branch: master
Commits on master
- e68aebf bring hip graph back (#2385) 2 years ago
- 46b05da simple display_name (#2416) 2 years ago
- 12023b6 onnx ops cleanup (#2413) 2 years ago
- 8f89e21 torch and numpy don't share ops anymore (#2412) 2 years ago
- 193be14 that had bugs, force an order (#2411) 2 years ago
- 65f4e69 beautiful_mnist.py link 2 years ago
- 1b3b8de update readme examples 2 years ago
- 5bb720a Cocoa is no longer used 2 years ago
- 095e2ce add name support to fetch (#2407) 2 years ago
- 397c093 fix wait in jit (#2408) 2 years ago
- b927942 Move HIP render logic to its dedicated place (#2394) 2 years ago
- 6d67278 Update Whisper to use fetch helper (#2401) 2 years ago
- 0505c5e remove force_wait, refactor to graph (#2405) 2 years ago
- c5d585e Fix Triton README broken link (#2406) 2 years ago
- b27c845 minor cleanup for View strides (#2404) 2 years ago
- 64aa2f4 clean up to_shape_strides (#2402) 2 years ago
- e4026dc don't pass lazybuffer to rawbuffer (#2400) 2 years ago
- aefa97a Remove runtime imports in realize (#2157) 2 years ago
- 4f8f0ac minor cleanups, remove dead files (#2398) 2 years ago
- 66c75f3 remove triton (#2396) 2 years ago
- 8656eeb jit doesn't use named tensors (#2393) 2 years ago
- 80e4ad8 faster get_recursive_parents (#2392) 2 years ago
- 8798d12 autopad shapetracker for BEAM (#2375) 2 years ago
- 162db46 hotfix: fix hip WMMA casting hack 2 years ago
- 6ceecc9 hotfix: scalar 2 years ago
- 0eda545 dtypes.float.vec(sz) (#2386) 2 years ago
- cbb8486 ResNet training changes (update benchmark) (#2390) 2 years ago
- 2dec869 hotfix: default remains gen 1 llama 2 years ago
- 7f0cc4a whisper: support audio >30s (#2378) 2 years ago
- 7220f5c fixed hf convert and now it's working with tinyllama (#2374) 2 years ago