Skip to content

Commit 57d7a3e

Browse files
committed
[feat][gpu] Q4 quantization, Metal GPU shaders, ANE kernel fusion, memory safety
Made-with: Cursor
1 parent abc9fa3 commit 57d7a3e

8 files changed

Lines changed: 3339 additions & 257 deletions

File tree

.gitignore

Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -25,6 +25,9 @@ training/test_*
2525
# Inference binaries and runtime data
2626
inference/qwen_ane
2727
inference/qwen05b.bin
28+
inference/qwen05b_f32.bin
29+
inference/qwen05b_f16.bin
30+
inference/qwen05b_q8.bin
2831
inference/.venv/
2932
inference/benchmark_results.json
3033

@@ -59,6 +62,7 @@ web/
5962
training/tinystories_data00.bin
6063
training/ane_stories110M_ckpt.bin
6164
*.bin
65+
*.metallib
6266
!training/download_data.sh
6367

6468
# Secrets / env

0 commit comments

Comments
 (0)