Add tencent/HY-MT1.5-1.8B recipes

[hanbitmyths]

Summary

• Add Olive recipes for tencent/HY-MT1.5-1.8B across baseline, CPU, CUDA, and WebGPU targets.
• Add WMT18 Chinese-to-English lm-eval configs and documentation.
• Document the selected full INT4 group_size 32 recipes and quality/size results.

Validation

• Parsed all HY-MT JSON recipe files successfully.
• Confirmed no files over 100 MB in the recipe tree.
• Ran git diff --check for the recipe changes.

[Copilot]

Pull request overview

This PR adds a complete Olive “recipe pack” for tencent/HY-MT1.5-1.8B, covering baseline (PyTorch) evaluation plus CPU/CUDA/WebGPU export paths with optional WMT18 zh→en lm-eval evaluation configs and accompanying documentation.

Changes:

• Added baseline (PyTorch) WMT18 zh→en evaluation recipe + docs/requirements.
• Added CPU/CUDA/WebGPU export recipes for fp32/fp16 and INT4 (GPTQ→RTN→ModelBuilder) variants, plus “with_eval” counterparts.
• Added per-target info.yaml, requirements.txt, and READMEs, plus upstream license reference.

Reviewed changes

Copilot reviewed 26 out of 26 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
tencent-HY-MT1.5-1.8B/LICENSE	Adds upstream license reference for the model.
tencent-HY-MT1.5-1.8B/baseline/tencent-HY-MT1.5-1.8B_pytorch_with_eval.json	Baseline PyTorch lm-eval config for WMT18 zh→en.
tencent-HY-MT1.5-1.8B/baseline/requirements.txt	Baseline dependencies for running the evaluation recipe.
tencent-HY-MT1.5-1.8B/baseline/README.md	Baseline usage + recorded evaluation results.
tencent-HY-MT1.5-1.8B/baseline/info.yaml	Registers baseline recipe metadata.
tencent-HY-MT1.5-1.8B/cpu/tencent-HY-MT1.5-1.8B_cpu_fp32.json	CPU fp32 export recipe (ModelBuilder + TieWordEmbeddings + external data).
tencent-HY-MT1.5-1.8B/cpu/tencent-HY-MT1.5-1.8B_cpu_fp32_with_eval.json	CPU fp32 export + WMT18 evaluation.
tencent-HY-MT1.5-1.8B/cpu/tencent-HY-MT1.5-1.8B_cpu_int4.json	CPU INT4 export recipe (GPTQ→RTN→ModelBuilder + surgeries + external data).
tencent-HY-MT1.5-1.8B/cpu/tencent-HY-MT1.5-1.8B_cpu_int4_with_eval.json	CPU INT4 export + WMT18 evaluation.
tencent-HY-MT1.5-1.8B/cpu/requirements.txt	CPU dependencies (ORT GenAI, etc.).
tencent-HY-MT1.5-1.8B/cpu/README.md	CPU instructions + notes and evaluation/size results.
tencent-HY-MT1.5-1.8B/cpu/info.yaml	Registers CPU recipes metadata.
tencent-HY-MT1.5-1.8B/cuda/tencent-HY-MT1.5-1.8B_cuda_fp16.json	CUDA fp16 export recipe (ModelBuilder + TieWordEmbeddings).
tencent-HY-MT1.5-1.8B/cuda/tencent-HY-MT1.5-1.8B_cuda_fp16_with_eval.json	CUDA fp16 export + WMT18 evaluation (ORT GenAI).
tencent-HY-MT1.5-1.8B/cuda/tencent-HY-MT1.5-1.8B_cuda_int4.json	CUDA INT4 export (GPTQ→RTN→ModelBuilder + surgeries, CUDA graph enabled).
tencent-HY-MT1.5-1.8B/cuda/tencent-HY-MT1.5-1.8B_cuda_int4_with_eval.json	CUDA INT4 export + WMT18 evaluation (ORT GenAI, CUDA graph enabled).
tencent-HY-MT1.5-1.8B/cuda/requirements.txt	CUDA dependencies (onnxruntime-gpu + onnxruntime-genai-cuda).
tencent-HY-MT1.5-1.8B/cuda/README.md	CUDA instructions + notes and evaluation/size results.
tencent-HY-MT1.5-1.8B/cuda/info.yaml	Registers CUDA recipes metadata.
tencent-HY-MT1.5-1.8B/webgpu/tencent-HY-MT1.5-1.8B_webgpu_fp16.json	WebGPU fp16 export recipe (ModelBuilder + TieWordEmbeddings).
tencent-HY-MT1.5-1.8B/webgpu/tencent-HY-MT1.5-1.8B_webgpu_fp16_with_eval.json	WebGPU fp16 export + WMT18 evaluation config.
tencent-HY-MT1.5-1.8B/webgpu/tencent-HY-MT1.5-1.8B_webgpu_int4.json	WebGPU INT4 export (GPTQ→RTN→ModelBuilder + surgeries).
tencent-HY-MT1.5-1.8B/webgpu/tencent-HY-MT1.5-1.8B_webgpu_int4_with_eval.json	WebGPU INT4 export + WMT18 evaluation (ORT GenAI model class configured).
tencent-HY-MT1.5-1.8B/webgpu/requirements.txt	WebGPU dependencies (pre-release enabled, ORT webgpu).
tencent-HY-MT1.5-1.8B/webgpu/README.md	WebGPU instructions + notes and evaluation/size results.
tencent-HY-MT1.5-1.8B/webgpu/info.yaml	Registers WebGPU recipes metadata.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.