CERBERUS/ollama
1
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-07-05 23:31:01 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions
ollama/x/models/nn
History
Exact
Patrick Devine 82e0ddb6fe
mlxrunner: harden linear/embedding layers against over-promotion (#16682) 
Adding/Multiplying a tensor by a scalar w/ a different data type
can cause the tensor to be promoted and cause performance issues.

This change adds several guards against over-promotion.
2026-06-11 13:56:25 -07:00
..
nn.go mlxrunner: harden linear/embedding layers against over-promotion (#16682) 2026-06-11 13:56:25 -07:00
nn_test.go mlx: allow the embedding layer to use the nvfp4 global scale (#16527) 2026-06-04 17:40:01 -07:00
recurrent.go nn/recurrent: return per-boundary states from the gated-delta kernels 2026-06-09 00:39:19 -07:00
recurrent_test.go nn/recurrent: return per-boundary states from the gated-delta kernels 2026-06-09 00:39:19 -07:00
rope.go mlxrunner: move YaRN RoPE helpers into x/models/nn 2026-05-22 09:32:09 -07:00
sdpa.go mlxrunner: decouple models from attention cache storage layout 2026-04-27 20:04:46 -07:00
sdpa_test.go mlxrunner: decouple models from attention cache storage layout 2026-04-27 20:04:46 -07:00