chenyu c07daf40e7 move attention upcast (#7830) ...

still upcast before softmax, but faster because intermediate buffer can be stored in half (as long as qk is within half range).

2024-11-22 17:10:51 -05:00

..

external

less flaky benchmarks (#7855)

2024-11-22 16:39:39 -05:00

imported

node cleanup + local metal test speed [pr] (#6880)

2024-10-04 18:14:23 +08:00

models

move is_dtype_supported to device [pr] (#7575)

2024-11-07 20:38:03 +08:00

testextra

Hook memoryview via class instead of a function (#7627)

2024-11-11 09:07:06 +08:00

unit

move attention upcast (#7830)

2024-11-22 17:10:51 -05:00

__init__.py

All devices are equal! (#196)

2020-12-15 23:44:08 -08:00

Dockerfile

new cloud is cloudy [pr] (#7631)

2024-11-11 20:18:04 +08:00

helpers.py

move is_dtype_supported to device [pr] (#7575)

2024-11-07 20:38:03 +08:00

test_arange.py

Tensor.cummax (#7854)

2024-11-22 15:55:02 -05:00

test_assign.py

Add replicate mode to Tensor.pad (#7608)

2024-11-18 10:55:38 -05:00

test_compile_failures.py

enable a few tests that work now (#7721)

2024-11-15 14:30:52 -05:00

test_const_folding.py

Remove UnaryOps, BinaryOps, TernaryOps, MetaOps [pr] (#7725)

2024-11-16 20:56:56 +08:00

test_conv_shapetracker.py

s/UOps/Ops (#7500)

2024-11-03 11:26:10 +08:00

test_conv.py

db in wal mode (#5388)

2024-07-12 20:43:36 -07:00

test_copy_speed.py

node cleanup + local metal test speed [pr] (#6880)

2024-10-04 18:14:23 +08:00

test_device_speed.py

remove iter from uopgraph (#6110)

2024-08-16 15:58:29 -07:00

test_dtype_alu.py

move is_dtype_supported to device [pr] (#7575)

2024-11-07 20:38:03 +08:00

test_dtype.py

clean up DTYPES_DICT [pr] (#7845)

2024-11-22 10:01:34 -05:00

test_fusion_op.py

Ops.VALID cleanup, move recursive tests [pr] (#7713)

2024-11-15 20:22:46 +08:00

test_fuzz_shape_ops.py

faster ci [pr] (#7348)

2024-10-29 14:01:44 +08:00

test_gc.py

seed in tensor (#6869)

2024-10-06 14:46:58 -04:00

test_graph.py

view doesn't have buffer, fix the tests [pr] (#7841)

2024-11-22 20:41:55 +08:00

test_hcq.py

_prg to match _buf [pr] (#7816)

2024-11-21 12:44:48 +08:00

test_image_dtype.py

image dtype fusion tests [pr] (#7530)

2024-11-04 22:00:16 +08:00

test_jit.py

skip test_jit_batch_split if JIT >= 2 (#7561)

2024-11-05 14:59:04 -05:00

test_kernel_cache.py

test_kernel_cache_in_action: fix test (#7792)

2024-11-19 13:34:56 -05:00

test_lazybuffer.py

assert view degrade to const tests post scheduler graph_rewrite [pr] (#7822)

2024-11-21 19:00:41 +08:00

test_linearizer_dumb.py

Remove UnaryOps, BinaryOps, TernaryOps, MetaOps [pr] (#7725)

2024-11-16 20:56:56 +08:00

test_linearizer_failures.py

Remove UnaryOps, BinaryOps, TernaryOps, MetaOps [pr] (#7725)

2024-11-16 20:56:56 +08:00

test_linearizer_overflows.py

Remove UnaryOps, BinaryOps, TernaryOps, MetaOps [pr] (#7725)

2024-11-16 20:56:56 +08:00

test_linearizer.py

metal bf16 tc support [pr] (#7408)

2024-11-20 14:39:08 -05:00

test_masked_st.py

multitensor start (#2676)

2023-12-07 17:07:05 -08:00

test_metal.py

METAL raise RuntimeError with no compiler and bad src (#7603)

2024-11-08 17:09:12 -05:00

test_method_cache.py

simple LoadOps.ASSIGN (#3745)

2024-03-14 20:44:34 -07:00

test_multitensor.py

fix Tensor(MultiLazyBuffer) with different dtype should fail (#7757)

2024-11-17 21:05:45 -05:00

test_net_speed.py

nv mockgpu (#4600)

2024-05-15 23:46:08 +03:00

test_nn.py

Raise error when model architecture does not match state dict (#7772)

2024-11-20 00:11:54 +08:00

test_ocl.py

Fix GPT2 with OpenCL backend. (#6821)

2024-10-01 16:57:22 +08:00

test_ops.py

Tensor.cummax (#7854)

2024-11-22 15:55:02 -05:00

test_optim.py

move is_dtype_supported to device [pr] (#7575)

2024-11-07 20:38:03 +08:00

test_pickle.py

Revert "s/UPat/Pat (#7506)" [pr] (#7517)

2024-11-03 16:33:02 -05:00

test_profiler.py

_prg to match _buf [pr] (#7816)

2024-11-21 12:44:48 +08:00

test_randomness.py

add tiny test for randomness + remove ulong buffers (#7648)

2024-11-12 12:45:52 +08:00

test_rearrange_einops.py

Missing features from rearrange (#6184)

2024-08-19 11:19:07 -07:00

test_renderer_failures.py

BufferSpec and ProgramSpec [pr] (#7814)

2024-11-21 12:18:05 +08:00

test_sample.py

UOp.variable (#7010)

2024-10-12 18:20:44 +08:00

test_schedule.py

add UOp.const_with_shape [pr] (#7825)

2024-11-21 21:13:23 +08:00

test_search.py

Remove UnaryOps, BinaryOps, TernaryOps, MetaOps [pr] (#7725)

2024-11-16 20:56:56 +08:00

test_setitem.py

fix advanced setitem with 1 in shape (#7797)

2024-11-19 20:04:59 -05:00

test_specific_conv.py

move is_dtype_supported to device [pr] (#7575)

2024-11-07 20:38:03 +08:00

test_speed_v_torch.py

fix HALF=1 in test_speed_v_torch (#7642)

2024-11-11 14:29:37 -05:00

test_subbuffer.py

move lazy to engine [pr] (#6886)

2024-10-04 23:19:26 +08:00

test_symbolic_jit.py

UOp.variable (#7010)

2024-10-12 18:20:44 +08:00

test_symbolic_ops.py

UOp.variable (#7010)

2024-10-12 18:20:44 +08:00

test_symbolic_shapetracker.py

fix view add with symbolic shape (#7569)

2024-11-06 11:39:20 -05:00

test_tensor_data.py

BEAM_COMPARE=2 validates the correctness of BEAM kernels (#5458)

2024-07-13 13:53:43 -07:00

test_tensor_variable.py

combine pad2d with pad (#7677)

2024-11-14 17:56:02 +08:00

test_tensor.py

simpler rand_like (#7680)

2024-11-13 12:28:41 -05:00

test_tiny.py

add tiny test for randomness + remove ulong buffers (#7648)

2024-11-12 12:45:52 +08:00

test_to_numpy.py

Apply ruff linting rules to tests (#2473)

2023-11-27 21:24:06 -08:00

test_transcendental.py

move is_dtype_supported to device [pr] (#7575)

2024-11-07 20:38:03 +08:00

test_uop_graph.py

Remove UnaryOps, BinaryOps, TernaryOps, MetaOps [pr] (#7725)

2024-11-16 20:56:56 +08:00

test_uops_stats.py

update tests (#7554)

2024-11-05 11:35:15 -05:00

test_uops.py

BufferSpec and ProgramSpec [pr] (#7814)

2024-11-21 12:18:05 +08:00

test_viz.py

Remove UnaryOps, BinaryOps, TernaryOps, MetaOps [pr] (#7725)

2024-11-16 20:56:56 +08:00

test_winograd.py

fix wino conv output dtype for half inputs (#7829)

2024-11-21 12:13:54 -05:00

test_zero_copy.py

remove numpy from device (#3123)

2024-01-14 19:36:05 -08:00