Anime Image Tagger Development & Optimization
Budget: $15 – $25 USD
I'm reaching out because I need an experienced developer to help get my deep learning project fully functional and production-ready.
The project is an Anime Image Tagger system built with PyTorch that implements:
- Vision Transformer architecture (1B-3B parameters)
- Multi-teacher knowledge distillation
- Hierarchical multi-label classification for 200k tags
- HDF5-based data pipeline
- Progressive model scaling capabilities
- ONNX export and inference optimization
I have a comprehensive codebase already developed, but I need assistance with:
1. Debugging and ensuring all components work together seamlessly
2. Optimizing the training pipeline for efficiency
3. Resolving any dependency/configuration issues
4. Getting the full training-to-deployment pipeline operational
Could you please share:
- A link to your GitHub profile so I can review your previous work
- Your experience with similar computer vision/deep learning projects, particularly:
- Vision Transformers or large-scale image classification
- Multi-label classification systems
- PyTorch model training and optimization
- Working with large-scale datasets and HDF5
- Model deployment and inference optimization
The project uses PyTorch, Transformers, ONNX, FastAPI, and various scientific computing libraries. Experience with distributed training and GPU optimization would be a plus.
The project is an Anime Image Tagger system built with PyTorch that implements:
- Vision Transformer architecture (1B-3B parameters)
- Multi-teacher knowledge distillation
- Hierarchical multi-label classification for 200k tags
- HDF5-based data pipeline
- Progressive model scaling capabilities
- ONNX export and inference optimization
I have a comprehensive codebase already developed, but I need assistance with:
1. Debugging and ensuring all components work together seamlessly
2. Optimizing the training pipeline for efficiency
3. Resolving any dependency/configuration issues
4. Getting the full training-to-deployment pipeline operational
Could you please share:
- A link to your GitHub profile so I can review your previous work
- Your experience with similar computer vision/deep learning projects, particularly:
- Vision Transformers or large-scale image classification
- Multi-label classification systems
- PyTorch model training and optimization
- Working with large-scale datasets and HDF5
- Model deployment and inference optimization
The project uses PyTorch, Transformers, ONNX, FastAPI, and various scientific computing libraries. Experience with distributed training and GPU optimization would be a plus.