Metadata-Version: 2.4
Name: pluto-dl
Version: 0.0.1
Summary: Coming soon
Description-Content-Type: text/markdown

# pluto:

```
pluto/
│
├── main/
│   ├── tensors.py
│   │   Core Tensor object
│   │   - storage / shape / stride system
│   │   - versioning
│   │   - autograd metadata
│
│   ├── tensor_math.py
│   │   Elementwise math ops
│   │   - Add
│   │   - Sub
│   │   - Mul
│   │   - Div
│   │   - Pow
│   │   - Neg
│   │   - Dot
│   │   - MatMul
│
│   ├── tensor_reduction.py
│   │   Reduction operations
│   │   - Sum
│   │   - Mean
│   │   - Prod
│   │   - Max
│   │   - ArgMax
│
│   ├── tensor_view.py
│   │   Shape/view transformations
│   │   - view
│   │   - reshape
│   │   - permute
│   │   - transpose
│   │   - squeeze / unsqueeze
│   │   - flatten
│   │   - expand
│
│   ├── tensor_copy.py
│   │   Copy / memory ops
│   │   - clone
│   │   - contiguous
│   │   - repeat
│
│   ├── tensor_index.py
│   │   Indexing + slicing
│   │   - normalize_slice
│   │   - basic_slicing
│   │   - select
│   │   - narrow
│   │   - limited_advanced_slicing
│   │
│   ├── tensor_broadcast.py
│   │   Broadcasting helpers
│   │   - infer_broadcast_shape
│   │   - are_broadcastable
│   │   - infer_broadcast_backward_dims
│
│   └── tensor_binds.py
│       Operator bindings
│       - + - * / ** @
│       - tensor methods
│       - reductions
│       - view ops
│
├── autograd/
│   ├── function.py
│   │   Autograd engine
│   │   - Function.apply
│   │   - Context
│   │   - backward graph execution
│   │   - gradient accumulation
│
│   └── utils.py
│       Autograd helpers
│       - SavedTensor (version check)
│       - normalize_grad_shape
│       - broadcast gradient handling
│
├── nn/
│   ├──module.py
│   Base module class
│   - parameter registration
│   - buffers
│   - train / eval
│   - state_dict / load_state_dict
│
├── linear.py
│   Linear layer
│   - weight
│   - bias
│   - forward
│
├── embedding.py
│   Embedding layer
│   - weight matrix
│   - lookup via EmbeddingLookup
│   - forward pass for token indices
│
├── activation.py
│   Activation functions
│   - ReLU
│   - GELU
│   - SiLU (Swish)
│   - Softmax
│   - LogSoftmax
│
├── loss_func.py
│   Loss functions
│   - CrossEntropyLoss
│   - MSELoss
│   - BCEWithLogitsLoss
│
├── norm.py
│   Normalization layers
│   - BatchNorm1d
│   - BatchNorm2d
│   - LayerNorm
│
└── optim.py
    Optimizers
    - SGD (momentum + nesterov)
    - Adam
    - AdamW
    - gradient clipping utility
```
