Update app.py

app.py

@@ -1,204 +1,120 @@
 import gradio as gr
-from gradio_leaderboard import Leaderboard, ColumnFilter, SelectColumns
-import pandas as pd
-from apscheduler.schedulers.background import BackgroundScheduler
-from huggingface_hub import snapshot_download
-
-from src.about import (
-    CITATION_BUTTON_LABEL,
-    CITATION_BUTTON_TEXT,
-    EVALUATION_QUEUE_TEXT,
-    INTRODUCTION_TEXT,
-    LLM_BENCHMARKS_TEXT,
-    TITLE,
-)
-from src.display.css_html_js import custom_css
-from src.display.utils import (
-    BENCHMARK_COLS,
-    COLS,
-    EVAL_COLS,
-    EVAL_TYPES,
-    AutoEvalColumn,
-    ModelType,
-    fields,
-    WeightType,
-    Precision
-)
-from src.envs import API, EVAL_REQUESTS_PATH, EVAL_RESULTS_PATH, QUEUE_REPO, REPO_ID, RESULTS_REPO, TOKEN
-from src.populate import get_evaluation_queue_df, get_leaderboard_df
-from src.submission.submit import add_new_eval
-
-
-def restart_space():
-    API.restart_space(repo_id=REPO_ID)
-
-### Space initialisation
-try:
-    print(EVAL_REQUESTS_PATH)
-    snapshot_download(
-        repo_id=QUEUE_REPO, local_dir=EVAL_REQUESTS_PATH, repo_type="dataset", tqdm_class=None, etag_timeout=30, token=TOKEN
-    )
-except Exception:
-    restart_space()
+import torch
+from transformers import AutoModel, AutoTokenizer
+from PIL import Image
+import os
+
+# --- Model Loading ---
+# This section loads the model and tokenizer from Hugging Face.
+# It's set to use bfloat16 for efficiency.
+# Using a try-except block to handle potential errors during model loading.
 try:
-    print(EVAL_RESULTS_PATH)
-    snapshot_download(
-        repo_id=RESULTS_REPO, local_dir=EVAL_RESULTS_PATH, repo_type="dataset", tqdm_class=None, etag_timeout=30, token=TOKEN
-    )
-except Exception:
-    restart_space()
-
-
-LEADERBOARD_DF = get_leaderboard_df(EVAL_RESULTS_PATH, EVAL_REQUESTS_PATH, COLS, BENCHMARK_COLS)
-
-(
-    finished_eval_queue_df,
-    running_eval_queue_df,
-    pending_eval_queue_df,
-) = get_evaluation_queue_df(EVAL_REQUESTS_PATH, EVAL_COLS)
-
-def init_leaderboard(dataframe):
-    if dataframe is None or dataframe.empty:
-        raise ValueError("Leaderboard DataFrame is empty or None.")
-    return Leaderboard(
-        value=dataframe,
-        datatype=[c.type for c in fields(AutoEvalColumn)],
-        select_columns=SelectColumns(
-            default_selection=[c.name for c in fields(AutoEvalColumn) if c.displayed_by_default],
-            cant_deselect=[c.name for c in fields(AutoEvalColumn) if c.never_hidden],
-            label="Select Columns to Display:",
-        ),
-        search_columns=[AutoEvalColumn.model.name, AutoEvalColumn.license.name],
-        hide_columns=[c.name for c in fields(AutoEvalColumn) if c.hidden],
-        filter_columns=[
-            ColumnFilter(AutoEvalColumn.model_type.name, type="checkboxgroup", label="Model types"),
-            ColumnFilter(AutoEvalColumn.precision.name, type="checkboxgroup", label="Precision"),
-            ColumnFilter(
-                AutoEvalColumn.params.name,
-                type="slider",
-                min=0.01,
-                max=150,
-                label="Select the number of parameters (B)",
-            ),
-            ColumnFilter(
-                AutoEvalColumn.still_on_hub.name, type="boolean", label="Deleted/incomplete", default=True
-            ),
-        ],
-        bool_checkboxgroup_label="Hide models",
-        interactive=False,
-    )
-
-
-demo = gr.Blocks(css=custom_css)
-with demo:
-    gr.HTML(TITLE)
-    gr.Markdown(INTRODUCTION_TEXT, elem_classes="markdown-text")
-
-    with gr.Tabs(elem_classes="tab-buttons") as tabs:
-        with gr.TabItem("🏅 LLM Benchmark", elem_id="llm-benchmark-tab-table", id=0):
-            leaderboard = init_leaderboard(LEADERBOARD_DF)
-
-        with gr.TabItem("📝 About", elem_id="llm-benchmark-tab-table", id=2):
-            gr.Markdown(LLM_BENCHMARKS_TEXT, elem_classes="markdown-text")
-
-        with gr.TabItem("🚀 Submit here! ", elem_id="llm-benchmark-tab-table", id=3):
-            with gr.Column():
-                with gr.Row():
-                    gr.Markdown(EVALUATION_QUEUE_TEXT, elem_classes="markdown-text")
-
-                with gr.Column():
-                    with gr.Accordion(
-                        f"✅ Finished Evaluations ({len(finished_eval_queue_df)})",
-                        open=False,
-                    ):
-                        with gr.Row():
-                            finished_eval_table = gr.components.Dataframe(
-                                value=finished_eval_queue_df,
-                                headers=EVAL_COLS,
-                                datatype=EVAL_TYPES,
-                                row_count=5,
-                            )
-                    with gr.Accordion(
-                        f"🔄 Running Evaluation Queue ({len(running_eval_queue_df)})",
-                        open=False,
-                    ):
-                        with gr.Row():
-                            running_eval_table = gr.components.Dataframe(
-                                value=running_eval_queue_df,
-                                headers=EVAL_COLS,
-                                datatype=EVAL_TYPES,
-                                row_count=5,
-                            )
-
-                    with gr.Accordion(
-                        f"⏳ Pending Evaluation Queue ({len(pending_eval_queue_df)})",
-                        open=False,
-                    ):
-                        with gr.Row():
-                            pending_eval_table = gr.components.Dataframe(
-                                value=pending_eval_queue_df,
-                                headers=EVAL_COLS,
-                                datatype=EVAL_TYPES,
-                                row_count=5,
-                            )
+    model_id = "ByteDance-Seed/BAGEL-7B-MoT"
+    tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
+    model = AutoModel.from_pretrained(
+        model_id,
+        torch_dtype=torch.bfloat16,
+        low_cpu_mem_usage=True,
+        trust_remote_code=True
+    ).cuda().eval()
+    print("Model loaded successfully.")
+except Exception as e:
+    print(f"Error loading model: {e}")
+    # Display an error in the Gradio interface if the model fails to load
+    with gr.Blocks() as demo:
+        gr.Markdown("# 🚨 Error")
+        gr.Markdown(f"Failed to load the BAGEL-7B-MoT model. Please check the logs in the Hugging Face Space for more details. Error: {e}")
+    demo.launch()
+    # Exit if the model cannot be loaded
+    exit()
+
+# --- Core Functions for Each Task ---
+
+def generate_image_from_text(text_prompt):
+    """Generates an image based on a text prompt."""
+    if not text_prompt:
+        return None, "Please provide a text prompt."
+    try:
+        inputs = tokenizer(text=text_prompt, return_tensors='pt')
+        inputs = {k: v.cuda() for k, v in inputs.items()}
+        # Generate the image
+        image = model.generate_image(**inputs)[0]
+        return image, "Image generated successfully."
+    except Exception as e:
+        return None, f"An error occurred: {e}"
+
+def understand_image(image, question):
+    """Answers a question about an uploaded image."""
+    if image is None or not question:
+        return "Please upload an image and ask a question."
+    try:
+        # The model expects a list of PIL images
+        pil_image = Image.fromarray(image).convert('RGB')
+        inputs = tokenizer(text=question, images=[pil_image], return_tensors='pt')
+        inputs = {k: v.cuda() for k, v in inputs.items()}
+        # Generate the textual response
+        generated_ids = model.generate(**inputs)
+        response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
+        return response
+    except Exception as e:
+        return f"An error occurred: {e}"
+
+def edit_image(image, instruction):
+    """Edits an image based on a given instruction."""
+    if image is None or not instruction:
+        return None, "Please upload an image and provide an editing instruction."
+    try:
+        pil_image = Image.fromarray(image).convert('RGB')
+        # For image editing, the task needs to be specified
+        inputs = tokenizer(text=instruction, images=[pil_image], return_tensors='pt', task='image-editing')
+        inputs = {k: v.cuda() for k, v in inputs.items()}
+        # Generate the edited image
+        edited_image = model.generate_image(**inputs)[0]
+        return edited_image, "Image edited successfully."
+    except Exception as e:
+        return None, f"An error occurred: {e}"
+
+# --- Gradio Interface ---
+# We use Gradio Blocks to create a tabbed interface for the three functionalities.
+
+with gr.Blocks(theme=gr.themes.Soft(), title="Multimodal BAGEL App") as demo:
+    gr.Markdown("# 🎨 Multimodal BAGEL App")
+    gr.Markdown("A prototype showcasing the capabilities of the `ByteDance-Seed/BAGEL-7B-MoT` model. Deployed on Hugging Face Spaces.")
+
+    with gr.Tabs():
+        # --- Text-to-Image Tab ---
+        with gr.TabItem("Text-to-Image Generation"):
             with gr.Row():
-                gr.Markdown("# ✉️✨ Submit your model here!", elem_classes="markdown-text")
-
+                with gr.Column(scale=1):
+                    t2i_prompt = gr.Textbox(lines=4, label="Prompt", placeholder="e.g., A photo of a bagel on a beach at sunset.")
+                    t2i_button = gr.Button("Generate Image", variant="primary")
+                with gr.Column(scale=1):
+                    t2i_output_image = gr.Image(label="Generated Image", show_label=True)
+                    t2i_status = gr.Textbox(label="Status", interactive=False)
+            t2i_button.click(generate_image_from_text, inputs=[t2i_prompt], outputs=[t2i_output_image, t2i_status])
+
+        # --- Image Understanding Tab ---
+        with gr.TabItem("Image Understanding"):
             with gr.Row():
-                with gr.Column():
-                    model_name_textbox = gr.Textbox(label="Model name")
-                    revision_name_textbox = gr.Textbox(label="Revision commit", placeholder="main")
-                    model_type = gr.Dropdown(
-                        choices=[t.to_str(" : ") for t in ModelType if t != ModelType.Unknown],
-                        label="Model type",
-                        multiselect=False,
-                        value=None,
-                        interactive=True,
-                    )
-
-                with gr.Column():
-                    precision = gr.Dropdown(
-                        choices=[i.value.name for i in Precision if i != Precision.Unknown],
-                        label="Precision",
-                        multiselect=False,
-                        value="float16",
-                        interactive=True,
-                    )
-                    weight_type = gr.Dropdown(
-                        choices=[i.value.name for i in WeightType],
-                        label="Weights type",
-                        multiselect=False,
-                        value="Original",
-                        interactive=True,
-                    )
-                    base_model_name_textbox = gr.Textbox(label="Base model (for delta or adapter weights)")
-
-            submit_button = gr.Button("Submit Eval")
-            submission_result = gr.Markdown()
-            submit_button.click(
-                add_new_eval,
-                [
-                    model_name_textbox,
-                    base_model_name_textbox,
-                    revision_name_textbox,
-                    precision,
-                    weight_type,
-                    model_type,
-                ],
-                submission_result,
-            )
-
-    with gr.Row():
-        with gr.Accordion("📙 Citation", open=False):
-            citation_button = gr.Textbox(
-                value=CITATION_BUTTON_TEXT,
-                label=CITATION_BUTTON_LABEL,
-                lines=20,
-                elem_id="citation-button",
-                show_copy_button=True,
-            )
-
-scheduler = BackgroundScheduler()
-scheduler.add_job(restart_space, "interval", seconds=1800)
-scheduler.start()
-demo.queue(default_concurrency_limit=40).launch()
+                with gr.Column(scale=1):
+                    iu_input_image = gr.Image(type="numpy", label="Upload Image")
+                    iu_question = gr.Textbox(label="Question", placeholder="e.g., What is in this image?")
+                    iu_button = gr.Button("Ask", variant="primary")
+                with gr.Column(scale=1):
+                    iu_answer = gr.Textbox(label="Answer", lines=10, interactive=False)
+            iu_button.click(understand_image, inputs=[iu_input_image, iu_question], outputs=[iu_answer])
+
+        # --- Image Editing Tab ---
+        with gr.TabItem("Image Editing"):
+            with gr.Row():
+                with gr.Column(scale=1):
+                    ie_input_image = gr.Image(type="numpy", label="Upload Image to Edit")
+                    ie_instruction = gr.Textbox(label="Editing Instruction", placeholder="e.g., Make the sky a vibrant pink.")
+                    ie_button = gr.Button("Apply Edit", variant="primary")
+                with gr.Column(scale=1):
+                    ie_output_image = gr.Image(label="Edited Image")
+                    ie_status = gr.Textbox(label="Status", interactive=False)
+            ie_button.click(edit_image, inputs=[ie_input_image, ie_instruction], outputs=[ie_output_image, ie_status])
+
+# Launch the Gradio app
+demo.launch()