mirror of
https://github.com/tinygrad/tinygrad.git
synced 2026-06-24 02:14:17 +00:00
* two stage cumsum in tensor.py * 2 more kernels for llama cumsum * gpt-2 and llama use fast multinomial |
||
|---|---|---|
| .. | ||
| test_disk_cache.py | ||
| test_disk_tensor.py | ||
| test_flopcounter.py | ||
| test_helpers.py | ||
| test_shapetracker.py | ||
| test_shm_tensor.py | ||
| test_symbolic.py | ||