ROCm AMDMIGraphX issues

ROCm / AMDMIGraphX

AMD's graph optimization engine.

https://rocm.docs.amd.com/projects/AMDMIGraphX/en/latest/

MIT License

185 stars 86 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Fix example to support correct Transformer version (#3556)

#3625 causten opened 1 day ago
2
Onnxruntime Weekly Sync 2024-11-15

#3624 github-actions[bot] opened 1 day ago
0
Try 3582 mlir

#3623 causten opened 2 days ago
3
Investigate memory failures when workspace size is set to 0 in compile hipblaslt pass

#3622 ahsan-ca opened 2 days ago
0
Fix failure generating crashes

#3621 causten opened 2 days ago
1
MLIR updates for 6.3

#3620 causten closed 2 days ago
0
Add FLUX e2e example

#3619 shivadbhavsar opened 2 days ago
4
Add end2end example for SD3

#3618 kahmed10 opened 3 days ago
4
Investigate llama2 int4 memory usage

#3617 turneram opened 3 days ago
0
Enable MIGRAPHX_MLIR_TUNING_CFG generation when defined

#3616 richagadgil opened 3 days ago
1
Fixed multilevel pointer conversion tidy check errors and re-enable t…

#3615 taylding-amd opened 3 days ago
3
Dont use mixed layouts with convolution (#3587)

#3614 causten closed 1 day ago
1
Serialize output arg

#3613 pfultz2 closed 3 days ago
2
Quantized distillgpt2 accuracy issue when quant params are inputs

#3612 shivadbhavsar opened 4 days ago
0
Fix broken link develop

#3611 prbasyal-amd closed 3 days ago
1
Track broadcast axes in the shape_transform_descriptor

#3610 pfultz2 opened 5 days ago
0
Always output a packed type for q/dq

#3609 pfultz2 closed 2 days ago
3
rocMLIR Weekly Sync 2024-11-10

#3608 github-actions[bot] closed 3 days ago
4
Windows fixes for generic_float

#3607 pfultz2 closed 5 days ago
3
Use the float8 and generic_float template for specialization

#3606 pfultz2 closed 4 days ago
3
Prefill fix for multiple output cases

#3605 shivadbhavsar closed 4 days ago
4
Generic "truncate float" class for quantization

#3604 richagadgil closed 3 days ago
3
Fix tidy check by getting rid of unused variables.

#3603 taylding-amd closed 3 days ago
4
Workaround to get correct results from llama2 mxr

#3602 turneram closed 4 days ago
3
UAI llama2 script recompiles for each prompt even with prefill approach

#3601 turneram opened 1 week ago
0
Onnxruntime Weekly Sync 2024-11-08

#3600 github-actions[bot] closed 3 days ago
0
fix compilation on Windows

#3599 apwojcik closed 6 days ago
0
Dangling quantizelinear from horizontal fusion, BERT and DistilGPT2

#3598 CharlieL7 opened 1 week ago
0
Missing constant propagation: `Literal` -> `Multibroadcast` -> `Quantizelinear`

#3597 CharlieL7 opened 1 week ago
0
GroupQueryAttention produces incorrect results when loaded from mxr

#3596 turneram opened 1 week ago
4
Integrate new CK API and operations with MIGraphX

#3595 mirza-halilcevic opened 1 week ago
0
Compile pass for hipBLASLt

#3594 ahsan-ca opened 1 week ago
8
Update finalize function for hipBLASLt to find tuned solution only when solution index is 0.

#3593 ahsan-ca closed 3 days ago
5
Remove hipblaslt version check in `hip_gemm_impl.cpp` for OCP FP8 types

#3592 CharlieL7 opened 1 week ago
0
Generic "truncate_float" class for bf16 and fp16 quantization

#3591 richagadgil opened 1 week ago
2
Removed MIGRAPHX_ENABLE_HIPRTC_WORKAROUNDS environment variable

#3590 tcgu-amd closed 1 week ago
1
[Issue]: Slower Inference Time on SDXL Model Using MIGraphX Than PyTorch

#3589 zixianwang2022 closed 2 days ago
11
PagedAttention support in MIGraphX

#3588 gyulaz-htec opened 1 week ago
0
Dont use mixed layouts with convolution

#3587 pfultz2 closed 4 days ago
3
rocMLIR Weekly Sync 2024-11-03

#3586 github-actions[bot] opened 1 week ago
1
[Issue]: `migraphx.quantize_int8()` not performing quantization on GPU

#3585 Squidwarder closed 1 week ago
4
use reshape to handle flatten

#3584 shivadbhavsar closed 1 week ago
3
Onnxruntime Weekly Sync 2024-11-01

#3583 github-actions[bot] closed 1 week ago
0
Add support for unsigned types with mlir

#3582 pfultz2 opened 2 weeks ago
5
update compile options to use -fnative-half-arguments-and-returns

#3581 kahmed10 closed 2 weeks ago
4
Fix bug in common dims

#3580 pfultz2 closed 2 weeks ago
5
Bug in find_nested_shape_transforms

#3579 shivadbhavsar closed 1 week ago
2
Implement BF16 using generic_float class

#3578 richagadgil opened 2 weeks ago
4
Enable GEMM/dot for FP8 using hipblasLT

#3577 CharlieL7 closed 3 days ago
8
Make check failures on Navi48

#3576 ahsan-ca opened 2 weeks ago
1