Block a user
tacit
synced commits to refs/pull/510/merge at tacit/nanochat from mirror
2026-02-10 20:49:58 +00:00
e569b59f92
delete torchao dependency, create our own exact API-matched version of Float8Linear, document it very well. for some poorly understood reason, the performance is not only ~identical but actually runs 3% faster. despite of it being significantly simpler and much less code. i don't fully understand why/how atm
tacit
synced commits to refs/pull/509/merge at tacit/nanochat from mirror
2026-02-10 20:49:58 +00:00
e569b59f92
delete torchao dependency, create our own exact API-matched version of Float8Linear, document it very well. for some poorly understood reason, the performance is not only ~identical but actually runs 3% faster. despite of it being significantly simpler and much less code. i don't fully understand why/how atm
tacit
synced commits to refs/pull/501/merge at tacit/nanochat from mirror
2026-02-10 20:49:58 +00:00
e569b59f92
delete torchao dependency, create our own exact API-matched version of Float8Linear, document it very well. for some poorly understood reason, the performance is not only ~identical but actually runs 3% faster. despite of it being significantly simpler and much less code. i don't fully understand why/how atm
tacit
synced commits to refs/pull/489/merge at tacit/nanochat from mirror
2026-02-10 20:49:58 +00:00
e569b59f92
delete torchao dependency, create our own exact API-matched version of Float8Linear, document it very well. for some poorly understood reason, the performance is not only ~identical but actually runs 3% faster. despite of it being significantly simpler and much less code. i don't fully understand why/how atm
ff46300720
tune miniseries just a bit, fairly cosmetic, keep to even depths where the math works out nicely in model sizing
tacit
synced commits to refs/pull/485/merge at tacit/nanochat from mirror
2026-02-10 20:49:58 +00:00
e569b59f92
delete torchao dependency, create our own exact API-matched version of Float8Linear, document it very well. for some poorly understood reason, the performance is not only ~identical but actually runs 3% faster. despite of it being significantly simpler and much less code. i don't fully understand why/how atm
tacit
synced commits to refs/pull/492/merge at tacit/nanochat from mirror
2026-02-10 20:49:58 +00:00
e569b59f92
delete torchao dependency, create our own exact API-matched version of Float8Linear, document it very well. for some poorly understood reason, the performance is not only ~identical but actually runs 3% faster. despite of it being significantly simpler and much less code. i don't fully understand why/how atm
tacit
synced commits to refs/pull/486/merge at tacit/nanochat from mirror
2026-02-10 20:49:58 +00:00
e569b59f92
delete torchao dependency, create our own exact API-matched version of Float8Linear, document it very well. for some poorly understood reason, the performance is not only ~identical but actually runs 3% faster. despite of it being significantly simpler and much less code. i don't fully understand why/how atm
tacit
synced commits to refs/pull/442/merge at tacit/nanochat from mirror
2026-02-10 20:49:57 +00:00
e569b59f92
delete torchao dependency, create our own exact API-matched version of Float8Linear, document it very well. for some poorly understood reason, the performance is not only ~identical but actually runs 3% faster. despite of it being significantly simpler and much less code. i don't fully understand why/how atm
ff46300720
tune miniseries just a bit, fairly cosmetic, keep to even depths where the math works out nicely in model sizing
aeff095e97
better comments/flow on all the hyperparameter transfer stuff, and change the WD scaling from my empirical 1/d^2 to a bit more principled version based on Tepoch. All of that theory is based on AdamW and could be suboptimal for Muon
e569b59f92
delete torchao dependency, create our own exact API-matched version of Float8Linear, document it very well. for some poorly understood reason, the performance is not only ~identical but actually runs 3% faster. despite of it being significantly simpler and much less code. i don't fully understand why/how atm
tacit
synced commits to refs/pull/437/merge at tacit/nanochat from mirror
2026-02-10 20:49:57 +00:00
e569b59f92
delete torchao dependency, create our own exact API-matched version of Float8Linear, document it very well. for some poorly understood reason, the performance is not only ~identical but actually runs 3% faster. despite of it being significantly simpler and much less code. i don't fully understand why/how atm
ff46300720
tune miniseries just a bit, fairly cosmetic, keep to even depths where the math works out nicely in model sizing
aeff095e97
better comments/flow on all the hyperparameter transfer stuff, and change the WD scaling from my empirical 1/d^2 to a bit more principled version based on Tepoch. All of that theory is based on AdamW and could be suboptimal for Muon
tacit
synced commits to refs/pull/483/merge at tacit/nanochat from mirror
2026-02-10 20:49:57 +00:00
e569b59f92
delete torchao dependency, create our own exact API-matched version of Float8Linear, document it very well. for some poorly understood reason, the performance is not only ~identical but actually runs 3% faster. despite of it being significantly simpler and much less code. i don't fully understand why/how atm
tacit
synced commits to refs/pull/425/merge at tacit/nanochat from mirror
2026-02-10 04:29:51 +00:00
ff46300720
tune miniseries just a bit, fairly cosmetic, keep to even depths where the math works out nicely in model sizing
aeff095e97
better comments/flow on all the hyperparameter transfer stuff, and change the WD scaling from my empirical 1/d^2 to a bit more principled version based on Tepoch. All of that theory is based on AdamW and could be suboptimal for Muon
ff46300720
tune miniseries just a bit, fairly cosmetic, keep to even depths where the math works out nicely in model sizing
aeff095e97
better comments/flow on all the hyperparameter transfer stuff, and change the WD scaling from my empirical 1/d^2 to a bit more principled version based on Tepoch. All of that theory is based on AdamW and could be suboptimal for Muon
ff46300720
tune miniseries just a bit, fairly cosmetic, keep to even depths where the math works out nicely in model sizing
aeff095e97
better comments/flow on all the hyperparameter transfer stuff, and change the WD scaling from my empirical 1/d^2 to a bit more principled version based on Tepoch. All of that theory is based on AdamW and could be suboptimal for Muon
tacit
synced commits to refs/pull/483/merge at tacit/nanochat from mirror
2026-02-10 04:29:51 +00:00
tacit
synced commits to refs/pull/141/merge at tacit/nanochat from mirror
2026-02-10 04:29:50 +00:00
ff46300720
tune miniseries just a bit, fairly cosmetic, keep to even depths where the math works out nicely in model sizing
ff46300720
tune miniseries just a bit, fairly cosmetic, keep to even depths where the math works out nicely in model sizing
aeff095e97
better comments/flow on all the hyperparameter transfer stuff, and change the WD scaling from my empirical 1/d^2 to a bit more principled version based on Tepoch. All of that theory is based on AdamW and could be suboptimal for Muon
tacit
synced commits to refs/pull/510/merge at tacit/nanochat from mirror
2026-02-09 20:19:56 +00:00
ff46300720
tune miniseries just a bit, fairly cosmetic, keep to even depths where the math works out nicely in model sizing
aeff095e97
better comments/flow on all the hyperparameter transfer stuff, and change the WD scaling from my empirical 1/d^2 to a bit more principled version based on Tepoch. All of that theory is based on AdamW and could be suboptimal for Muon
tacit
synced commits to refs/pull/485/merge at tacit/nanochat from mirror
2026-02-09 20:19:56 +00:00
ff46300720
tune miniseries just a bit, fairly cosmetic, keep to even depths where the math works out nicely in model sizing
aeff095e97
better comments/flow on all the hyperparameter transfer stuff, and change the WD scaling from my empirical 1/d^2 to a bit more principled version based on Tepoch. All of that theory is based on AdamW and could be suboptimal for Muon
tacit
synced and deleted reference 2026-02-09 20:19:56 +00:00
refs/tags/refs/pull/518/merge at tacit/nanochat from mirror