tacit

tacit synced commits to refs/pull/31/head at tacit/nanochat from mirror 2025-10-22 14:22:14 +00:00

6641aeed1d FEAT: Allow CPU-only execution in compute_init

tacit synced commits to refs/pull/24/merge at tacit/nanochat from mirror 2025-10-22 14:22:14 +00:00

8e1274c18c Merge 33142f99d3 into 2e938530ce

2e938530ce delete spurious torch.empty allocation in adamw

a088b7a6ec use enable_gqa of pytorch sdpa, allows us to delete some code, didnt realize it's available

94ee507054 quick fix base eval due to fewshot requirement

49cd02f283 fix: remove unnecessary tensor allocation in DistAdamW optimizer

Compare 5 commits »

tacit synced commits to refs/pull/15/merge at tacit/nanochat from mirror 2025-10-22 14:22:13 +00:00

df311a0e62 Merge 6ef9f77789 into 2e938530ce

2e938530ce delete spurious torch.empty allocation in adamw

a088b7a6ec use enable_gqa of pytorch sdpa, allows us to delete some code, didnt realize it's available

94ee507054 quick fix base eval due to fewshot requirement

33e8a27f91 Merge karpathy/cpu-mps-dev , adding the ability to run on CPU, on MPS, or on CUDA, with autodetect. Gnarly PR, nonzero chance I broke something.

Compare 28 commits »

tacit synced commits to refs/pull/145/merge at tacit/nanochat from mirror 2025-10-22 14:22:13 +00:00

b9b91290bb Merge 67d76b834a into 50bea28ef9

67d76b834a tidy up and doc simplification

e83d633179 Add training continuation script and update MacOS guide

b81d789992 Pass device batch size to base_loss script

1225ddf00e Add macOS memory-optimized training and documentation

Compare 5 commits »

tacit synced commits to refs/pull/145/head at tacit/nanochat from mirror 2025-10-22 14:22:13 +00:00

67d76b834a tidy up and doc simplification

e83d633179 Add training continuation script and update MacOS guide

b81d789992 Pass device batch size to base_loss script

1225ddf00e Add macOS memory-optimized training and documentation

Compare 4 commits »

tacit synced commits to refs/pull/13/merge at tacit/nanochat from mirror 2025-10-22 14:22:13 +00:00

3d8ab1a771 Merge 07f193ab0d into 2e938530ce

2e938530ce delete spurious torch.empty allocation in adamw

a088b7a6ec use enable_gqa of pytorch sdpa, allows us to delete some code, didnt realize it's available

94ee507054 quick fix base eval due to fewshot requirement

33e8a27f91 Merge karpathy/cpu-mps-dev , adding the ability to run on CPU, on MPS, or on CUDA, with autodetect. Gnarly PR, nonzero chance I broke something.

Compare 25 commits »

tacit synced and deleted reference refs/tags/refs/pull/61/merge at tacit/nanochat from mirror 2025-10-22 14:22:13 +00:00

tacit synced and deleted reference refs/tags/refs/pull/50/merge at tacit/nanochat from mirror 2025-10-22 14:22:13 +00:00

tacit synced and deleted reference refs/tags/refs/pull/133/merge at tacit/nanochat from mirror 2025-10-22 14:22:13 +00:00

tacit synced and deleted reference refs/tags/refs/pull/132/merge at tacit/nanochat from mirror 2025-10-22 14:22:13 +00:00

tacit synced and deleted reference refs/tags/refs/pull/126/merge at tacit/nanochat from mirror 2025-10-22 14:22:13 +00:00

tacit synced commits to refs/pull/94/merge at tacit/nanochat from mirror 2025-10-22 06:12:15 +00:00

aaae1ec2a9 Merge 037706a330 into 2e938530ce

2e938530ce delete spurious torch.empty allocation in adamw

49cd02f283 fix: remove unnecessary tensor allocation in DistAdamW optimizer

Compare 3 commits »

tacit synced commits to refs/pull/91/head at tacit/nanochat from mirror 2025-10-22 06:12:14 +00:00

067298c51b reverted back

tacit synced commits to refs/pull/90/merge at tacit/nanochat from mirror 2025-10-22 06:12:14 +00:00

ece5fea007 Merge 1a428f2b0b into 2e938530ce

2e938530ce delete spurious torch.empty allocation in adamw

a088b7a6ec use enable_gqa of pytorch sdpa, allows us to delete some code, didnt realize it's available

94ee507054 quick fix base eval due to fewshot requirement

33e8a27f91 Merge karpathy/cpu-mps-dev , adding the ability to run on CPU, on MPS, or on CUDA, with autodetect. Gnarly PR, nonzero chance I broke something.

Compare 28 commits »

tacit synced commits to refs/pull/64/merge at tacit/nanochat from mirror 2025-10-22 06:12:14 +00:00

7004d6dd15 Merge 0493c0b8c5 into 2e938530ce

2e938530ce delete spurious torch.empty allocation in adamw

a088b7a6ec use enable_gqa of pytorch sdpa, allows us to delete some code, didnt realize it's available

94ee507054 quick fix base eval due to fewshot requirement

33e8a27f91 Merge karpathy/cpu-mps-dev , adding the ability to run on CPU, on MPS, or on CUDA, with autodetect. Gnarly PR, nonzero chance I broke something.

Compare 28 commits »

tacit synced commits to refs/pull/46/merge at tacit/nanochat from mirror 2025-10-22 06:12:14 +00:00

11ac06079d Merge 5e8030ae16 into 2e938530ce

2e938530ce delete spurious torch.empty allocation in adamw

a088b7a6ec use enable_gqa of pytorch sdpa, allows us to delete some code, didnt realize it's available

94ee507054 quick fix base eval due to fewshot requirement

33e8a27f91 Merge karpathy/cpu-mps-dev , adding the ability to run on CPU, on MPS, or on CUDA, with autodetect. Gnarly PR, nonzero chance I broke something.

Compare 28 commits »

tacit synced commits to refs/pull/18/merge at tacit/nanochat from mirror 2025-10-22 06:12:14 +00:00

7271837622 Merge 49870bb540 into 2e938530ce

2e938530ce delete spurious torch.empty allocation in adamw

a088b7a6ec use enable_gqa of pytorch sdpa, allows us to delete some code, didnt realize it's available

94ee507054 quick fix base eval due to fewshot requirement

33e8a27f91 Merge karpathy/cpu-mps-dev , adding the ability to run on CPU, on MPS, or on CUDA, with autodetect. Gnarly PR, nonzero chance I broke something.

Compare 28 commits »

tacit synced commits to refs/pull/142/merge at tacit/nanochat from mirror 2025-10-22 06:12:14 +00:00

97c9e36d92 Merge 7a52f9bfbb into 2e938530ce

7a52f9bfbb Updates lockfile with CPU package support without overwriting other architectures

760af62e11 Git ignore eval_bundle

901b075605 Fix GPU-less CPU use on Linux with specific Torch indexes

Compare 4 commits »

tacit synced commits to refs/pull/91/merge at tacit/nanochat from mirror 2025-10-22 06:12:14 +00:00

f143fc516e Merge 067298c51b into 2e938530ce

067298c51b reverted back

Compare 2 commits »

tacit synced commits to refs/pull/142/head at tacit/nanochat from mirror 2025-10-22 06:12:14 +00:00

7a52f9bfbb Updates lockfile with CPU package support without overwriting other architectures

760af62e11 Git ignore eval_bundle

901b075605 Fix GPU-less CPU use on Linux with specific Torch indexes

Compare 3 commits »