Downloads · 30 days
0
WanTopic-V1-Minimal
ShihteSiao/WanTopic-V1-Minimal
WanTopic-V1-Minimal is a machine learning model from ShihteSiao. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
- Author: ShihteSiao - Version: v1.4-minimal - Architecture: Hierarchical Reasoning Model (HRM) - Minimal 256d - Parameters: 5,552,386 (5.6M) - Memory Usage: 0.0GB
Access
Public
Updated Sep 1, 2025
Repo size
45.7 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.pth22.9 MB · 100%
At a glance
- Access
- Public
- Created
- Sep 1, 2025
- Updated
- Sep 1, 2025
- SHA
- 4a5d5cc3
Model card
From the Hugging Face model README
萬題v1 (WanTopic-V1-Minimal)
Model Description
- Author: ShihteSiao
- Version: v1.4-minimal
- Architecture: Hierarchical Reasoning Model (HRM) - Minimal 256d
- Parameters: 5,552,386 (5.6M)
- Memory Usage: 0.0GB
Performance
- Test Accuracy: 50.0%
- Best Validation Loss: 0.0025
- Peak Reasoning Efficiency: 49.3%
- Max Reasoning Steps: 12
Architecture Details
- Hidden Size: 256
- Attention Heads: 4
- High-level Layers: 4
- Low-level Layers: 4
- Total Layers: 8
Training Configuration
- Batch Size: 4 × 8 = 32
- Learning Rate: 0.0002
- Epochs: 40
- Mixed Precision: True
- Label Smoothing: 0.0
Usage
import torch
# Load model
checkpoint = torch.load('WanTopic-V1-Stable_final_model.pth')
# Initialize model with checkpoint['model_config']
# Load state dict: model.load_state_dict(checkpoint['model_state_dict'])
Generated on: 2025-09-01 14:57:19
- Created
- Sep 1, 2025
- Updated
- Sep 1, 2025