Add MultiHeadAttention

[marko-fabo-htec]

Add support for MultiHeadAttention operator.
Description: https://github.com/microsoft/onnxruntime/blob/main/docs/ContribOperators.md#com.microsoft.MultiHeadAttention

TODOs:

• add key_padding_mask input handling
• add attention_bias input handling
• add past_key and past_value inputs handling
• add parse tests
• check performance

[codecov]

Codecov Report

Attention: Patch coverage is 81.25000% with 21 lines in your changes missing coverage. Please review.

Project coverage is 92.11%. Comparing base (952a257) to head (ec2e5b0).

Files with missing lines	Patch %	Lines
src/onnx/parse_multi_head_attention.cpp	81.25%	21 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #3650      +/-   ##
===========================================
- Coverage    92.17%   92.11%   -0.06%     
===========================================
  Files          513      514       +1     
  Lines        21583    21695     +112     
===========================================
+ Hits         19894    19985      +91     
- Misses        1689     1710      +21     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

---

🚨 Try these New Features:

• Flaky Tests Detection - Detect and resolve failed and flaky tests

[migraphx-bot]

Test	Batch	Rate new ec2e5b	Rate old 0f36aa	Diff	Compare
torchvision-resnet50	64	3,258.04	3,261.99	-0.12%	✅
torchvision-resnet50_fp16	64	6,989.64	6,984.41	0.07%	✅
torchvision-densenet121	32	2,436.29	2,434.46	0.08%	✅
torchvision-densenet121_fp16	32	4,105.31	4,068.77	0.90%	✅
torchvision-inceptionv3	32	1,628.76	1,630.14	-0.08%	✅
torchvision-inceptionv3_fp16	32	2,745.51	2,746.22	-0.03%	✅
cadene-inceptionv4	16	765.35	765.59	-0.03%	✅
cadene-resnext64x4	16	810.68	809.78	0.11%	✅
slim-mobilenet	64	7,470.53	7,474.57	-0.05%	✅
slim-nasnetalarge	64	208.49	208.58	-0.04%	✅
slim-resnet50v2	64	3,441.90	3,441.49	0.01%	✅
bert-mrpc-onnx	8	1,144.35	1,150.80	-0.56%	✅
bert-mrpc-tf	1	465.18	465.54	-0.08%	✅
pytorch-examples-wlang-gru	1	412.37	420.06	-1.83%	✅
pytorch-examples-wlang-lstm	1	402.31	381.98	5.32%	🔆
torchvision-resnet50_1	1	770.37	750.44	2.66%	✅
cadene-dpn92_1	1	396.31	398.35	-0.51%	✅
cadene-resnext101_1	1	382.22	382.96	-0.19%	✅
onnx-taau-downsample	1	346.19	346.08	0.03%	✅
dlrm-criteoterabyte	1	33.33	33.35	-0.05%	✅
dlrm-criteoterabyte_fp16	1	52.76	52.68	0.15%	✅
agentmodel	1	8,407.25	8,091.53	3.90%	🔆
unet_fp16	2	58.81	58.77	0.07%	✅
resnet50v1_fp16	1	920.85	943.16	-2.37%	✅
resnet50v1_int8	1	1,013.50	1,012.12	0.14%	✅
bert_base_cased_fp16	64	1,170.20	1,169.97	0.02%	✅
bert_large_uncased_fp16	32	363.56	363.75	-0.05%	✅
bert_large_fp16	1	200.48	199.03	0.73%	✅
distilgpt2_fp16	16	2,197.33	2,201.98	-0.21%	✅
yolov5s	1	529.38	539.79	-1.93%	✅
tinyllama	1	43.40	43.42	-0.03%	✅
vicuna-fastchat	1	179.23	175.75	1.98%	✅
whisper-tiny-encoder	1	415.99	418.02	-0.49%	✅
whisper-tiny-decoder	1	425.55	428.37	-0.66%	✅

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