance and explanation functionality!
+                """)
         with gr.Row():
-            gr.ChatInterface(
-                chat.interference
-            )
+                prompt = gr.Textbox(label="Input Message")
         with gr.Row():
-            gr.Slider(
-                label="Temperature",
-                value=0.7,
-                minimum=0.0,
-                maximum=1.0,
-                step=0.05,
-                interactive=True,
-                info="Higher values produce more diverse outputs",
-            ),
-            gr.Slider(
-                label="Max new tokens",
-                value=256,
-                minimum=0,
-                maximum=1024,
-                step=64,
-                interactive=True,
-                info="The maximum numbers of new tokens",
-            ),
-            gr.Slider(
-                label="Top-p (nucleus sampling)",
-                value=0.95,
-                minimum=0.0,
-                maximum=1,
-                step=0.05,
-                interactive=True,
-                info="Higher values sample more low-probability tokens",
-            ),
-            gr.Slider(
-                label="Repetition penalty",
-                value=1.1,
-                minimum=1.0,
-                maximum=2.0,
-                step=0.05,
-                interactive=True,
-                info="Penalize repeated tokens",
-            )
+            with gr.Column(scale=1):
+                clear_btn = gr.ClearButton([prompt, chatbot])
+            with gr.Column(scale=1):
+                submit_btn = gr.Button("Submit")
+
+        submit_btn.click(model.chat, [prompt, chatbot], [prompt, chatbot])
+        prompt.submit(model.chat, [prompt, chatbot], [prompt, chatbot])
+
+    with gr.Tab("Explanations"):
+        with gr.Row():
+            gr.Markdown(
+                """
+                ### Get Explanations for  
+                SHAP Visualization Dashboard adopted from [shapash](https://github.com/MAIF/shapash)
+                """)
+
 
     with gr.Tab("SHAP Dashboard"):
         with gr.Row():
@@ -83,6 +79,9 @@ with gr.Blocks() as ui:
                 Adopted from official [model paper](https://arxiv.org/abs/2310.06825) by Mistral AI
                 """)
 
+    with gr.Row():
+        with gr.Accordion("Credits, Data Protection and License", open=False):
+            gr.Markdown(value=load_md("credits_dataprotection_license.md"))
 
 if __name__ == "__main__":
     ui.launch(debug=True)

chatmodel.py

@@ -4,10 +4,24 @@ import gradio as gr
 
 token = os.environ.get("HGFTOKEN")
 
-client = InferenceClient(
+interference = InferenceClient(
     "mistralai/Mistral-7B-Instruct-v0.1"
 )
 
+model_temperature = 0.7
+model_max_new_tokens = 256
+model_top_p = 0.95
+model_repetition_penalty = 1.1
+
+def chat (prompt, history,):
+
+    formatted_prompt = format_prompt(prompt, history)
+    answer=respond(formatted_prompt)
+
+    history.append((prompt, answer))
+
+    return "",history
+
 def format_prompt(message, history):
     prompt = "<s>"
     for user_prompt, bot_response in history:
@@ -16,33 +30,20 @@ def format_prompt(message, history):
     prompt += f"[INST] {message} [/INST]"
     return prompt
 
-def interference(
-        prompt, history, temperature=0.7, max_new_tokens=256, top_p=0.95, repetition_penalty=1.1,
-):
-    temperature = float(temperature)
+def respond(formatted_prompt):
+    temperature = float(model_temperature)
     if temperature < 1e-2:
         temperature = 1e-2
-    top_p = float(top_p)
+    top_p = float(model_top_p)
 
     generate_kwargs = dict(
         temperature=temperature,
-        max_new_tokens=max_new_tokens,
+        max_new_tokens=model_max_new_tokens,
         top_p=top_p,
-        repetition_penalty=repetition_penalty,
+        repetition_penalty=model_repetition_penalty,
         do_sample=True,
         seed=42,
     )
 
-    formatted_prompt = format_prompt(prompt, history)
-
-    stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=False)
-    output = ""
-
-    for response in stream:
-        output += response.token.text
-        yield output
-    return output
-
-custom=[
-
-]
+    output = interference.text_generation(formatted_prompt, **generate_kwargs, stream=False, details=True, return_full_text=False).generated_text
+    return output

public/credits_dataprotection_license.md

@@ -0,0 +1,6 @@
+### Credits
+
+### Data Protection
+
+### License
+This Product is licensed under the MIT license.

requirements.txt

@@ -2,4 +2,5 @@ gradio
 transformers
 torch
 shap
-accelerate
+accelerate
+markdown