rocMLIR Weekly Sync 2024-10-27

[github-actions]

Update rocMLIR version

• Updated with changes from 2024-10-27
• Auto-generated by create-pull-request
• Update requirements.txt to rocMLIR@5f51701ee5ff5216125eb3e62deb2dce402cdd9c

[github-actions]

model	DEFAULT	MLIR_EXTENDED
bert-mrpc-onnx	✅	🔴
bert-mrpc-tf	✅	✅
pytorch-examples-wlang-gru	✅	✅
pytorch-examples-wlang-lstm	✅	✅
torchvision-resnet50_1	✅	✅
cadene-dpn92_1	✅	✅
cadene-resnext101_1	✅	✅
dlrm-criteoterabyte	✅	✅
agentmodel	✅	✅
unet	✅	🔴
resnet50v1	✅	✅
bert_base_cased_fp16	✅	✅
bert_large_uncased_fp16	🔴	✅
bert_large	✅	✅
yolov5s	✅	✅
tinyllama	✅	✅
vicuna-fastchat	✅	✅
whisper-tiny-encoder	✅	🔴
whisper-tiny-decoder	✅	✅
distilgpt2_fp16	✅	✅

✅ - within tolerance
🔴 - not within tolerance
❌ - error in execution

[migraphx-bot]

Test	Batch	Rate new 0ae791	Rate old 71fd27	Diff	Compare
torchvision-resnet50	64	3,258.64	3,258.47	0.01%	✅
torchvision-resnet50_fp16	64	6,987.78	6,989.35	-0.02%	✅
torchvision-densenet121	32	2,439.34	2,437.00	0.10%	✅
torchvision-densenet121_fp16	32	4,083.09	4,070.38	0.31%	✅
torchvision-inceptionv3	32	1,638.55	1,639.85	-0.08%	✅
torchvision-inceptionv3_fp16	32	2,763.05	2,763.41	-0.01%	✅
cadene-inceptionv4	16	775.94	776.50	-0.07%	✅
cadene-resnext64x4	16	811.32	808.24	0.38%	✅
slim-mobilenet	64	7,530.71	7,538.65	-0.11%	✅
slim-nasnetalarge	64	211.48	211.54	-0.03%	✅
slim-resnet50v2	64	3,505.27	3,507.21	-0.06%	✅
bert-mrpc-onnx	8	1,147.37	1,150.51	-0.27%	✅
bert-mrpc-tf	1	463.16	475.19	-2.53%	✅
pytorch-examples-wlang-gru	1	512.98	426.61	20.25%	🔆
pytorch-examples-wlang-lstm	1	398.24	376.26	5.84%	🔆
torchvision-resnet50_1	1	784.25	785.19	-0.12%	✅
cadene-dpn92_1	1	402.83	399.09	0.94%	✅
cadene-resnext101_1	1	383.36	383.01	0.09%	✅
onnx-taau-downsample	1	343.31	343.03	0.08%	✅
dlrm-criteoterabyte	1	33.32	33.33	-0.01%	✅
dlrm-criteoterabyte_fp16	1	52.73	52.73	0.01%	✅
agentmodel	1	7,986.77	8,178.96	-2.35%	✅
unet_fp16	2	59.06	58.92	0.24%	✅
resnet50v1_fp16	1	937.24	925.30	1.29%	✅
resnet50v1_int8	1	1,019.49	1,011.55	0.78%	✅
bert_base_cased_fp16	64	1,171.99	1,169.93	0.18%	✅
bert_large_uncased_fp16	32	363.42	363.31	0.03%	✅
bert_large_fp16	1	200.13	200.50	-0.18%	✅
distilgpt2_fp16	16	2,196.54	2,194.69	0.08%	✅
yolov5s	1	538.30	533.06	0.98%	✅
tinyllama	1	43.41	43.45	-0.09%	✅
vicuna-fastchat	1	174.67	172.29	1.38%	✅
whisper-tiny-encoder	1	418.52	417.95	0.14%	✅
whisper-tiny-decoder	1	435.64	425.65	2.35%	✅

Check results before merge 🔆

[migraphx-bot]

     ✅ bert-mrpc-onnx: PASSED: MIGraphX meets tolerance

     ✅ bert-mrpc-tf: PASSED: MIGraphX meets tolerance

     ✅ pytorch-examples-wlang-gru: PASSED: MIGraphX meets tolerance

     ✅ pytorch-examples-wlang-lstm: PASSED: MIGraphX meets tolerance

     ✅ torchvision-resnet50_1: PASSED: MIGraphX meets tolerance

     ✅ cadene-dpn92_1: PASSED: MIGraphX meets tolerance

     ✅ cadene-resnext101_1: PASSED: MIGraphX meets tolerance

     ✅ dlrm-criteoterabyte: PASSED: MIGraphX meets tolerance

     ✅ agentmodel: PASSED: MIGraphX meets tolerance

     ✅ unet: PASSED: MIGraphX meets tolerance

     ✅ resnet50v1: PASSED: MIGraphX meets tolerance

     ✅ bert_base_cased_fp16: PASSED: MIGraphX meets tolerance

🔴bert_large_uncased_fp16: FAILED: MIGraphX is not within tolerance - check verbose output

     ✅ bert_large: PASSED: MIGraphX meets tolerance

     ✅ yolov5s: PASSED: MIGraphX meets tolerance

     ✅ tinyllama: PASSED: MIGraphX meets tolerance

     ✅ vicuna-fastchat: PASSED: MIGraphX meets tolerance

     ✅ whisper-tiny-encoder: PASSED: MIGraphX meets tolerance

     ✅ whisper-tiny-decoder: PASSED: MIGraphX meets tolerance

     ✅ distilgpt2_fp16: PASSED: MIGraphX meets tolerance