branch: master
Commits on master
- 957ae9b Fix Tensor's __repr__ for printing out grad (#3673) 2 years ago
- 0f16729 RDNA3: restore launch bounds (#3672) 2 years ago
- d7452c2 clean up llvmir builder (#3671) 2 years ago
- 1143c62 tensor.py touchups (#3667) 2 years ago
- 69ca7f7 changes for teenygrad (#3665) 2 years ago
- 89b8b5d Fix missing import. (#3666) 2 years ago
- 8ae85b2 add inference_mode context manager with decorator support (#3621) 2 years ago
- b5cbf17 Fix `Tensor.cumsum` when axis of length 0 is selected (#3473) 2 years ago
- 915f987 use custom KernelOptError in kernel opt (#3661) 2 years ago
- ac02e73 ptx timing vs cuda timing (#3659) 2 years ago
- daa4034 No more metal flakiness (#3643) 2 years ago
- e25879d don't get new var_val for the same ast in fuzz_linearizer (#3657) 2 years ago
- 1130c73 add FUZZ_NTH to fuzz_linearizer (#3656) 2 years ago
- 9f66dcf PolynomialDecayWithWarmup + tests (#3649) 2 years ago
- 57df8e8 update fuzz_linearizer (#3648) 2 years ago
- b282a45 fix direct store float4 with same vin (#3652) 2 years ago
- a66ffec update kernel dataset to exclude the disktensor ones (#3651) 2 years ago
- fcf4a5c fix example that calls Tensor.__bool__ (#3650) 2 years ago
- 6e50582 working to improve ptx (#3647) 2 years ago
- 1853ec9 add tests for bfloat16 on HIP (#3638) 2 years ago
- 0cef284 fix typing FlopCounter.flops can be sint (#3646) 2 years ago
- 906cc3a cleanup tests Device[Device.DEFAULT] is always Compiled (#3645) 2 years ago
- bdd62c7 make the bf16 include dynamic (#3642) 2 years ago
- 4552248 fix Tensor.to preserves grad.data (#3636) 2 years ago
- d33311e remove parens of ALU if it has associative property (#3635) 2 years ago
- fe6b6e3 remove parentheses of GEP if it's from SSA (#3634) 2 years ago
- 0afaf70 lars optimizer + tests (#3631) 2 years ago
- b2e92d4 skip METAL sin test in test_dtype_alu (#3633) 2 years ago
- 8f10bfa ban __bool__ on Tensor (#3632) 2 years ago
- 81baf3e bring ptx back (#3623) 2 years ago