Re-structuration

CLAUDE.md

@@ -0,0 +1,91 @@
+# CLAUDE.md
+
+This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
+
+## Project Overview
+
+This is **TCID** (Transformer CI Dashboard) - a Gradio-based web dashboard that displays test results for Transformer models across AMD and NVIDIA hardware. The application fetches CI test data from HuggingFace datasets and presents it through interactive visualizations and detailed failure reports.
+
+## Architecture
+
+### Core Components
+
+- **`app.py`** - Main Gradio application with UI components, plotting functions, and data visualization logic
+- **`data.py`** - Data fetching module that retrieves test results from HuggingFace datasets for AMD and NVIDIA CI runs
+- **`styles.css`** - Complete dark theme styling for the Gradio interface
+- **`requirements.txt`** - Python dependencies (matplotlib only)
+
+### Data Flow
+
+1. **Data Loading**: `get_data()` in `data.py` fetches latest CI results from:
+   - AMD: `hf://datasets/optimum-amd/transformers_daily_ci`
+   - NVIDIA: `hf://datasets/hf-internal-testing/transformers_daily_ci`
+
+2. **Data Processing**: Results are joined and filtered to show only important models defined in `IMPORTANT_MODELS` list
+
+3. **Visualization**: Two main views:
+   - **Summary Page**: Horizontal bar charts showing test results for all models
+   - **Detail View**: Pie charts for individual models with failure details
+
+### UI Architecture
+
+- **Sidebar**: Model selection, refresh controls, CI job links
+- **Main Content**: Dynamic display switching between summary and detail views
+- **Auto-refresh**: Data reloads every 15 minutes via background threading
+
+## Running the Application
+
+### Development Commands
+
+```bash
+# Install dependencies
+pip install -r requirements.txt
+
+# Run the application
+python app.py
+```
+
+### HuggingFace Spaces Deployment
+
+This application is configured for HuggingFace Spaces deployment:
+- **Framework**: Gradio 5.38.0
+- **App file**: `app.py`
+- **Configuration**: See `README.md` header for Spaces metadata
+
+## Key Data Structures
+
+### Model Results DataFrame
+The joined DataFrame contains these columns:
+- `success_amd` / `success_nvidia` - Number of passing tests
+- `failed_multi_no_amd` / `failed_multi_no_nvidia` - Multi-GPU failure counts
+- `failed_single_no_amd` / `failed_single_no_nvidia` - Single-GPU failure counts
+- `failures_amd` / `failures_nvidia` - Detailed failure information objects
+- `job_link_amd` / `job_link_nvidia` - CI job URLs
+
+### Important Models List
+Predefined list in `data.py` focusing on significant models:
+- Classic models: bert, gpt2, t5, vit, clip, whisper
+- Modern models: llama, gemma3, qwen2, mistral3
+- Multimodal: qwen2_5_vl, llava, smolvlm, internvl
+
+## Styling and Theming
+
+The application uses a comprehensive dark theme with:
+- Fixed sidebar layout (300px width)
+- Black background throughout (`#000000`)
+- Custom scrollbars with dark styling
+- Monospace fonts for technical aesthetics
+- Gradient buttons and hover effects
+
+## Error Handling
+
+- **Data Loading Failures**: Falls back to predefined model list for testing
+- **Missing Model Data**: Shows "No data available" message in visualizations
+- **Empty Results**: Gracefully handles cases with no test results
+
+## Performance Considerations
+
+- **Memory Management**: Matplotlib configured to prevent memory warnings
+- **Interactive Mode**: Disabled to prevent figure accumulation
+- **Auto-reload**: Background threading with daemon timers
+- **Data Caching**: Global variables store loaded data between UI updates

app.py

@@ -1,76 +1,30 @@
 import matplotlib.pyplot as plt
 import matplotlib
-import numpy as np
 import pandas as pd
 import gradio as gr
 import threading
-import time
-from datetime import datetime
-from data import get_data
+
+from data import CIResults
+from utils import logger, generate_underlined_line
+from summary_page import create_summary_page
 
 # Configure matplotlib to prevent memory warnings and set dark background
-matplotlib.rcParams['figure.max_open_warning'] = 0
 matplotlib.rcParams['figure.facecolor'] = '#000000'
 matplotlib.rcParams['axes.facecolor'] = '#000000' 
 matplotlib.rcParams['savefig.facecolor'] = '#000000'
 plt.ioff()  # Turn off interactive mode to prevent figure accumulation
 
-# Global variables for data
-df = pd.DataFrame()
-available_models = []
-last_update_time = None
-
-def load_data():
-    """Load data from the data source."""
-    global df, available_models, last_update_time
-    try:
-        print(f"[{datetime.now().strftime('%Y-%m-%d %H:%M:%S')}] Loading data...")
-        new_df = get_data()
-        new_models = new_df.index.tolist()
-        
-        # Update global variables
-        df = new_df
-        available_models = new_models
-        last_update_time = datetime.now()
-        
-        print(f"[{datetime.now().strftime('%Y-%m-%d %H:%M:%S')}] Data loaded successfully: {len(available_models)} models")
-        print(f"Models: {available_models[:5]}{'...' if len(available_models) > 5 else ''}")
-        
-        return True
-    except Exception as e:
-        print(f"[{datetime.now().strftime('%Y-%m-%d %H:%M:%S')}] Error loading data: {e}")
-        return False
-
-def schedule_data_reload():
-    """Schedule the next data reload."""
-    def reload_data():
-        load_data()
-        # Schedule the next reload in 15 minutes (900 seconds)
-        timer = threading.Timer(900.0, reload_data)
-        timer.daemon = True  # Dies when main thread dies
-        timer.start()
-        print(f"[{datetime.now().strftime('%Y-%m-%d %H:%M:%S')}] Next data reload scheduled in 15 minutes")
-    
-    # Start the first reload timer
-    timer = threading.Timer(900.0, reload_data)
-    timer.daemon = True
-    timer.start()
-    print(f"[{datetime.now().strftime('%Y-%m-%d %H:%M:%S')}] Data auto-reload scheduled every 15 minutes")
 
 # Load data once at startup
-if not load_data():
-    print("WARNING: Failed to load data! Adding fallback models.")
-    available_models = ["auto", "bert", "clip", "llama", "t5"]  # Fallback models for testing
-
+Ci_results = CIResults()
+Ci_results.load_data()
 # Start the auto-reload scheduler
-schedule_data_reload()
+Ci_results.schedule_data_reload()
 
-def generate_underlined_line(text: str) -> str:
-    return text + "\n" + "─" * len(text) + "\n"
 
 def plot_model_stats(model_name: str) -> tuple[plt.Figure, str, str]:
     """Draws a pie chart of model's passed, failed, skipped, and error stats."""
-    if df.empty or model_name not in df.index:
+    if Ci_results.df.empty or model_name not in Ci_results.df.index:
         # Handle case where model data is not available
         fig, ax = plt.subplots(figsize=(10, 8), facecolor='#000000')
         ax.set_facecolor('#000000')
@@ -83,7 +37,7 @@ def plot_model_stats(model_name: str) -> tuple[plt.Figure, str, str]:
         ax.axis('off')
         return fig, "No data available", "No data available"
     
-    row = df.loc[model_name]
+    row = Ci_results.df.loc[model_name]
     
     # Handle missing values and get counts directly from dataframe
     success_amd = int(row.get('success_amd', 0)) if pd.notna(row.get('success_amd', 0)) else 0
@@ -276,139 +230,7 @@ def extract_failure_info(failures_obj, device: str, multi_count: int, single_cou
             return f"Failures detected on {device} (Multi: {multi_count}, Single: {single_count})\nDetails unavailable: {str(e)}"
         return f"Error processing failure info for {device}: {str(e)}"
 
-def create_summary_page() -> plt.Figure:
-    """Create a summary page with model names and both AMD/NVIDIA test stats bars."""
-    if df.empty:
-        fig, ax = plt.subplots(figsize=(16, 8), facecolor='#000000')
-        ax.set_facecolor('#000000')
-        ax.text(0.5, 0.5, 'No data available', 
-                horizontalalignment='center', verticalalignment='center',
-                transform=ax.transAxes, fontsize=20, color='#888888',
-                fontfamily='monospace', weight='normal')
-        ax.axis('off')
-        return fig
-    
-    fig, ax = plt.subplots(figsize=(16, len(available_models) * 2.5 + 2), facecolor='#000000')
-    ax.set_facecolor('#000000')
-    
-    colors = {
-        'passed': '#4CAF50',
-        'failed': '#E53E3E', 
-        'skipped': '#FFD54F',
-        'error': '#8B0000'
-    }
-    
-    visible_model_count = 0
-    max_y = 0
-    
-    for i, model_name in enumerate(available_models):
-        if model_name not in df.index:
-            continue
-            
-        row = df.loc[model_name]
-        
-        # Get values directly from dataframe
-        success_amd = int(row.get('success_amd', 0)) if pd.notna(row.get('success_amd', 0)) else 0
-        success_nvidia = int(row.get('success_nvidia', 0)) if pd.notna(row.get('success_nvidia', 0)) else 0
-        failed_multi_amd = int(row.get('failed_multi_no_amd', 0)) if pd.notna(row.get('failed_multi_no_amd', 0)) else 0
-        failed_multi_nvidia = int(row.get('failed_multi_no_nvidia', 0)) if pd.notna(row.get('failed_multi_no_nvidia', 0)) else 0
-        failed_single_amd = int(row.get('failed_single_no_amd', 0)) if pd.notna(row.get('failed_single_no_amd', 0)) else 0
-        failed_single_nvidia = int(row.get('failed_single_no_nvidia', 0)) if pd.notna(row.get('failed_single_no_nvidia', 0)) else 0
-        
-        # Calculate stats
-        amd_stats = {
-            'passed': success_amd,
-            'failed': failed_multi_amd + failed_single_amd,
-            'skipped': 0,
-            'error': 0
-        }
-        
-        nvidia_stats = {
-            'passed': success_nvidia,
-            'failed': failed_multi_nvidia + failed_single_nvidia,
-            'skipped': 0,
-            'error': 0
-        }
-        
-        amd_total = sum(amd_stats.values())
-        nvidia_total = sum(nvidia_stats.values())
-        
-        if amd_total == 0 and nvidia_total == 0:
-            continue
-            
-        # Position for this model - use visible model count for spacing
-        y_base = (2.2 + visible_model_count) * 1.8
-        y_model_name = y_base    # Model name above AMD bar
-        y_amd_bar = y_base + 0.45       # AMD bar
-        y_nvidia_bar = y_base + 0.97    # NVIDIA bar
-        max_y = max(max_y, y_nvidia_bar + 0.5)
-        
-        # Model name centered above the AMD bar
-        left_0 = 8
-        bar_length = 92
-        ax.text(bar_length / 2 + left_0, y_model_name, f"{model_name.lower()}", 
-               ha='center', va='center', color='#FFFFFF', 
-               fontsize=20, fontfamily='monospace', fontweight='bold')
-        
-        # AMD label and bar on the same level
-        if amd_total > 0:
-            ax.text(left_0 - 2, y_amd_bar, "amd", 
-                   ha='right', va='center', color='#CCCCCC', 
-                   fontsize=18, fontfamily='monospace', fontweight='normal')
-            
-            # AMD bar starts after labels
-            left = left_0
-            for category in ['passed', 'failed', 'skipped', 'error']:
-                if amd_stats[category] > 0:
-                    width = amd_stats[category] / amd_total * bar_length
-                    ax.barh(y_amd_bar, width, left=left, height=0.405, 
-                           color=colors[category], alpha=0.9)
-                    if width > 4:
-                        ax.text(left + width/2, y_amd_bar, str(amd_stats[category]), 
-                               ha='center', va='center', color='black', 
-                               fontweight='bold', fontsize=12, fontfamily='monospace')
-                    left += width
-        
-        # NVIDIA label and bar on the same level
-        if nvidia_total > 0:
-            ax.text(left_0 - 2, y_nvidia_bar, "nvidia", 
-                   ha='right', va='center', color='#CCCCCC', 
-                   fontsize=18, fontfamily='monospace', fontweight='normal')
-            
-            # NVIDIA bar starts after labels
-            left = left_0
-            for category in ['passed', 'failed', 'skipped', 'error']:
-                if nvidia_stats[category] > 0:
-                    width = nvidia_stats[category] / nvidia_total * bar_length
-                    ax.barh(y_nvidia_bar, width, left=left, height=0.405, 
-                           color=colors[category], alpha=0.9)
-                    if width > 4:
-                        ax.text(left + width/2, y_nvidia_bar, str(nvidia_stats[category]), 
-                               ha='center', va='center', color='black', 
-                               fontweight='bold', fontsize=12, fontfamily='monospace')
-                    left += width
-        
-        # Increment counter for next visible model
-        visible_model_count += 1
-    
-    # Style the axes to be completely invisible and span full width
-    ax.set_xlim(0, 100)
-    ax.set_ylim(-0.5, max_y)
-    ax.set_xlabel('')
-    ax.set_ylabel('')
-    ax.spines['bottom'].set_visible(False)
-    ax.spines['left'].set_visible(False)
-    ax.spines['top'].set_visible(False)
-    ax.spines['right'].set_visible(False)
-    ax.set_xticks([])
-    ax.set_yticks([])
-    ax.yaxis.set_inverted(True)
-    
-    # Remove all margins to make bars span full width
-    plt.tight_layout()
-    plt.subplots_adjust(left=0.02, right=0.98, top=0.98, bottom=0.02)
-    
-    return fig
+
 
 # Load CSS from external file
 def load_css():
@@ -416,7 +238,7 @@ def load_css():
         with open("styles.css", "r") as f:
             return f.read()
     except FileNotFoundError:
-        print("Warning: styles.css not found, using minimal default styles")
+        logger.warning("styles.css not found, using minimal default styles")
         return "body { background: #000; color: #fff; }"
 
 # Create the Gradio interface with sidebar and dark theme
@@ -429,8 +251,8 @@ with gr.Blocks(title="Model Test Results Dashboard", css=load_css()) as demo:
             gr.Markdown("**Transformer CI Dashboard**\n\n*Result overview by model and hardware*\n")
             
             # Data status indicator
-            if last_update_time:
-                status_text = f"📊 **Updated:** {last_update_time.strftime('%H:%M')}\n\n*Auto-refresh: 15min*"
+            if Ci_results.last_update_time:
+                status_text = f"📊 **Updated:** {Ci_results.last_update_time}\n\n*Auto-refresh: 15min*"
             else:
                 status_text = f"📊 **Loading...**\n\n*Auto-refresh: 15min*"
             status_display = gr.Markdown(status_text)
@@ -456,9 +278,9 @@ with gr.Blocks(title="Model Test Results Dashboard", css=load_css()) as demo:
             
             # Back to simple buttons that work
             # Model selector dropdown - much better for long lists
-            gr.Markdown(f"**Select Model ({len(available_models)}):**")
+            gr.Markdown(f"**Select Model ({len(Ci_results.available_models)}):**")
             
-            model_choices = [model.lower() for model in available_models] if available_models else ["auto", "bert", "clip", "llama"]
+            model_choices = [model.lower() for model in Ci_results.available_models] if Ci_results.available_models else ["auto", "bert", "clip", "llama"]
             model_dropdown = gr.Dropdown(
                 choices=model_choices,
                 value=model_choices[0] if model_choices else "auto",
@@ -471,7 +293,7 @@ with gr.Blocks(title="Model Test Results Dashboard", css=load_css()) as demo:
         with gr.Column(scale=4, elem_classes=["main-content"]):
             # Summary display (default view)
             summary_display = gr.Plot(
-                value=create_summary_page(),
+                value=create_summary_page(Ci_results.df, Ci_results.available_models),
                 label="", 
                 format="png",
                 elem_classes=["plot-container"],
@@ -522,7 +344,7 @@ with gr.Blocks(title="Model Test Results Dashboard", css=load_css()) as demo:
     # Summary button click handler
     def show_summary_and_update_links():
         """Show summary page and update CI links."""
-        return create_summary_page(), get_ci_links()
+        return create_summary_page(Ci_results.df, Ci_results.available_models), get_ci_links()
     
     summary_button.click(
         fn=show_summary_and_update_links,
@@ -535,8 +357,8 @@ with gr.Blocks(title="Model Test Results Dashboard", css=load_css()) as demo:
     # Function to get current status text
     def get_status_text():
         """Get current status text with last update time."""
-        if last_update_time:
-            return f"📊 **Updated:** {last_update_time.strftime('%H:%M')}\n\n*Auto-refresh: 15min*"
+        if Ci_results.last_update_time:
+            return f"📊 **Updated:** {Ci_results.last_update_time}\n\n*Auto-refresh: 15min*"
         else:
             return f"📊 **Loading...**\n\n*Auto-refresh: 15min*"
     
@@ -545,7 +367,7 @@ with gr.Blocks(title="Model Test Results Dashboard", css=load_css()) as demo:
         """Get CI job links from the most recent data."""
         try:
             # Check if df exists and is not empty
-            if 'df' not in globals() or df is None or df.empty:
+            if Ci_results.df is None or Ci_results.df.empty:
                 return "🔗 **CI Jobs:** *Loading...*"
             
             # Get links from any available model (they should be the same for all models in a run)
@@ -554,8 +376,8 @@ with gr.Blocks(title="Model Test Results Dashboard", css=load_css()) as demo:
             nvidia_multi_link = None
             nvidia_single_link = None
             
-            for model_name in df.index:
-                row = df.loc[model_name]
+            for model_name in Ci_results.df.index:
+                row = Ci_results.df.loc[model_name]
                 
                 # Extract AMD links
                 if pd.notna(row.get('job_link_amd')) and (not amd_multi_link or not amd_single_link):
@@ -603,19 +425,14 @@ with gr.Blocks(title="Model Test Results Dashboard", css=load_css()) as demo:
                 
             return links_md
         except Exception as e:
-            print(f"Error getting CI links: {e}")
+            logger.error(f"getting CI links: {e}")
             return "🔗 **CI Jobs:** *Error loading links*"
     
     # Refresh button click handler
     def refresh_data_and_status():
         """Manual data refresh triggered by user."""
-        success = load_data()
-        if success:
-            # Return updated summary page, status, and CI links
-            return create_summary_page(), get_status_text(), get_ci_links()
-        else:
-            # Return current summary page, status, and CI links if reload failed
-            return create_summary_page(), get_status_text(), get_ci_links()
+        Ci_results.load_data()
+        return create_summary_page(Ci_results.df, Ci_results.available_models), get_status_text(), get_ci_links()
     
     refresh_button.click(
         fn=refresh_data_and_status,

data.py

@@ -1,5 +1,9 @@
-from huggingface_hub import HfApi, HfFileSystem, login
+from huggingface_hub import HfFileSystem
 import pandas as pd
+from utils import logger
+import os
+from datetime import datetime
+import threading
 
 fs = HfFileSystem()
 
@@ -29,39 +33,26 @@ IMPORTANT_MODELS = [
 ]
 
 
-def get_data():
-    files_amd = fs.glob(
-        "hf://datasets/optimum-amd/transformers_daily_ci/**/runs/**/ci_results_run_models_gpu/model_results.json"
-    )
-    files_amd.sort(reverse=True)
-
-    df_amd = pd.read_json(f"hf://{files_amd[0]}", orient="index")
-    df_amd.index.name = "model_name"
-    df_amd["failed_multi_no_amd"] = df_amd["failures"].apply(
-        lambda x: len(x["multi"]) if "multi" in x else 0
-    )
-    df_amd["failed_single_no_amd"] = df_amd["failures"].apply(
-        lambda x: len(x["single"]) if "single" in x else 0
-    )
-
-    files_nvidia = fs.glob(
-        "hf://datasets/hf-internal-testing/transformers_daily_ci/**/ci_results_run_models_gpu/model_results.json"
-    )
-    files_nvidia.sort(reverse=True)
-
-    df_nvidia = pd.read_json(
-        f"https://huggingface.co/datasets/hf-internal-testing/transformers_daily_ci/raw/main/{files_nvidia[0].lstrip('datasets/hf-internal-testing/transformers_daily_ci/')}",
-        orient="index",
-    )
-    df_nvidia.index.name = "model_name"
-    df_nvidia["failed_multi_no_nvidia"] = df_nvidia["failures"].apply(
-        lambda x: len(x["multi"]) if "multi" in x else 0
-    )
-    df_nvidia["failed_single_no_nvidia"] = df_nvidia["failures"].apply(
-        lambda x: len(x["single"]) if "single" in x else 0
-    )
-    df_nvidia
+def read_one_dataframe(json_path: str, device_label: str) -> pd.DataFrame:
+    df = pd.read_json(json_path, orient="index")
+    df.index.name = "model_name"
+    df[f"failed_multi_no_{device_label}"] = df["failures"].apply(lambda x: len(x["multi"]) if "multi" in x else 0)
+    df[f"failed_single_no_{device_label}"] = df["failures"].apply(lambda x: len(x["single"]) if "single" in x else 0)
+    return df
 
+def get_distant_data() -> pd.DataFrame:
+    # Retrieve AMD dataframe
+    amd_src = "hf://datasets/optimum-amd/transformers_daily_ci/**/runs/**/ci_results_run_models_gpu/model_results.json"
+    files_amd = sorted(fs.glob(amd_src), reverse=True)
+    df_amd = read_one_dataframe(f"hf://{files_amd[0]}", "amd")
+    # Retrieve NVIDIA dataframe
+    nvidia_src = "hf://datasets/hf-internal-testing/transformers_daily_ci/**/ci_results_run_models_gpu/model_results.json"
+    files_nvidia = sorted(fs.glob(nvidia_src), reverse=True)
+    # NOTE: should this be removeprefix instead of lstrip? 
+    nvidia_path = files_nvidia[0].lstrip('datasets/hf-internal-testing/transformers_daily_ci/')
+    nvidia_path = "https://huggingface.co/datasets/hf-internal-testing/transformers_daily_ci/raw/main/" + nvidia_path
+    df_nvidia = read_one_dataframe(nvidia_path, "nvidia")
+    # Join both dataframes
     joined = df_amd.join(df_nvidia, rsuffix="_nvidia", lsuffix="_amd", how="outer")
     joined = joined[
         [
@@ -77,10 +68,58 @@ def get_data():
             "job_link_nvidia",
         ]
     ]
-
     joined.index = joined.index.str.replace("^models_", "", regex=True)
-
+    # Fitler out all but important models
     important_models_lower = [model.lower() for model in IMPORTANT_MODELS]
     filtered_joined = joined[joined.index.str.lower().isin(important_models_lower)]
-
     return filtered_joined
+
+
+def get_sample_data() -> pd.DataFrame:
+    path = os.path.join(os.path.dirname(__file__), "sample_data.csv")
+    df = pd.read_csv(path)
+    df = df.set_index("model_name")
+    return df
+
+
+
+class CIResults:
+
+    def __init__(self):
+        self.df = pd.DataFrame()
+        self.available_models = []
+        self.last_update_time = ""
+
+    def load_data(self) -> None:
+        """Load data from the data source."""
+        # Try loading the distant data, and fall back on sample data for local tinkering
+        try:
+            logger.info("Loading distant data...")
+            new_df = get_distant_data()
+        except Exception as e:
+            logger.error(f"Loading data failed: {e}")
+            logger.warning("Falling back on sample data.")
+            new_df = get_sample_data()
+        # Update attributes
+        self.df = new_df
+        self.available_models = new_df.index.tolist()
+        self.last_update_time = datetime.now().strftime('%H:%M')
+        # Log and return distant load status
+        logger.info(f"Data loaded successfully: {len(self.available_models)} models")
+        logger.info(f"Models: {self.available_models[:5]}{'...' if len(self.available_models) > 5 else ''}")
+
+    def schedule_data_reload(self):
+        """Schedule the next data reload."""
+        def reload_data():
+            self.load_data()
+            # Schedule the next reload in 15 minutes (900 seconds)
+            timer = threading.Timer(900.0, reload_data)
+            timer.daemon = True  # Dies when main thread dies
+            timer.start()
+            logger.info("Next data reload scheduled in 15 minutes")
+
+        # Start the first reload timer
+        timer = threading.Timer(900.0, reload_data)
+        timer.daemon = True
+        timer.start()
+        logger.info("Data auto-reload scheduled every 15 minutes")

sample_data.csv

@@ -0,0 +1,22 @@
+model_name,success_amd,success_nvidia,failed_multi_no_amd,failed_multi_no_nvidia,failed_single_no_amd,failed_single_no_nvidia,failures_amd,failures_nvidia,job_link_amd,job_link_nvidia
+sample_auto,80,226,0,0,0,0,{},{},"{'multi': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447501262', 'single': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447500785'}","{'single': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526561673', 'multi': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526561472'}"
+sample_bert,239,527,2,2,2,2,"{'multi': [{'line': 'tests/models/bert/test_modeling_bert.py::BertModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}, {'line': 'tests/models/bert/test_modeling_bert.py::BertModelTest::test_sdpa_padding_matches_padding_free_with_position_ids', 'trace': '(line 4201)  AssertionError: Tensor-likes are not equal!'}], 'single': [{'line': 'tests/models/bert/test_modeling_bert.py::BertModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}, {'line': 'tests/models/bert/test_modeling_bert.py::BertModelTest::test_sdpa_padding_matches_padding_free_with_position_ids', 'trace': '(line 4201)  AssertionError: Tensor-likes are not equal!'}]}","{'single': [{'line': 'tests/models/bert/test_modeling_bert.py::BertModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}, {'line': 'tests/models/bert/test_modeling_bert.py::BertModelTest::test_sdpa_padding_matches_padding_free_with_position_ids', 'trace': '(line 4216)  AssertionError: Tensor-likes are not equal!'}], 'multi': [{'line': 'tests/models/bert/test_modeling_bert.py::BertModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}, {'line': 'tests/models/bert/test_modeling_bert.py::BertModelTest::test_sdpa_padding_matches_padding_free_with_position_ids', 'trace': '(line 4216)  AssertionError: Tensor-likes are not equal!'}]}","{'multi': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447501282', 'single': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447500788'}","{'single': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526561709', 'multi': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526561482'}"
+clip,288,660,0,0,0,0,{},{},"{'single': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447500866', 'multi': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447501323'}","{'multi': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526561994', 'single': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526562125'}"
+detr,69,177,4,0,4,0,"{'multi': [{'line': 'tests/models/detr/test_modeling_detr.py::DetrModelIntegrationTestsTimmBackbone::test_inference_no_head', 'trace': '(line 595)  AssertionError: Tensor-likes are not close!'}, {'line': 'tests/models/detr/test_modeling_detr.py::DetrModelIntegrationTestsTimmBackbone::test_inference_object_detection_head', 'trace': '(line 619)  AssertionError: Tensor-likes are not close!'}, {'line': 'tests/models/detr/test_modeling_detr.py::DetrModelIntegrationTestsTimmBackbone::test_inference_panoptic_segmentation_head', 'trace': '(line 667)  AssertionError: Tensor-likes are not close!'}, {'line': 'tests/models/detr/test_modeling_detr.py::DetrModelIntegrationTests::test_inference_no_head', 'trace': '(line 741)  AssertionError: Tensor-likes are not close!'}], 'single': [{'line': 'tests/models/detr/test_modeling_detr.py::DetrModelIntegrationTestsTimmBackbone::test_inference_no_head', 'trace': '(line 595)  AssertionError: Tensor-likes are not close!'}, {'line': 'tests/models/detr/test_modeling_detr.py::DetrModelIntegrationTestsTimmBackbone::test_inference_object_detection_head', 'trace': '(line 619)  AssertionError: Tensor-likes are not close!'}, {'line': 'tests/models/detr/test_modeling_detr.py::DetrModelIntegrationTestsTimmBackbone::test_inference_panoptic_segmentation_head', 'trace': '(line 667)  AssertionError: Tensor-likes are not close!'}, {'line': 'tests/models/detr/test_modeling_detr.py::DetrModelIntegrationTests::test_inference_no_head', 'trace': '(line 741)  AssertionError: Tensor-likes are not close!'}]}",{},"{'multi': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447501397', 'single': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447500969'}","{'single': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526562517', 'multi': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526562397'}"
+gemma3,349,499,8,8,7,7,"{'single': [{'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3ModelTest::test_generate_compilation_all_outputs', 'trace': ""(line 317)  torch._dynamo.exc.Unsupported: isinstance(NestedUserFunctionVariable(), TorchInGraphFunctionVariable(<class 'torch.nn.parameter.Parameter'>)): can't determine type of NestedUserFunctionVariable()""}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3Vision2TextModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3Vision2TextModelTest::test_generate_compilation_all_outputs', 'trace': ""(line 317)  torch._dynamo.exc.Unsupported: isinstance(NestedUserFunctionVariable(), TorchInGraphFunctionVariable(<class 'torch.nn.parameter.Parameter'>)): can't determine type of NestedUserFunctionVariable()""}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3IntegrationTest::test_model_4b_batch', 'trace': ""(line 675)  AssertionError: Lists differ: ['use[374 chars]t scenes:\\n\\n*   **Image 1** shows a cow on a beach.\\n'] != ['use[374 chars]t scenes. \\n\\n*   **Image 1** shows a cow standing on a beach']""}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3IntegrationTest::test_model_4b_batch_crops', 'trace': ""(line 675)  AssertionError: Lists differ: ['use[251 chars]. The sky is blue with some white clouds. It’s[405 chars]h a'] != ['use[251 chars]. There are clouds in the blue sky above.', 'u[398 chars]h a']""}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3IntegrationTest::test_model_4b_bf16', 'trace': ""(line 675)  AssertionError: Lists differ: ['use[154 chars]each next to a turquoise ocean. There are some[16 chars]lue'] != ['use[154 chars]each with turquoise water and a distant coastl[28 chars]oks']""}], 'multi': [{'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3ModelTest::test_generate_compilation_all_outputs', 'trace': ""(line 317)  torch._dynamo.exc.Unsupported: isinstance(NestedUserFunctionVariable(), TorchInGraphFunctionVariable(<class 'torch.nn.parameter.Parameter'>)): can't determine type of NestedUserFunctionVariable()""}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3ModelTest::test_sdpa_padding_matches_padding_free_with_position_ids', 'trace': '(line 4204)  AssertionError: Tensor-likes are not close!'}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3Vision2TextModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3Vision2TextModelTest::test_generate_compilation_all_outputs', 'trace': ""(line 317)  torch._dynamo.exc.Unsupported: isinstance(NestedUserFunctionVariable(), TorchInGraphFunctionVariable(<class 'torch.nn.parameter.Parameter'>)): can't determine type of NestedUserFunctionVariable()""}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3IntegrationTest::test_model_4b_batch', 'trace': ""(line 675)  AssertionError: Lists differ: ['use[374 chars]t scenes:\\n\\n*   **Image 1** shows a cow on a beach.\\n'] != ['use[374 chars]t scenes. \\n\\n*   **Image 1** shows a cow standing on a beach']""}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3IntegrationTest::test_model_4b_batch_crops', 'trace': ""(line 675)  AssertionError: Lists differ: ['use[251 chars]. The sky is blue with some white clouds. It’s[405 chars]h a'] != ['use[251 chars]. There are clouds in the blue sky above.', 'u[398 chars]h a']""}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3IntegrationTest::test_model_4b_bf16', 'trace': ""(line 675)  AssertionError: Lists differ: ['use[154 chars]each next to a turquoise ocean. There are some[16 chars]lue'] != ['use[154 chars]each with turquoise water and a distant coastl[28 chars]oks']""}]}","{'single': [{'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3ModelTest::test_sdpa_padding_matches_padding_free_with_position_ids', 'trace': '(line 4216)  AssertionError: Tensor-likes are not equal!'}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3Vision2TextModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3IntegrationTest::test_export_text_only_with_hybrid_cache', 'trace': ""(line 1642)  torch._dynamo.exc.TorchRuntimeError: Dynamo failed to run FX node with fake tensors: call_function <built-in function scaled_dot_product_attention>(*(FakeTensor(..., size=(1, 4, 1, 256), grad_fn=<AddBackward0>), FakeTensor(..., size=(1, 4, 4096, 256), grad_fn=<CloneBackward0>), FakeTensor(..., size=(1, 4, 4096, 256), grad_fn=<CloneBackward0>)), **{'attn_mask': FakeTensor(..., size=(1, 1, 1, 512), dtype=torch.bool), 'dropout_p': 0.0, 'scale': 0.0625, 'is_causal': False}): got RuntimeError('Attempting to broadcast a dimension of length 512 at -1! Mismatching argument at index 1 had torch.Size([1, 1, 1, 512]); but expected shape should be broadcastable to [1, 4, 1, 4096]')""}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3IntegrationTest::test_generation_beyond_sliding_window_1_sdpa', 'trace': '(line 81)  RuntimeError: The expanded size of the tensor (4826) must match the existing size (4807) at non-singleton dimension 3.  Target sizes: [2, 4, 4807, 4826].  Tensor sizes: [2, 1, 4807, 4807]'}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3IntegrationTest::test_generation_beyond_sliding_window_2_eager', 'trace': '(line 265)  RuntimeError: The size of tensor a (4826) must match the size of tensor b (4807) at non-singleton dimension 3'}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3IntegrationTest::test_model_4b_batch_crops', 'trace': '(line 81)  RuntimeError: The expanded size of the tensor (1646) must match the existing size (1617) at non-singleton dimension 3.  Target sizes: [2, 8, 1617, 1646].  Tensor sizes: [2, 1, 1617, 1617]'}], 'multi': [{'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3ModelTest::test_sdpa_padding_matches_padding_free_with_position_ids', 'trace': '(line 4219)  AssertionError: Tensor-likes are not close!'}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3Vision2TextModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3Vision2TextModelTest::test_model_parallelism', 'trace': '(line 925)  RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:1 and cuda:0!'}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3IntegrationTest::test_export_text_only_with_hybrid_cache', 'trace': ""(line 1642)  torch._dynamo.exc.TorchRuntimeError: Dynamo failed to run FX node with fake tensors: call_function <built-in function scaled_dot_product_attention>(*(FakeTensor(..., size=(1, 4, 1, 256), grad_fn=<AddBackward0>), FakeTensor(..., size=(1, 4, 4096, 256), grad_fn=<CloneBackward0>), FakeTensor(..., size=(1, 4, 4096, 256), grad_fn=<CloneBackward0>)), **{'attn_mask': FakeTensor(..., size=(1, 1, 1, 512), dtype=torch.bool), 'dropout_p': 0.0, 'scale': 0.0625, 'is_causal': False}): got RuntimeError('Attempting to broadcast a dimension of length 512 at -1! Mismatching argument at index 1 had torch.Size([1, 1, 1, 512]); but expected shape should be broadcastable to [1, 4, 1, 4096]')""}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3IntegrationTest::test_generation_beyond_sliding_window_1_sdpa', 'trace': '(line 81)  RuntimeError: The expanded size of the tensor (4826) must match the existing size (4807) at non-singleton dimension 3.  Target sizes: [2, 4, 4807, 4826].  Tensor sizes: [2, 1, 4807, 4807]'}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3IntegrationTest::test_generation_beyond_sliding_window_2_eager', 'trace': '(line 265)  RuntimeError: The size of tensor a (4826) must match the size of tensor b (4807) at non-singleton dimension 3'}, {'line': 'tests/models/gemma3/test_modeling_gemma3.py::Gemma3IntegrationTest::test_model_4b_batch_crops', 'trace': '(line 81)  RuntimeError: The expanded size of the tensor (1646) must match the existing size (1617) at non-singleton dimension 3.  Target sizes: [2, 8, 1617, 1646].  Tensor sizes: [2, 1, 1617, 1617]'}]}","{'single': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447501046', 'multi': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447501545'}","{'single': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526563053', 'multi': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526562857'}"
+gemma3n,0,286,0,2,0,1,{},"{'multi': [{'line': 'tests/models/gemma3n/test_modeling_gemma3n.py::Gemma3nTextModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}, {'line': 'tests/models/gemma3n/test_modeling_gemma3n.py::Gemma3nTextModelTest::test_multi_gpu_data_parallel_forward', 'trace': ""(line 1305)  AttributeError: 'DynamicCache' object has no attribute 'layers'""}], 'single': [{'line': 'tests/models/gemma3n/test_modeling_gemma3n.py::Gemma3nTextModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}]}","{'single': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447501047', 'multi': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447501538'}","{'multi': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526562955', 'single': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526563061'}"
+got_ocr2,145,254,2,2,2,1,"{'multi': [{'line': 'tests/models/got_ocr2/test_modeling_got_ocr2.py::GotOcr2ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}, {'line': 'tests/models/got_ocr2/test_modeling_got_ocr2.py::GotOcr2ModelTest::test_generate_compilation_all_outputs', 'trace': ""(line 317)  torch._dynamo.exc.Unsupported: isinstance(NestedUserFunctionVariable(), TorchInGraphFunctionVariable(<class 'torch.nn.parameter.Parameter'>)): can't determine type of NestedUserFunctionVariable()""}], 'single': [{'line': 'tests/models/got_ocr2/test_modeling_got_ocr2.py::GotOcr2ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}, {'line': 'tests/models/got_ocr2/test_modeling_got_ocr2.py::GotOcr2ModelTest::test_generate_compilation_all_outputs', 'trace': ""(line 317)  torch._dynamo.exc.Unsupported: isinstance(NestedUserFunctionVariable(), TorchInGraphFunctionVariable(<class 'torch.nn.parameter.Parameter'>)): can't determine type of NestedUserFunctionVariable()""}]}","{'multi': [{'line': 'tests/models/got_ocr2/test_modeling_got_ocr2.py::GotOcr2ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}, {'line': 'tests/models/got_ocr2/test_modeling_got_ocr2.py::GotOcr2ModelTest::test_multi_gpu_data_parallel_forward', 'trace': ""(line 1305)  AttributeError: 'DynamicCache' object has no attribute 'layers'""}], 'single': [{'line': 'tests/models/got_ocr2/test_modeling_got_ocr2.py::GotOcr2ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}]}","{'multi': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447501556', 'single': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447501063'}","{'multi': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526562995', 'single': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526563212'}"
+gpt2,249,487,1,1,1,1,"{'single': [{'line': 'tests/models/gpt2/test_modeling_gpt2.py::GPT2ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}], 'multi': [{'line': 'tests/models/gpt2/test_modeling_gpt2.py::GPT2ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}]}","{'multi': [{'line': 'tests/models/gpt2/test_modeling_gpt2.py::GPT2ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}], 'single': [{'line': 'tests/models/gpt2/test_modeling_gpt2.py::GPT2ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}]}","{'single': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447501087', 'multi': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447501566'}","{'multi': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526563001', 'single': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526563255'}"
+internvl,249,356,4,3,4,2,"{'single': [{'line': 'tests/models/internvl/test_modeling_internvl.py::InternVLModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}, {'line': 'tests/models/internvl/test_modeling_internvl.py::InternVLModelTest::test_generate_compilation_all_outputs', 'trace': ""(line 317)  torch._dynamo.exc.Unsupported: isinstance(NestedUserFunctionVariable(), TorchInGraphFunctionVariable(<class 'torch.nn.parameter.Parameter'>)): can't determine type of NestedUserFunctionVariable()""}, {'line': 'tests/models/internvl/test_modeling_internvl.py::InternVLLlamaIntegrationTest::test_llama_small_model_integration_forward', 'trace': '(line 687)  AssertionError: False is not true : Actual logits: tensor([ -9.8828,  -0.5005,   1.4697, -10.3438, -10.3438], dtype=torch.float16)'}, {'line': 'tests/models/internvl/test_modeling_internvl.py::InternVLLlamaIntegrationTest::test_llama_small_model_integration_interleaved_images_videos', 'trace': ""(line 675)  AssertionError: 'user[118 chars]nse. Upon closer inspection, the differences b[31 chars]. **' != 'user[118 chars]nse. After re-examining the images, I can see [13 chars]e no'""}], 'multi': [{'line': 'tests/models/internvl/test_modeling_internvl.py::InternVLModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}, {'line': 'tests/models/internvl/test_modeling_internvl.py::InternVLModelTest::test_generate_compilation_all_outputs', 'trace': ""(line 317)  torch._dynamo.exc.Unsupported: isinstance(NestedUserFunctionVariable(), TorchInGraphFunctionVariable(<class 'torch.nn.parameter.Parameter'>)): can't determine type of NestedUserFunctionVariable()""}, {'line': 'tests/models/internvl/test_modeling_internvl.py::InternVLLlamaIntegrationTest::test_llama_small_model_integration_forward', 'trace': '(line 687)  AssertionError: False is not true : Actual logits: tensor([ -9.8828,  -0.5005,   1.4697, -10.3438, -10.3438], dtype=torch.float16)'}, {'line': 'tests/models/internvl/test_modeling_internvl.py::InternVLLlamaIntegrationTest::test_llama_small_model_integration_interleaved_images_videos', 'trace': ""(line 675)  AssertionError: 'user[118 chars]nse. Upon closer inspection, the differences b[31 chars]. **' != 'user[118 chars]nse. After re-examining the images, I can see [13 chars]e no'""}]}","{'multi': [{'line': 'tests/models/internvl/test_modeling_internvl.py::InternVLModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}, {'line': 'tests/models/internvl/test_modeling_internvl.py::InternVLModelTest::test_flex_attention_with_grads', 'trace': '(line 439)  torch._inductor.exc.InductorError: RuntimeError: No valid triton configs. OutOfResources: out of resource: shared memory, Required: 106496, Hardware limit: 101376. Reducing block sizes or `num_stages` may help.'}, {'line': 'tests/models/internvl/test_modeling_internvl.py::InternVLModelTest::test_multi_gpu_data_parallel_forward', 'trace': ""(line 1305)  AttributeError: 'DynamicCache' object has no attribute 'layers'""}], 'single': [{'line': 'tests/models/internvl/test_modeling_internvl.py::InternVLModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}, {'line': 'tests/models/internvl/test_modeling_internvl.py::InternVLModelTest::test_flex_attention_with_grads', 'trace': '(line 439)  torch._inductor.exc.InductorError: RuntimeError: No valid triton configs. OutOfResources: out of resource: shared memory, Required: 106496, Hardware limit: 101376. Reducing block sizes or `num_stages` may help.'}]}","{'single': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447501143', 'multi': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447501636'}","{'multi': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526563553', 'single': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526563712'}"
+llama,229,478,4,2,4,1,"{'multi': [{'line': 'tests/models/llama/test_modeling_llama.py::LlamaModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}, {'line': 'tests/models/llama/test_modeling_llama.py::LlamaModelTest::test_generate_compilation_all_outputs', 'trace': ""(line 317)  torch._dynamo.exc.Unsupported: isinstance(NestedUserFunctionVariable(), TorchInGraphFunctionVariable(<class 'torch.nn.parameter.Parameter'>)): can't determine type of NestedUserFunctionVariable()""}, {'line': 'tests/models/llama/test_modeling_llama.py::LlamaModelTest::test_torch_compile_for_training', 'trace': '(line 951)  AssertionError: expected size 2==2, stride 20==64 at dim=0; expected size 2==2, stride 10==32 at dim=1; expected size 10==32, stride 1==1 at dim=2'}, {'line': 'tests/models/llama/test_modeling_llama.py::LlamaIntegrationTest::test_model_7b_logits_bf16', 'trace': '(line 687)  AssertionError: False is not true'}], 'single': [{'line': 'tests/models/llama/test_modeling_llama.py::LlamaModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}, {'line': 'tests/models/llama/test_modeling_llama.py::LlamaModelTest::test_generate_compilation_all_outputs', 'trace': ""(line 317)  torch._dynamo.exc.Unsupported: isinstance(NestedUserFunctionVariable(), TorchInGraphFunctionVariable(<class 'torch.nn.parameter.Parameter'>)): can't determine type of NestedUserFunctionVariable()""}, {'line': 'tests/models/llama/test_modeling_llama.py::LlamaModelTest::test_torch_compile_for_training', 'trace': '(line 951)  AssertionError: expected size 2==2, stride 20==64 at dim=0; expected size 2==2, stride 10==32 at dim=1; expected size 10==32, stride 1==1 at dim=2'}, {'line': 'tests/models/llama/test_modeling_llama.py::LlamaIntegrationTest::test_model_7b_logits_bf16', 'trace': '(line 687)  AssertionError: False is not true'}]}","{'multi': [{'line': 'tests/models/llama/test_modeling_llama.py::LlamaModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}, {'line': 'tests/models/llama/test_modeling_llama.py::LlamaModelTest::test_multi_gpu_data_parallel_forward', 'trace': ""(line 1305)  AttributeError: 'DynamicCache' object has no attribute 'layers'""}], 'single': [{'line': 'tests/models/llama/test_modeling_llama.py::LlamaModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}]}","{'multi': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447501675', 'single': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447501165'}","{'multi': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526563871', 'single': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526564103'}"
+llava,201,346,5,4,4,3,"{'single': [{'line': 'tests/models/llava/test_modeling_llava.py::LlavaForConditionalGenerationModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}, {'line': 'tests/models/llava/test_modeling_llava.py::LlavaForConditionalGenerationModelTest::test_flex_attention_with_grads', 'trace': '(line 687)  AssertionError: False is not true'}, {'line': 'tests/models/llava/test_modeling_llava.py::LlavaForConditionalGenerationModelTest::test_generate_compilation_all_outputs', 'trace': ""(line 317)  torch._dynamo.exc.Unsupported: isinstance(NestedUserFunctionVariable(), TorchInGraphFunctionVariable(<class 'torch.nn.parameter.Parameter'>)): can't determine type of NestedUserFunctionVariable()""}, {'line': 'tests/models/llava/test_modeling_llava.py::LlavaForConditionalGenerationIntegrationTest::test_batched_generation', 'trace': '(line 548)  importlib.metadata.PackageNotFoundError: No package metadata was found for bitsandbytes'}], 'multi': [{'line': 'tests/models/llava/test_modeling_llava.py::LlavaForConditionalGenerationModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}, {'line': 'tests/models/llava/test_modeling_llava.py::LlavaForConditionalGenerationModelTest::test_flex_attention_with_grads', 'trace': '(line 687)  AssertionError: False is not true'}, {'line': 'tests/models/llava/test_modeling_llava.py::LlavaForConditionalGenerationModelTest::test_generate_compilation_all_outputs', 'trace': ""(line 317)  torch._dynamo.exc.Unsupported: isinstance(NestedUserFunctionVariable(), TorchInGraphFunctionVariable(<class 'torch.nn.parameter.Parameter'>)): can't determine type of NestedUserFunctionVariable()""}, {'line': 'tests/models/llava/test_modeling_llava.py::LlavaForConditionalGenerationModelTest::test_sdpa_padding_matches_padding_free_with_position_ids', 'trace': '(line 4182)  IndexError: The shape of the mask [3, 23] at index 1 does not match the shape of the indexed tensor [3, 3, 8, 8] at index 1'}, {'line': 'tests/models/llava/test_modeling_llava.py::LlavaForConditionalGenerationIntegrationTest::test_batched_generation', 'trace': '(line 548)  importlib.metadata.PackageNotFoundError: No package metadata was found for bitsandbytes'}]}","{'multi': [{'line': 'tests/models/llava/test_modeling_llava.py::LlavaForConditionalGenerationModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}, {'line': 'tests/models/llava/test_modeling_llava.py::LlavaForConditionalGenerationModelTest::test_flex_attention_with_grads', 'trace': '(line 687)  AssertionError: False is not true'}, {'line': 'tests/models/llava/test_modeling_llava.py::LlavaForConditionalGenerationModelTest::test_multi_gpu_data_parallel_forward', 'trace': ""(line 1305)  AttributeError: 'DynamicCache' object has no attribute 'layers'""}, {'line': 'tests/models/llava/test_modeling_llava.py::LlavaForConditionalGenerationModelTest::test_sdpa_padding_matches_padding_free_with_position_ids', 'trace': '(line 4197)  IndexError: The shape of the mask [3, 23] at index 1 does not match the shape of the indexed tensor [3, 3, 8, 8] at index 1'}], 'single': [{'line': 'tests/models/llava/test_modeling_llava.py::LlavaForConditionalGenerationModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}, {'line': 'tests/models/llava/test_modeling_llava.py::LlavaForConditionalGenerationModelTest::test_flex_attention_with_grads', 'trace': '(line 687)  AssertionError: False is not true'}, {'line': 'tests/models/llava/test_modeling_llava.py::LlavaForConditionalGenerationModelTest::test_sdpa_padding_matches_padding_free_with_position_ids', 'trace': '(line 4197)  IndexError: The shape of the mask [3, 23] at index 1 does not match the shape of the indexed tensor [3, 3, 8, 8] at index 1'}]}","{'single': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447501186', 'multi': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447501727'}","{'multi': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526564002', 'single': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526564108'}"
+mistral3,197,286,3,2,3,1,"{'multi': [{'line': 'tests/models/mistral3/test_modeling_mistral3.py::Mistral3ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}, {'line': 'tests/models/mistral3/test_modeling_mistral3.py::Mistral3ModelTest::test_generate_compilation_all_outputs', 'trace': ""(line 317)  torch._dynamo.exc.Unsupported: isinstance(NestedUserFunctionVariable(), TorchInGraphFunctionVariable(<class 'torch.nn.parameter.Parameter'>)): can't determine type of NestedUserFunctionVariable()""}, {'line': 'tests/models/mistral3/test_modeling_mistral3.py::Mistral3IntegrationTest::test_mistral3_integration_batched_generate', 'trace': '(line 675)  AssertionError: \'Calm waters reflect\\nWooden path to distant shore\\nSilence in the scene\' != ""Wooden path to calm,\\nReflections whisper secrets,\\nNature\'s peace unfolds.""'}], 'single': [{'line': 'tests/models/mistral3/test_modeling_mistral3.py::Mistral3ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}, {'line': 'tests/models/mistral3/test_modeling_mistral3.py::Mistral3ModelTest::test_generate_compilation_all_outputs', 'trace': ""(line 317)  torch._dynamo.exc.Unsupported: isinstance(NestedUserFunctionVariable(), TorchInGraphFunctionVariable(<class 'torch.nn.parameter.Parameter'>)): can't determine type of NestedUserFunctionVariable()""}, {'line': 'tests/models/mistral3/test_modeling_mistral3.py::Mistral3IntegrationTest::test_mistral3_integration_batched_generate', 'trace': '(line 675)  AssertionError: \'Calm waters reflect\\nWooden path to distant shore\\nSilence in the scene\' != ""Wooden path to calm,\\nReflections whisper secrets,\\nNature\'s peace unfolds.""'}]}","{'single': [{'line': 'tests/models/mistral3/test_modeling_mistral3.py::Mistral3ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}], 'multi': [{'line': 'tests/models/mistral3/test_modeling_mistral3.py::Mistral3ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}, {'line': 'tests/models/mistral3/test_modeling_mistral3.py::Mistral3ModelTest::test_multi_gpu_data_parallel_forward', 'trace': ""(line 1305)  AttributeError: 'DynamicCache' object has no attribute 'layers'""}]}","{'multi': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447500305', 'single': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447499780'}","{'single': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526561480', 'multi': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526561618'}"
+modernbert,132,164,5,5,5,5,"{'single': [{'line': 'tests/models/modernbert/test_modeling_modernbert.py::ModernBertModelIntegrationTest::test_export', 'trace': ""(line 675)  AssertionError: Lists differ: ['mechanic', 'lawyer', 'teacher', 'waiter', 'doctor'] != ['lawyer', 'mechanic', 'teacher', 'doctor', 'waiter']""}, {'line': 'tests/models/modernbert/test_modeling_modernbert.py::ModernBertModelIntegrationTest::test_inference_masked_lm', 'trace': '(line 401)  AssertionError: Tensor-likes are not close!'}, {'line': 'tests/models/modernbert/test_modeling_modernbert.py::ModernBertModelIntegrationTest::test_inference_no_head', 'trace': '(line 423)  AssertionError: Tensor-likes are not close!'}, {'line': 'tests/models/modernbert/test_modeling_modernbert.py::ModernBertModelIntegrationTest::test_inference_sequence_classification', 'trace': '(line 469)  AssertionError: Tensor-likes are not close!'}, {'line': 'tests/models/modernbert/test_modeling_modernbert.py::ModernBertModelIntegrationTest::test_inference_token_classification', 'trace': '(line 446)  AssertionError: Tensor-likes are not close!'}], 'multi': [{'line': 'tests/models/modernbert/test_modeling_modernbert.py::ModernBertModelIntegrationTest::test_export', 'trace': ""(line 675)  AssertionError: Lists differ: ['mechanic', 'lawyer', 'teacher', 'waiter', 'doctor'] != ['lawyer', 'mechanic', 'teacher', 'doctor', 'waiter']""}, {'line': 'tests/models/modernbert/test_modeling_modernbert.py::ModernBertModelIntegrationTest::test_inference_masked_lm', 'trace': '(line 401)  AssertionError: Tensor-likes are not close!'}, {'line': 'tests/models/modernbert/test_modeling_modernbert.py::ModernBertModelIntegrationTest::test_inference_no_head', 'trace': '(line 423)  AssertionError: Tensor-likes are not close!'}, {'line': 'tests/models/modernbert/test_modeling_modernbert.py::ModernBertModelIntegrationTest::test_inference_sequence_classification', 'trace': '(line 469)  AssertionError: Tensor-likes are not close!'}, {'line': 'tests/models/modernbert/test_modeling_modernbert.py::ModernBertModelIntegrationTest::test_inference_token_classification', 'trace': '(line 446)  AssertionError: Tensor-likes are not close!'}]}","{'multi': [{'line': 'tests/models/modernbert/test_modeling_modernbert.py::ModernBertModelIntegrationTest::test_export', 'trace': ""(line 675)  AssertionError: Lists differ: ['mechanic', 'lawyer', 'teacher', 'waiter', 'doctor'] != ['lawyer', 'mechanic', 'teacher', 'doctor', 'waiter']""}, {'line': 'tests/models/modernbert/test_modeling_modernbert.py::ModernBertModelIntegrationTest::test_inference_masked_lm', 'trace': '(line 401)  AssertionError: Tensor-likes are not close!'}, {'line': 'tests/models/modernbert/test_modeling_modernbert.py::ModernBertModelIntegrationTest::test_inference_no_head', 'trace': '(line 423)  AssertionError: Tensor-likes are not close!'}, {'line': 'tests/models/modernbert/test_modeling_modernbert.py::ModernBertModelIntegrationTest::test_inference_sequence_classification', 'trace': '(line 469)  AssertionError: Tensor-likes are not close!'}, {'line': 'tests/models/modernbert/test_modeling_modernbert.py::ModernBertModelIntegrationTest::test_inference_token_classification', 'trace': '(line 446)  AssertionError: Tensor-likes are not close!'}], 'single': [{'line': 'tests/models/modernbert/test_modeling_modernbert.py::ModernBertModelIntegrationTest::test_export', 'trace': ""(line 675)  AssertionError: Lists differ: ['mechanic', 'lawyer', 'teacher', 'waiter', 'doctor'] != ['lawyer', 'mechanic', 'teacher', 'doctor', 'waiter']""}, {'line': 'tests/models/modernbert/test_modeling_modernbert.py::ModernBertModelIntegrationTest::test_inference_masked_lm', 'trace': '(line 401)  AssertionError: Tensor-likes are not close!'}, {'line': 'tests/models/modernbert/test_modeling_modernbert.py::ModernBertModelIntegrationTest::test_inference_no_head', 'trace': '(line 423)  AssertionError: Tensor-likes are not close!'}, {'line': 'tests/models/modernbert/test_modeling_modernbert.py::ModernBertModelIntegrationTest::test_inference_sequence_classification', 'trace': '(line 469)  AssertionError: Tensor-likes are not close!'}, {'line': 'tests/models/modernbert/test_modeling_modernbert.py::ModernBertModelIntegrationTest::test_inference_token_classification', 'trace': '(line 446)  AssertionError: Tensor-likes are not close!'}]}","{'single': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447499811', 'multi': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447500326'}","{'multi': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526561668', 'single': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526561515'}"
+qwen2,213,438,3,3,3,2,"{'multi': [{'line': 'tests/models/qwen2/test_modeling_qwen2.py::Qwen2ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}, {'line': 'tests/models/qwen2/test_modeling_qwen2.py::Qwen2ModelTest::test_generate_compilation_all_outputs', 'trace': ""(line 317)  torch._dynamo.exc.Unsupported: isinstance(NestedUserFunctionVariable(), TorchInGraphFunctionVariable(<class 'torch.nn.parameter.Parameter'>)): can't determine type of NestedUserFunctionVariable()""}, {'line': 'tests/models/qwen2/test_modeling_qwen2.py::Qwen2IntegrationTest::test_export_static_cache', 'trace': ""(line 1638)  torch._dynamo.exc.TorchRuntimeError: Failed running call_method index_copy_(*(FakeTensor(..., size=(1, 2, 26, 64), dtype=torch.bfloat16), 2, FakeTensor(..., device='cuda:0', size=(1,), dtype=torch.int64), FakeTensor(..., device='cuda:0', size=(1, 2, 1, 64), dtype=torch.bfloat16,""}], 'single': [{'line': 'tests/models/qwen2/test_modeling_qwen2.py::Qwen2ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}, {'line': 'tests/models/qwen2/test_modeling_qwen2.py::Qwen2ModelTest::test_generate_compilation_all_outputs', 'trace': ""(line 317)  torch._dynamo.exc.Unsupported: isinstance(NestedUserFunctionVariable(), TorchInGraphFunctionVariable(<class 'torch.nn.parameter.Parameter'>)): can't determine type of NestedUserFunctionVariable()""}, {'line': 'tests/models/qwen2/test_modeling_qwen2.py::Qwen2IntegrationTest::test_export_static_cache', 'trace': ""(line 1638)  torch._dynamo.exc.TorchRuntimeError: Failed running call_method index_copy_(*(FakeTensor(..., size=(1, 2, 26, 64), dtype=torch.bfloat16), 2, FakeTensor(..., device='cuda:0', size=(1,), dtype=torch.int64), FakeTensor(..., device='cuda:0', size=(1, 2, 1, 64), dtype=torch.bfloat16,""}]}","{'multi': [{'line': 'tests/models/qwen2/test_modeling_qwen2.py::Qwen2ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}, {'line': 'tests/models/qwen2/test_modeling_qwen2.py::Qwen2ModelTest::test_multi_gpu_data_parallel_forward', 'trace': ""(line 1305)  AttributeError: 'DynamicCache' object has no attribute 'layers'""}, {'line': 'tests/models/qwen2/test_modeling_qwen2.py::Qwen2IntegrationTest::test_export_static_cache', 'trace': ""(line 1642)  torch._dynamo.exc.TorchRuntimeError: Dynamo failed to run FX node with fake tensors: call_method index_copy_(*(FakeTensor(..., size=(1, 2, 26, 64), dtype=torch.bfloat16), 2, FakeTensor(..., device='cuda:0', size=(1,), dtype=torch.int64), FakeTensor(..., device='cuda:0', size=(1, 2, 1, 64), dtype=torch.bfloat16,""}], 'single': [{'line': 'tests/models/qwen2/test_modeling_qwen2.py::Qwen2ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}, {'line': 'tests/models/qwen2/test_modeling_qwen2.py::Qwen2IntegrationTest::test_export_static_cache', 'trace': ""(line 1642)  torch._dynamo.exc.TorchRuntimeError: Dynamo failed to run FX node with fake tensors: call_method index_copy_(*(FakeTensor(..., size=(1, 2, 26, 64), dtype=torch.bfloat16), 2, FakeTensor(..., device='cuda:0', size=(1,), dtype=torch.int64), FakeTensor(..., device='cuda:0', size=(1, 2, 1, 64), dtype=torch.bfloat16,""}]}","{'multi': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447500458', 'single': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447499989'}","{'multi': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526562376', 'single': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526562270'}"
+qwen2_5_omni,168,277,2,5,1,1,"{'single': [{'line': 'tests/models/qwen2_5_omni/test_modeling_qwen2_5_omni.py::Qwen2_5OmniModelIntegrationTest::test_small_model_integration_test_batch', 'trace': '(line 675)  AssertionError: Lists differ: [""sys[96 chars]ant\\nsystem\\nYou are a helpful assistant.\\nuse[129 chars]er.""] != [""sys[96 chars]ant\\nThe sound is glass shattering, and the do[198 chars]er.""]'}], 'multi': [{'line': 'tests/models/qwen2_5_omni/test_modeling_qwen2_5_omni.py::Qwen2_5OmniThinkerForConditionalGenerationModelTest::test_model_parallelism', 'trace': '(line 675)  AssertionError: Items in the second set but not the first:'}, {'line': 'tests/models/qwen2_5_omni/test_modeling_qwen2_5_omni.py::Qwen2_5OmniModelIntegrationTest::test_small_model_integration_test_batch', 'trace': '(line 675)  AssertionError: Lists differ: [""sys[96 chars]ant\\nsystem\\nYou are a helpful assistant.\\nuse[129 chars]er.""] != [""sys[96 chars]ant\\nThe sound is glass shattering, and the do[198 chars]er.""]'}]}","{'multi': [{'line': 'tests/models/qwen2_5_omni/test_modeling_qwen2_5_omni.py::Qwen2_5OmniThinkerForConditionalGenerationModelTest::test_model_parallelism', 'trace': '(line 675)  AssertionError: Items in the second set but not the first:'}, {'line': 'tests/models/qwen2_5_omni/test_modeling_qwen2_5_omni.py::Qwen2_5OmniThinkerForConditionalGenerationModelTest::test_multi_gpu_data_parallel_forward', 'trace': ""(line 1305)  AttributeError: 'DynamicCache' object has no attribute 'layers'""}, {'line': 'tests/models/qwen2_5_omni/test_modeling_qwen2_5_omni.py::Qwen2_5OmniModelIntegrationTest::test_small_model_integration_test_batch', 'trace': '(line 675)  AssertionError: Lists differ: [""sys[96 chars]ant\\nsystem\\nYou are a helpful assistant.\\nuse[129 chars]er.""] != [""sys[96 chars]ant\\nThe sound is glass shattering, and the do[198 chars]er.""]'}, {'line': 'tests/models/qwen2_5_omni/test_modeling_qwen2_5_omni.py::Qwen2_5OmniModelIntegrationTest::test_small_model_integration_test_multiturn', 'trace': '(line 849)  torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 20.00 MiB. GPU 1 has a total capacity of 22.18 GiB of which 6.50 MiB is free. Process 51940 has 22.17 GiB memory in use. Of the allocated memory 21.74 GiB is allocated by PyTorch, and 27.83 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation.  See documentation for Memory Management  (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)'}, {'line': 'tests/models/qwen2_5_omni/test_modeling_qwen2_5_omni.py::Qwen2_5OmniModelIntegrationTest::test_small_model_integration_test_w_audio', 'trace': '(line 1000)  torch.OutOfMemoryError: CUDA out of memory. Tried to allocate 20.00 MiB. GPU 1 has a total capacity of 22.18 GiB of which 8.50 MiB is free. Process 51940 has 22.17 GiB memory in use. Of the allocated memory 21.75 GiB is allocated by PyTorch, and 17.78 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation.  See documentation for Memory Management  (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables)'}], 'single': [{'line': 'tests/models/qwen2_5_omni/test_modeling_qwen2_5_omni.py::Qwen2_5OmniModelIntegrationTest::test_small_model_integration_test_batch', 'trace': '(line 675)  AssertionError: Lists differ: [""sys[96 chars]ant\\nsystem\\nYou are a helpful assistant.\\nuse[129 chars]er.""] != [""sys[96 chars]ant\\nThe sound is glass shattering, and the do[198 chars]er.""]'}]}","{'single': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447499993', 'multi': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447500491'}","{'multi': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526562375', 'single': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526562289'}"
+qwen2_5_vl,204,311,1,1,2,1,"{'single': [{'line': 'tests/models/qwen2_5_vl/test_modeling_qwen2_5_vl.py::Qwen2_5_VLIntegrationTest::test_small_model_integration_test', 'trace': ""(line 700)  requests.exceptions.ConnectionError: HTTPSConnectionPool(host='qianwen-res.oss-accelerate-overseas.aliyuncs.com', port=443): Max retries exceeded with url: /Qwen2-VL/demo_small.jpg (Caused by NewConnectionError('<urllib3.connection.HTTPSConnection object at 0x7b289312aad0>: Failed to establish a new connection: [Errno -2] Name or service not known'))""}, {'line': 'tests/models/qwen2_5_vl/test_modeling_qwen2_5_vl.py::Qwen2_5_VLIntegrationTest::test_small_model_integration_test_batch_different_resolutions', 'trace': ""(line 675)  AssertionError: Lists differ: ['sys[314 chars]ion\\n addCriterion\\n\\n addCriterion\\n\\n addCri[75 chars]n\\n'] != ['sys[314 chars]ion\\nThe dog in the picture appears to be a La[81 chars] is']""}], 'multi': [{'line': 'tests/models/qwen2_5_vl/test_modeling_qwen2_5_vl.py::Qwen2_5_VLIntegrationTest::test_small_model_integration_test_batch_different_resolutions', 'trace': ""(line 675)  AssertionError: Lists differ: ['sys[314 chars]ion\\n addCriterion\\n\\n addCriterion\\n\\n addCri[75 chars]n\\n'] != ['sys[314 chars]ion\\nThe dog in the picture appears to be a La[81 chars] is']""}]}","{'multi': [{'line': 'tests/models/qwen2_5_vl/test_modeling_qwen2_5_vl.py::Qwen2_5_VLIntegrationTest::test_small_model_integration_test_batch_different_resolutions', 'trace': ""(line 675)  AssertionError: Lists differ: ['sys[314 chars]ion\\n addCriterion\\n\\n addCriterion\\n\\n addCri[75 chars]n\\n'] != ['sys[314 chars]ion\\nThe dog in the picture appears to be a La[81 chars] is']""}], 'single': [{'line': 'tests/models/qwen2_5_vl/test_modeling_qwen2_5_vl.py::Qwen2_5_VLIntegrationTest::test_small_model_integration_test_batch_different_resolutions', 'trace': ""(line 675)  AssertionError: Lists differ: ['sys[314 chars]ion\\n addCriterion\\n\\n addCriterion\\n\\n addCri[75 chars]n\\n'] != ['sys[314 chars]ion\\nThe dog in the picture appears to be a La[81 chars] is']""}]}","{'single': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447499984', 'multi': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447500447'}","{'multi': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526562382', 'single': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526562290'}"
+smolvlm,323,499,1,1,1,1,"{'multi': [{'line': 'tests/models/smolvlm/test_modeling_smolvlm.py::SmolVLMForConditionalGenerationModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}], 'single': [{'line': 'tests/models/smolvlm/test_modeling_smolvlm.py::SmolVLMForConditionalGenerationModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}]}","{'single': [{'line': 'tests/models/smolvlm/test_modeling_smolvlm.py::SmolVLMForConditionalGenerationModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}], 'multi': [{'line': 'tests/models/smolvlm/test_modeling_smolvlm.py::SmolVLMForConditionalGenerationModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}]}","{'multi': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447500533', 'single': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447500052'}","{'single': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526562675', 'multi': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526562798'}"
+t5,254,592,4,3,3,2,"{'multi': [{'line': 'tests/models/t5/test_modeling_t5.py::T5ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}, {'line': 'tests/models/t5/test_modeling_t5.py::T5ModelTest::test_multi_gpu_data_parallel_forward', 'trace': ""(line 130)  TypeError: EncoderDecoderCache.__init__() missing 1 required positional argument: 'cross_attention_cache'""}, {'line': 'tests/models/t5/test_modeling_t5.py::T5ModelIntegrationTests::test_export_t5_summarization', 'trace': ""(line 885)  torch._dynamo.exc.TorchRuntimeError: Failed running call_function <built-in function add>(*(FakeTensor(..., size=(1, 8, 1, 1234)), FakeTensor(..., device='cuda:1', size=(1, 1, 1, 1234))), **{}):""}, {'line': 'tests/models/t5/test_modeling_t5.py::T5ModelIntegrationTests::test_small_integration_test', 'trace': '(line 687)  AssertionError: False is not true'}], 'single': [{'line': 'tests/models/t5/test_modeling_t5.py::T5ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4125)  KeyError: 'eager'""}, {'line': 'tests/models/t5/test_modeling_t5.py::T5ModelIntegrationTests::test_export_t5_summarization', 'trace': ""(line 885)  torch._dynamo.exc.TorchRuntimeError: Failed running call_function <built-in function add>(*(FakeTensor(..., size=(1, 8, 1, 1234)), FakeTensor(..., device='cuda:0', size=(1, 1, 1, 1234))), **{}):""}, {'line': 'tests/models/t5/test_modeling_t5.py::T5ModelIntegrationTests::test_small_integration_test', 'trace': '(line 687)  AssertionError: False is not true'}]}","{'multi': [{'line': 'tests/models/t5/test_modeling_t5.py::T5ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}, {'line': 'tests/models/t5/test_modeling_t5.py::T5ModelTest::test_multi_gpu_data_parallel_forward', 'trace': ""(line 131)  TypeError: EncoderDecoderCache.__init__() missing 1 required positional argument: 'cross_attention_cache'""}, {'line': 'tests/models/t5/test_modeling_t5.py::T5ModelIntegrationTests::test_export_t5_summarization', 'trace': ""(line 687)  AttributeError: 'dict' object has no attribute 'batch_size'""}], 'single': [{'line': 'tests/models/t5/test_modeling_t5.py::T5ModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}, {'line': 'tests/models/t5/test_modeling_t5.py::T5ModelIntegrationTests::test_export_t5_summarization', 'trace': ""(line 687)  AttributeError: 'dict' object has no attribute 'batch_size'""}]}","{'multi': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447500560', 'single': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447500103'}","{'multi': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526563047', 'single': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526562939'}"
+vit,135,217,0,0,0,0,{},{},"{'multi': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447500654', 'single': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447500177'}","{'multi': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526563537', 'single': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526563397'}"
+wav2vec2,0,672,0,4,0,4,{},"{'multi': [{'line': 'tests/models/wav2vec2/test_modeling_wav2vec2.py::Wav2Vec2ModelIntegrationTest::test_inference_mms_1b_all', 'trace': '(line 989)  RuntimeError: Dataset scripts are no longer supported, but found common_voice_11_0.py'}, {'line': 'tests/models/wav2vec2/test_modeling_wav2vec2.py::Wav2Vec2ModelIntegrationTest::test_wav2vec2_with_lm', 'trace': '(line 989)  RuntimeError: Dataset scripts are no longer supported, but found common_voice_11_0.py'}, {'line': 'tests/models/wav2vec2/test_modeling_wav2vec2.py::Wav2Vec2ModelIntegrationTest::test_wav2vec2_with_lm_invalid_pool', 'trace': '(line 675)  AssertionError: Traceback (most recent call last):'}, {'line': 'tests/models/wav2vec2/test_modeling_wav2vec2.py::Wav2Vec2ModelIntegrationTest::test_wav2vec2_with_lm_pool', 'trace': '(line 989)  RuntimeError: Dataset scripts are no longer supported, but found common_voice_11_0.py'}], 'single': [{'line': 'tests/models/wav2vec2/test_modeling_wav2vec2.py::Wav2Vec2ModelIntegrationTest::test_inference_mms_1b_all', 'trace': '(line 989)  RuntimeError: Dataset scripts are no longer supported, but found common_voice_11_0.py'}, {'line': 'tests/models/wav2vec2/test_modeling_wav2vec2.py::Wav2Vec2ModelIntegrationTest::test_wav2vec2_with_lm', 'trace': '(line 989)  RuntimeError: Dataset scripts are no longer supported, but found common_voice_11_0.py'}, {'line': 'tests/models/wav2vec2/test_modeling_wav2vec2.py::Wav2Vec2ModelIntegrationTest::test_wav2vec2_with_lm_invalid_pool', 'trace': '(line 675)  AssertionError: Traceback (most recent call last):'}, {'line': 'tests/models/wav2vec2/test_modeling_wav2vec2.py::Wav2Vec2ModelIntegrationTest::test_wav2vec2_with_lm_pool', 'trace': '(line 989)  RuntimeError: Dataset scripts are no longer supported, but found common_voice_11_0.py'}]}","{'multi': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447500676', 'single': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447500194'}","{'multi': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526563711', 'single': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526563582'}"
+whisper,0,1010,0,11,0,8,{},"{'single': [{'line': 'tests/models/whisper/test_modeling_whisper.py::WhisperModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}, {'line': 'tests/models/whisper/test_modeling_whisper.py::WhisperModelIntegrationTests::test_large_batched_generation_multilingual', 'trace': '(line 756)  RuntimeError: The frame has 0 channels, expected 1. If you are hitting this, it may be because you are using a buggy FFmpeg version. FFmpeg4 is known to fail here in some valid scenarios. Try to upgrade FFmpeg?'}, {'line': 'tests/models/whisper/test_modeling_whisper.py::WhisperModelIntegrationTests::test_small_longform_timestamps_generation', 'trace': '(line 756)  RuntimeError: The frame has 0 channels, expected 1. If you are hitting this, it may be because you are using a buggy FFmpeg version. FFmpeg4 is known to fail here in some valid scenarios. Try to upgrade FFmpeg?'}, {'line': 'tests/models/whisper/test_modeling_whisper.py::WhisperModelIntegrationTests::test_tiny_longform_timestamps_generation', 'trace': '(line 756)  RuntimeError: The frame has 0 channels, expected 1. If you are hitting this, it may be because you are using a buggy FFmpeg version. FFmpeg4 is known to fail here in some valid scenarios. Try to upgrade FFmpeg?'}, {'line': 'tests/models/whisper/test_modeling_whisper.py::WhisperModelIntegrationTests::test_whisper_longform_multi_batch_hard', 'trace': '(line 675)  AssertionError: Lists differ: ["" Fo[272 chars]ting of classics, Sicilian, nade door variatio[8147 chars]le!\'] != ["" Fo[272 chars]ting a classic Sicilian, nade door variation o[8150 chars]le!\']'}, {'line': 'tests/models/whisper/test_modeling_whisper.py::WhisperModelIntegrationTests::test_whisper_longform_multi_batch_hard_prev_cond', 'trace': '(line 675)  AssertionError: Lists differ: ["" Fo[422 chars]to a fisher shows in lip-nitsky attack that cu[7903 chars]le!""] != ["" Fo[422 chars]to a Fisher shows in lip-nitsky attack that cu[7918 chars]le.""]'}, {'line': 'tests/models/whisper/test_modeling_whisper.py::WhisperModelIntegrationTests::test_whisper_shortform_single_batch_prev_cond', 'trace': '(line 675)  AssertionError: Lists differ: ["" Fo[268 chars]ating, so soft, it would make JD power and her[196 chars]ke.""] != ["" Fo[268 chars]ating so soft, it would make JD power and her [195 chars]ke.""]'}, {'line': 'tests/models/whisper/test_modeling_whisper.py::WhisperStandaloneDecoderModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}], 'multi': [{'line': 'tests/models/whisper/test_modeling_whisper.py::WhisperModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}, {'line': 'tests/models/whisper/test_modeling_whisper.py::WhisperModelTest::test_multi_gpu_data_parallel_forward', 'trace': ""(line 131)  TypeError: EncoderDecoderCache.__init__() missing 1 required positional argument: 'cross_attention_cache'""}, {'line': 'tests/models/whisper/test_modeling_whisper.py::WhisperModelIntegrationTests::test_generate_with_forced_decoder_ids', 'trace': '(line 713)  requests.exceptions.ReadTimeout: (ReadTimeoutError(""HTTPSConnectionPool(host=\'huggingface.co\', port=443): Read timed out. (read timeout=10)""), \'(Request ID: 13cb0b08-c261-4ca3-a58f-91a2f3e327ed)\')'}, {'line': 'tests/models/whisper/test_modeling_whisper.py::WhisperModelIntegrationTests::test_large_batched_generation_multilingual', 'trace': '(line 756)  RuntimeError: The frame has 0 channels, expected 1. If you are hitting this, it may be because you are using a buggy FFmpeg version. FFmpeg4 is known to fail here in some valid scenarios. Try to upgrade FFmpeg?'}, {'line': 'tests/models/whisper/test_modeling_whisper.py::WhisperModelIntegrationTests::test_small_longform_timestamps_generation', 'trace': '(line 756)  RuntimeError: The frame has 0 channels, expected 1. If you are hitting this, it may be because you are using a buggy FFmpeg version. FFmpeg4 is known to fail here in some valid scenarios. Try to upgrade FFmpeg?'}, {'line': 'tests/models/whisper/test_modeling_whisper.py::WhisperModelIntegrationTests::test_tiny_longform_timestamps_generation', 'trace': '(line 756)  RuntimeError: The frame has 0 channels, expected 1. If you are hitting this, it may be because you are using a buggy FFmpeg version. FFmpeg4 is known to fail here in some valid scenarios. Try to upgrade FFmpeg?'}, {'line': 'tests/models/whisper/test_modeling_whisper.py::WhisperModelIntegrationTests::test_whisper_longform_multi_batch_hard', 'trace': '(line 675)  AssertionError: Lists differ: ["" Fo[272 chars]ting of classics, Sicilian, nade door variatio[8147 chars]le!\'] != ["" Fo[272 chars]ting a classic Sicilian, nade door variation o[8150 chars]le!\']'}, {'line': 'tests/models/whisper/test_modeling_whisper.py::WhisperModelIntegrationTests::test_whisper_longform_multi_batch_hard_prev_cond', 'trace': '(line 675)  AssertionError: Lists differ: ["" Fo[422 chars]to a fisher shows in lip-nitsky attack that cu[7903 chars]le!""] != ["" Fo[422 chars]to a Fisher shows in lip-nitsky attack that cu[7918 chars]le.""]'}, {'line': 'tests/models/whisper/test_modeling_whisper.py::WhisperModelIntegrationTests::test_whisper_shortform_single_batch_prev_cond', 'trace': '(line 675)  AssertionError: Lists differ: ["" Fo[268 chars]ating, so soft, it would make JD power and her[196 chars]ke.""] != ["" Fo[268 chars]ating so soft, it would make JD power and her [195 chars]ke.""]'}, {'line': 'tests/models/whisper/test_modeling_whisper.py::WhisperStandaloneDecoderModelTest::test_eager_padding_matches_padding_free_with_position_ids', 'trace': ""(line 4140)  KeyError: 'eager'""}, {'line': 'tests/models/whisper/test_modeling_whisper.py::WhisperStandaloneDecoderModelTest::test_multi_gpu_data_parallel_forward', 'trace': ""(line 1305)  AttributeError: 'DynamicCache' object has no attribute 'layers'""}]}","{'multi': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447500690', 'single': 'https://github.com/huggingface/transformers/actions/runs/16433423306/job/46447500204'}","{'single': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526563737', 'multi': 'https://github.com/huggingface/transformers/actions/runs/16460401119/job/46526563862'}"

summary_page.py

@@ -0,0 +1,138 @@
+import matplotlib.pyplot as plt
+import pandas as pd
+
+def create_summary_page(df: pd.DataFrame, available_models: list[str]) -> plt.Figure:
+    """Create a summary page with model names and both AMD/NVIDIA test stats bars."""
+    if df.empty:
+        fig, ax = plt.subplots(figsize=(16, 8), facecolor='#000000')
+        ax.set_facecolor('#000000')
+        ax.text(0.5, 0.5, 'No data available', 
+                horizontalalignment='center', verticalalignment='center',
+                transform=ax.transAxes, fontsize=20, color='#888888',
+                fontfamily='monospace', weight='normal')
+        ax.axis('off')
+        return fig
+    
+    fig, ax = plt.subplots(figsize=(16, len(available_models) * 2.5 + 2), facecolor='#000000')
+    ax.set_facecolor('#000000')
+    
+    colors = {
+        'passed': '#4CAF50',
+        'failed': '#E53E3E', 
+        'skipped': '#FFD54F',
+        'error': '#8B0000',
+        'empty': "#5B5B5B"
+    }
+    
+    visible_model_count = 0
+    max_y = 0
+    
+    for i, model_name in enumerate(available_models):
+        if model_name not in df.index:
+            continue
+            
+        row = df.loc[model_name]
+        
+        # Get values directly from dataframe
+        success_amd = int(row.get('success_amd', 0)) if pd.notna(row.get('success_amd', 0)) else 0
+        success_nvidia = int(row.get('success_nvidia', 0)) if pd.notna(row.get('success_nvidia', 0)) else 0
+        failed_multi_amd = int(row.get('failed_multi_no_amd', 0)) if pd.notna(row.get('failed_multi_no_amd', 0)) else 0
+        failed_multi_nvidia = int(row.get('failed_multi_no_nvidia', 0)) if pd.notna(row.get('failed_multi_no_nvidia', 0)) else 0
+        failed_single_amd = int(row.get('failed_single_no_amd', 0)) if pd.notna(row.get('failed_single_no_amd', 0)) else 0
+        failed_single_nvidia = int(row.get('failed_single_no_nvidia', 0)) if pd.notna(row.get('failed_single_no_nvidia', 0)) else 0
+        
+        # Calculate stats
+        amd_stats = {
+            'passed': success_amd,
+            'failed': failed_multi_amd + failed_single_amd,
+            'skipped': 0,
+            'error': 0
+        }
+        
+        nvidia_stats = {
+            'passed': success_nvidia,
+            'failed': failed_multi_nvidia + failed_single_nvidia,
+            'skipped': 0,
+            'error': 0
+        }
+        
+        amd_total = sum(amd_stats.values())
+        nvidia_total = sum(nvidia_stats.values())
+        
+        if amd_total == 0 and nvidia_total == 0:
+            continue
+            
+        # Position for this model - use visible model count for spacing
+        y_base = (2.2 + visible_model_count) * 1.8
+        y_model_name = y_base    # Model name above AMD bar
+        y_amd_bar = y_base + 0.45       # AMD bar
+        y_nvidia_bar = y_base + 0.97    # NVIDIA bar
+        max_y = max(max_y, y_nvidia_bar + 0.5)
+        
+        # Model name centered above the AMD bar
+        left_0 = 8
+        bar_length = 92
+        ax.text(bar_length / 2 + left_0, y_model_name, f"{i} - {y_model_name} - {model_name.lower()}", 
+               ha='center', va='center', color='#FFFFFF', 
+               fontsize=20, fontfamily='monospace', fontweight='bold')
+        
+        # AMD label and bar on the same level
+        ax.text(left_0 - 2, y_amd_bar, "amd", ha='right', va='center', color='#CCCCCC', fontsize=18, fontfamily='monospace', fontweight='normal')
+        if amd_total > 0:           
+            # AMD bar starts after labels
+            left = left_0
+            for category in ['passed', 'failed', 'skipped', 'error']:
+                if amd_stats[category] > 0:
+                    width = amd_stats[category] / amd_total * bar_length
+                    ax.barh(y_amd_bar, width, left=left, height=0.405, 
+                           color=colors[category], alpha=0.9)
+                    if width > 4:
+                        ax.text(left + width/2, y_amd_bar, str(amd_stats[category]), 
+                               ha='center', va='center', color='black', 
+                               fontweight='bold', fontsize=12, fontfamily='monospace')
+                    left += width
+        else:
+            ax.barh(y_amd_bar, bar_length, left=left_0, height=0.405, color=colors['empty'], alpha=0.9)
+            ax.text(left_0 + bar_length/2, y_amd_bar, "No data", ha='center', va='center', color='black', fontweight='bold', fontsize=12, fontfamily='monospace')
+        
+        # NVIDIA label and bar on the same level
+        ax.text(left_0 - 2, y_nvidia_bar, "nvidia", ha='right', va='center', color='#CCCCCC', fontsize=18, fontfamily='monospace', fontweight='normal')
+
+        if nvidia_total > 0:            
+            # NVIDIA bar starts after labels
+            left = left_0
+            for category in ['passed', 'failed', 'skipped', 'error']:
+                if nvidia_stats[category] > 0:
+                    width = nvidia_stats[category] / nvidia_total * bar_length
+                    ax.barh(y_nvidia_bar, width, left=left, height=0.405, 
+                           color=colors[category], alpha=0.9)
+                    if width > 4:
+                        ax.text(left + width/2, y_nvidia_bar, str(nvidia_stats[category]), 
+                               ha='center', va='center', color='black', 
+                               fontweight='bold', fontsize=12, fontfamily='monospace')
+                    left += width
+        else:
+            ax.barh(y_nvidia_bar, bar_length, left=left_0, height=0.405, color=colors['empty'], alpha=0.9)
+            ax.text(left_0 + bar_length/2, y_nvidia_bar, "No data", ha='center', va='center', color='black', fontweight='bold', fontsize=12, fontfamily='monospace')
+        
+        # Increment counter for next visible model
+        visible_model_count += 1
+    
+    # Style the axes to be completely invisible and span full width
+    ax.set_xlim(0, 100)
+    ax.set_ylim(0, max_y)
+    ax.set_xlabel('')
+    ax.set_ylabel('')
+    ax.spines['bottom'].set_visible(False)
+    ax.spines['left'].set_visible(False)
+    ax.spines['top'].set_visible(False)
+    ax.spines['right'].set_visible(False)
+    ax.set_xticks([])
+    ax.set_yticks([])
+    ax.yaxis.set_inverted(True)
+    
+    # Remove all margins to make bars span full width
+    plt.tight_layout()
+    plt.subplots_adjust(left=0.02, right=0.98, top=0.98, bottom=0.02)
+    
+    return fig

utils.py

@@ -0,0 +1,51 @@
+import logging
+import sys
+from datetime import datetime
+
+
+class TimestampFormatter(logging.Formatter):
+    """Custom formatter that matches the existing timestamp format used in print statements."""
+    
+    def format(self, record):
+        # Create timestamp in the same format as existing print statements
+        timestamp = datetime.now().strftime('%Y-%m-%d %H:%M:%S')
+        
+        # Format the message with timestamp prefix
+        if record.levelno == logging.WARNING:
+            return f"WARNING: {record.getMessage()}"
+        elif record.levelno == logging.ERROR:
+            return f"Error {record.getMessage()}"
+        else:
+            return f"[{timestamp}] {record.getMessage()}"
+
+
+def setup_logger(name="tcid", level=logging.INFO):
+    """Set up logger with custom timestamp formatting to match existing print format."""
+    logger = logging.getLogger(name)
+    
+    # Avoid adding multiple handlers if logger already exists
+    if logger.handlers:
+        return logger
+    
+    logger.setLevel(level)
+    
+    # Create console handler
+    handler = logging.StreamHandler(sys.stdout)
+    handler.setLevel(level)
+    
+    # Set custom formatter
+    formatter = TimestampFormatter()
+    handler.setFormatter(formatter)
+    
+    logger.addHandler(handler)
+    
+    return logger
+
+
+# Create default logger instance
+logger = setup_logger()
+
+
+
+def generate_underlined_line(text: str) -> str:
+    return text + "\n" + "─" * len(text) + "\n"