johnmarktaylor91 / torchlens
Inspect, visualize and intervene on any operation in any neural network. Capture every activation and gradient, the full computational graph and 800+ metadata fields without changing model code. Supports PyTorch, JAX, TensorFlow, MLX, PaddlePaddle and tinygrad. Validated on 6400+ models. Built for researchers and their AI agents.
- Repository stars
- 662All-time #51,985
- Fork
- 31
- Last push
- updated 1 days ago
- Repository created
- since 2022
About
Inspect, visualize and intervene on any operation in any neural network. Capture every activation and gradient, the full computational graph and 800+ metadata fields without changing model code. Supports PyTorch, JAX, TensorFlow, MLX, PaddlePaddle and tinygrad. Validated on 6400+ models. Built for researchers and their AI agents.Summarised by AI from the README; the original prevails
Author description: Inspect, visualize and intervene on any operation in any neural network. Capture every activation and gradient, the full computational graph and 800+ metadata fields without changing model code. Supports PyTorch, JAX, TensorFlow, MLX, PaddlePaddle and tinygrad. Validated on 6400+ models. Built for researchers and their AI agents.
Basic listing based on official public information; functionality has not been independently evaluated.
- Pricing
- Pricing not verifiedCheck current official software, hosting, model and infrastructure pricing.