mrq 6d42c9ae23 how foolish of me, not having a softmax as float32 (maybe addresses an emergent regression where bfloat16 training shits the bed where float16+loss scaling doesnt)		2025-04-07 22:51:52 -05:00
..
arch	diagnosed both hf/llama.cpp versions to probably just being a faulty export method (to-do: migrate vall_e.models.base to vall_e.export --hf)	2025-04-05 22:05:39 -05:00
__init__.py	nothing could go wrong part 2 (reverted and rewrote commits since there was a nasty regression)	2025-03-25 23:06:16 -05:00
ar_nar_v2.py	tweaks	2025-04-05 10:27:07 -05:00
ar_nar.py	tweaks	2025-04-05 10:27:07 -05:00
base_v2.py	how foolish of me, not having a softmax as float32 (maybe addresses an emergent regression where bfloat16 training shits the bed where float16+loss scaling doesnt)	2025-04-07 22:51:52 -05:00
base.py	diagnosed both hf/llama.cpp versions to probably just being a faulty export method (to-do: migrate vall_e.models.base to vall_e.export --hf)	2025-04-05 22:05:39 -05:00
lora.py	naive model offloading support (handles automatically splitting parts of the model to requested device per memory constraints, either inferred or requested in the yaml, input tensors are automatically migrated to the right device, it SEEMS to work for training under the test trainer when split between GPU and CPU) (this was specifically only because that Flux imagegen model released so I can test it there)	2024-08-01 20:12:06 -05:00