branch: master
Commits on master
- d3e244d prev speed improvements (#5252) 1 year ago
- 21d41f0 nv follows HCQCompatAllocRes protocol (#5275) 1 year ago
- d3e4e21 add return type for HCQCompatAllocator _alloc (#5267) 1 year ago
- 191463a add timing to SDXL (#5273) 1 year ago
- b2c3a28 nn.RMSNorm (#5272) 1 year ago
- 9a2a82a test stable diffusion unet in ci (#5268) 1 year ago
- ce52b10 add a flag DISABLE_LOOP_COLLAPSE (#5270) 1 year ago
- e53b164 small changes from lowerer (#5266) 1 year ago
- 7be776f add _alloc_signal/_free_signal to hcq (#5264) 1 year ago
- 9a25ee0 pixed unet call params (#5262) 1 year ago
- 59bc837 refactor gated load rendering [run_process_replay] (#5259) 1 year ago
- e050603 nv close fds after mapping (#5246) 1 year ago
- d3cfb6c refactor UOps.LOAD barrier [run_process_replay] (#5258) 1 year ago
- a1044e6 iterate over scoped uops once [run_process_replay] (#5255) 1 year ago
- dfbee4f feat: add blobfile to testing (#5254) 1 year ago
- 8c9c1cf Pulled CLIP and UNet into Seperate Files (#5253) 1 year ago
- 5808c37 hotfix disable flaky llama3 beam benchmark on green (#5249) 1 year ago
- b9122ec revert stable diffusion validation with threefry (#5248) 1 year ago
- 57e8964 hcq spec test (#5226) 1 year ago
- d7839fd Add x!=0 -> (bool)x pattern [run_process_replay] [no_assert] (#5237) 1 year ago
- 14980f7 hotfix: unbreak llama 1 year ago
- 146eb3a hotfix: add repeat_interleave docs 1 year ago
- 3df47bc OpenELM + repeat_interleave (#5234) 1 year ago
- 7b7b751 simple hip backend for debugging (#5201) 1 year ago
- 88763eb fix stable_diffusion with fp16 (#5239) 1 year ago
- 649641a fix tqdm with generator without `__len__` (#5238) 1 year ago
- fd53b6d tqdm supports fractional blocks (#5233) 1 year ago
- ae10ae4 simplify tqdm scale math (#5231) 1 year ago
- ad1ca7d [Feature] Added BinaryOps.AND/BinaryOps.OR (#5223) 1 year ago
- 50b05dd tqdm minor cleanup (#5229) 1 year ago