nmt_results API & Inference Endpoint

Author

Zesheng Shi
System developer, model trainer, and repository maintainer for this WMT26 submission.

See AUTHORS.md for the repository authorship record.

Submitted branch: Full-parameter SFT (not the LoRA branch). Under matched data, seed, and evaluation conditions, full-parameter SFT outperforms LoRA SFT on all 14 directions (see Validation results below and docs/wmt26_hymt2_1_8b_paper_emnlp.pdf).

Model summary

Table with columns: Field, Value
Field	Value
Base model	`Tencent-Hunyuan/Hy-MT2-1.8B` (downloaded from ModelScope)
Adaptation lineage	Hy-MT2-1.8B → CPT 1-epoch (monolingual continued pretraining) → Full-parameter SFT
Architecture	`HunYuanDenseV1ForCausalLM` (`model_type: hunyuan_v1_dense`)
Parameters	1.8B (≤ 20B → complies with shared-task size limit)
Precision	bfloat16
Context cutoff	4,096 train / 2,048 max new tokens at inference
Languages	Chinese (ZH) ↔ 7 Southeast Asian languages: ID, TH, KM, MS, VI, MY, LO
License	Apache-2.0 (inherited from the Hy-MT2-1.8B base model)

⚠️ Loading requirement (important)

HunYuanDenseV1ForCausalLM is a native transformers architecture — present in transformers>=4.57.6 (verified in the live venv venvs/hymt2-1_8b-wmt26; model_doc/hunyuan_v1_dense). The config.json contains no auto_map, so trust_remote_code is NOT required and no custom modeling_*.py is needed. The config.json declares transformers_version: "4.57.6", which matches the actual runtime — install exactly this version (or newer) to be safe.

bash
pip install "transformers>=4.57.6" torch accelerate safetensors
# exact pinned versions for full reproducibility: see code/requirements.txt

Covered translation directions (14)

Table with columns: Family, Directions
Family	Directions
Southeast Asian → Chinese	ID–ZH, TH–ZH, KM–ZH, MS–ZH, VI–ZH, MY–ZH, LO–ZH
Chinese → Southeast Asian	ZH–ID, ZH–TH, ZH–KM, ZH–MS, ZH–VI, ZH–MY, ZH–LO

Quick start

python
from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained(
    ".",                      # this repository
    torch_dtype="auto",
    device_map="auto",
    # trust_remote_code=False  (default) — architecture is native (verified on transformers 4.57.6)
)
tokenizer = AutoTokenizer.from_pretrained(".")

src_lang, tgt_lang = "Thai", "Chinese"
src_text = "สวัสดีครับ ยินดีที่ได้รู้จัก"

messages = [{
    "role": "user",
    "content": (
        f"Translate the following {src_lang} into {tgt_lang}. "
        f"Output only the translation and do not add explanations.\n{src_text}"
    ),
}]
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
out = model.generate(**inputs, max_new_tokens=2048, do_sample=False)  # temperature=0 / greedy
print(tokenizer.decode(out[0][inputs.input_ids.shape[1]:], skip_special_tokens=True))

The production inference script (code/generate_wmt26_translations.py) applies the same instruction contract and chat template, shards the validation set across GPUs, and stops on the end-of-sequence marker (2,048 is an upper bound, not a target length).

Repository layout

markdown
hymt2-1_8b-wmt26-full-sft-hf/
├── model.safetensors            # full-parameter SFT weights (bf16, ~3.8 GB)
├── config.json                  # architecture = HunYuanDenseV1ForCausalLM
├── generation_config.json
├── tokenizer.json
├── tokenizer_config.json
├── special_tokens_map.json
├── chat_template.jinja
├── eval_results.json            # validation metrics for this checkpoint
├── AUTHORS.md                   # authorship and repository-maintainer record
├── README.md                    # this file
├── code/                        # inference + reproduction (authored in train/wmt26, uploaded as-is)
│   ├── generate_wmt26_translations.py   # native Transformers inference
│   ├── start_wmt26_vllm_4gpu.sh         # optional 4×80G vLLM serving (throughput only)
│   ├── record_wmt26_environment.py      # dumps environment manifest (deps/versions)
│   ├── hymt2_1_8b_wmt26_full_sft.yaml    # exact full-SFT training config
│   ├── WMT26_RUNBOOK_CN.md               # end-to-end reproduction runbook
│   ├── WMT26_EVAL_HANDOFF_CN.md          # evaluation handoff notes
│   └── requirements.txt                 # pinned dependency versions (live venv: venvs/hymt2-1_8b-wmt26)
├── docs/
│   └── wmt26_hymt2_1_8b_paper_emnlp.pdf  # system description (EMNLP-format)
└── evidence/                    # version-provenance archive (kept locally; lora-* contrast branches NOT uploaded)
    ├── evaluation/full-sft-best/        # confirms this repo == the 2026-08-20 blind-test submission
    └── records/                        # training logs + environment manifest (mirrored to code/)

Training / data-prep / scoring helper scripts that are not required for inference or reproduction are intentionally omitted from this repository; they live in the project's train/wmt26/ source tree.

Training configuration (full-parameter SFT)

Table with columns: Setting, Value
Setting	Value
Initialization	CPT 1-epoch checkpoint
Trainable parameters	All model parameters
Learning rate	1×10⁻⁵
Epochs	3
Per-device batch size	4
Gradient accumulation	4
Context cutoff	4,096 tokens
Precision	bf16
Seed	2026

Validation results (26,474-example reconstructed validation set)

Table with columns: Direction, BLEU, COMET, Q
Direction	BLEU	COMET	Q
ID–ZH	65.99	85.98	75.98
ZH–ID	52.94	88.70	70.82
TH–ZH	47.74	83.87	65.80
ZH–TH	13.37	85.23	49.30
KM–ZH

COMET is reported on a 0–100 scale; the auxiliary quality score Q = (BLEU + COMET) / 2. These are validation-set numbers; the organizer-held blind test is kept separate and was not used for checkpoint selection, prompt design, or decoding-parameter tuning.

Evaluation notes

Quality metrics (sacreBLEU, COMET): computed by code/ in a separate scoring environment to avoid mixing training and scoring dependencies.
Throughput: measured by the shared-task organizers on a 4×80 GB platform. The bundled start_wmt26_vllm_4gpu.sh is provided for diagnostic throughput only and is not required for the quality comparison.

System description

The full methodology, data construction (leakage-controlled), staged CPT→SFT design, and direction-level analysis are described in docs/wmt26_hymt2_1_8b_paper_emnlp.pdf ("Hy-MT2-1.8B for WMT26: Staged Multilingual Adaptation for Chinese–Southeast Asian Translation").

Field

Value

Base model

Tencent-Hunyuan/Hy-MT2-1.8B (downloaded from ModelScope)

Adaptation lineage

Hy-MT2-1.8B → CPT 1-epoch (monolingual continued pretraining) → Full-parameter SFT

Architecture

HunYuanDenseV1ForCausalLM (model_type: hunyuan_v1_dense)

Parameters

1.8B (≤ 20B → complies with shared-task size limit)

Precision

bfloat16

Context cutoff

4,096 train / 2,048 max new tokens at inference

Languages

Chinese (ZH) ↔ 7 Southeast Asian languages: ID, TH, KM, MS, VI, MY, LO

License

Apache-2.0 (inherited from the Hy-MT2-1.8B base model)

Family

Directions

Southeast Asian → Chinese

ID–ZH, TH–ZH, KM–ZH, MS–ZH, VI–ZH, MY–ZH, LO–ZH

Chinese → Southeast Asian

ZH–ID, ZH–TH, ZH–KM, ZH–MS, ZH–VI, ZH–MY, ZH–LO

python

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained(
    ".",                      # this repository
    torch_dtype="auto",
    device_map="auto",
    # trust_remote_code=False  (default) — architecture is native (verified on transformers 4.57.6)
)
tokenizer = AutoTokenizer.from_pretrained(".")

src_lang, tgt_lang = "Thai", "Chinese"
src_text = "สวัสดีครับ ยินดีที่ได้รู้จัก"

messages = [{
    "role": "user",
    "content": (
        f"Translate the following {src_lang} into {tgt_lang}. "
        f"Output only the translation and do not add explanations.\n{src_text}"
    ),
}]
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
out = model.generate(**inputs, max_new_tokens=2048, do_sample=False)  # temperature=0 / greedy
print(tokenizer.decode(out[0][inputs.input_ids.shape[1]:], skip_special_tokens=True))

markdown

hymt2-1_8b-wmt26-full-sft-hf/
├── model.safetensors            # full-parameter SFT weights (bf16, ~3.8 GB)
├── config.json                  # architecture = HunYuanDenseV1ForCausalLM
├── generation_config.json
├── tokenizer.json
├── tokenizer_config.json
├── special_tokens_map.json
├── chat_template.jinja
├── eval_results.json            # validation metrics for this checkpoint
├── AUTHORS.md                   # authorship and repository-maintainer record
├── README.md                    # this file
├── code/                        # inference + reproduction (authored in train/wmt26, uploaded as-is)
│   ├── generate_wmt26_translations.py   # native Transformers inference
│   ├── start_wmt26_vllm_4gpu.sh         # optional 4×80G vLLM serving (throughput only)
│   ├── record_wmt26_environment.py      # dumps environment manifest (deps/versions)
│   ├── hymt2_1_8b_wmt26_full_sft.yaml    # exact full-SFT training config
│   ├── WMT26_RUNBOOK_CN.md               # end-to-end reproduction runbook
│   ├── WMT26_EVAL_HANDOFF_CN.md          # evaluation handoff notes
│   └── requirements.txt                 # pinned dependency versions (live venv: venvs/hymt2-1_8b-wmt26)
├── docs/
│   └── wmt26_hymt2_1_8b_paper_emnlp.pdf  # system description (EMNLP-format)
└── evidence/                    # version-provenance archive (kept locally; lora-* contrast branches NOT uploaded)
    ├── evaluation/full-sft-best/        # confirms this repo == the 2026-08-20 blind-test submission
    └── records/                        # training logs + environment manifest (mirrored to code/)

Setting

Value

Initialization

CPT 1-epoch checkpoint

Trainable parameters

All model parameters

Learning rate

1×10⁻⁵

Epochs

Per-device batch size

Gradient accumulation

Context cutoff

4,096 tokens

Precision

bf16

Seed

2026

Direction

BLEU

COMET

ID–ZH

65.99

85.98

75.98

ZH–ID

52.94

88.70

70.82

TH–ZH

47.74

83.87

65.80

ZH–TH

13.37

85.23

49.30

KM–ZH

nmt_results

README

Author

Model summary

⚠️ Loading requirement (important)

Covered translation directions (14)

Quick start

Repository layout

Training configuration (full-parameter SFT)

Validation results (26,474-example reconstructed validation set)

Evaluation notes

System description

Explore FriendliAI today

README

Author

Model summary

⚠️ Loading requirement (important)

Covered translation directions (14)

Quick start

Repository layout

Training configuration (full-parameter SFT)

Validation results (26,474-example reconstructed validation set)

Evaluation notes

System description