Training UNet3D

[jla524]

Changes made:

• Implement iterator for training set
• Implement transformations
• Implement dice cross entropy loss

TODO:

• Support batch size > 1
• Implement LR warmup
• Run up to 4000 epochs
• Train the model (currently running out of memory on M1 Pro)
• Validate results (goal is 0.908 Mean DICE score)

[geohot]

Bounty locked, good luck!

[tinyb0t]

Changes made in tinygrad/:

------------------------------------------------------------
files                             insertions       deletions
------------------------------------------------------------
tinygrad/tensor.py                         1               1
------------------------------------------------------------
lines added in the tinygrad folder: 0

[geohot]

No update in a week