Downloads · 30 days
9
16% of all-time downloads
mishbahul11/nerds-nits-nllb600m-ar-en-x-finetuned-wmt2026
nerds-nits-nllb600m-ar-en-x-finetuned-wmt2026 is a translation model from mishbahul11. Use it when you need text moved from one language to another. The card lists the license as mit.
This model is a fully fine-tuned version of facebook/nllb-200-distilled-600M, adapted for Arabic→English translation as part of the NERDS-NITS submission to the WMT 2026 Low-Resource Arabic–Asian Language Translation…
Downloads · 30 days
9
16% of all-time downloads
All-time downloads
57
Public
Parameters
615M
2.5 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors2.5 GB · 99%
From the Hugging Face model README
This model is a fully fine-tuned version of facebook/nllb-200-distilled-600M, adapted for Arabic→English translation as part of the NERDS-NITS submission to the WMT 2026 Low-Resource Arabic–Asian Language Translation Shared Task.
This model forms the Arabic→English pivot leg of the NERDS-NITS pivot-based translation framework, used for the Arabic→Hindi and Arabic→Bengali directions, where its English output is passed to pretrained IndicTrans2 checkpoints handling the English→Indic leg.
AutoTokenizer (inherited from the original NLLB-200-600M checkpoint)facebook/nllb-200-distilled-600MSeq2SeqTrainer, minimizing sequence-to-sequence cross-entropy lossarb_Arab, English — eng_LatnThis model was selected as the Arabic→English pivot leg after comparing two fine-tuning strategies: LoRA fine-tuning of NLLB-200-1.3B, and full fine-tuning of NLLB-200-600M. The fully fine-tuned 600M model outperformed the 1.3B LoRA model across BLEU, chrF, and COMET on the held-out development set, and was therefore chosen as the pivot leg for the Arabic-to-Indic direction in this submission.
Limitations:
| Hyperparameter | Value |
|---|---|
| Learning rate | 2e-5 |
| Epochs | 3 |
| Effective batch size | 64 |
| Fine-tuning method | Full fine-tuning |
| Hardware | Single NVIDIA T4 (16GB VRAM) |
| Precision | fp16 / bf16 mixed precision |
| LR scheduler | Cosine |
| Early stopping | Patience = 2 (on validation loss) |
This checkpoint and its English→Arabic counterpart were each fully fine-tuned from the same NLLB-200-600M initialization using identical hyperparameters, so that any performance difference between directions reflects genuine asymmetries in the translation task rather than differences in training configuration.
Fine-tuned on the WMT 2026 Arabic–English parallel corpus, consisting of 19,950 training pairs, with 1,050 dev pairs and 500 devtest pairs.
import torch
from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
model_name = "mishbahul11/nerds-nits-nllb600m-ar-en-x-finetuned-wmt2026"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
# Arabic -> English
tokenizer.src_lang = "arb_Arab"
inputs = tokenizer("مرحبا بالعالم", return_tensors="pt")
generated_tokens = model.generate(
**inputs,
forced_bos_token_id=tokenizer.convert_tokens_to_ids("eng_Latn"),
num_beams=5,
max_new_tokens=192,
)
print(tokenizer.batch_decode(generated_tokens, skip_special_tokens=True))
# Example: Arabic -> English (this model) -> Hindi (IndicTrans2)
from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
ar_en_model_name = "mishbahul11/nerds-nits-nllb600m-ar-en-x-finetuned-wmt2026"
en_hi_model_name = "ai4bharat/indictrans2-en-indic-dist-200M"
# Step 1: Arabic -> English (this model)
tokenizer_ar_en = AutoTokenizer.from_pretrained(ar_en_model_name)
model_ar_en = AutoModelForSeq2SeqLM.from_pretrained(ar_en_model_name)
tokenizer_ar_en.src_lang = "arb_Arab"
inputs = tokenizer_ar_en("مرحبا بالعالم", return_tensors="pt")
tokens = model_ar_en.generate(
**inputs,
forced_bos_token_id=tokenizer_ar_en.convert_tokens_to_ids("eng_Latn"),
num_beams=5,
)
english_text = tokenizer_ar_en.batch_decode(tokens, skip_special_tokens=True)[0]
# Step 2: English -> Hindi (via IndicTrans2)
# See ai4bharat/indictrans2-en-indic-dist-200M model card for full usage details
This model serves as the Arabic→English pivot leg; downstream pivot performance (combined with IndicTrans2's English→Indic leg) for Arabic→Hindi and Arabic→Bengali is reported below.
| Direction | Best System | BLEU | chrF | COMET |
|---|---|---|---|---|
| AR→HI | NLLB + IndicTrans (pivot) | 28.64 | 51.8 | 0.7186 |
| AR→BN | NLLB + IndicTrans (pivot) | 17.2 | 48.1 | 0.8075 |
| Direction | Best System | BLEU | chrF2 | TER | COMET | mBERT |
|---|---|---|---|---|---|---|
| AR→HI | NLLB + IndicTrans (pivot) | 27.7756 | 52.9529 | 63.7534 | 70.0510 | 85.0161 |
| AR→BN | NLLB + IndicTrans (pivot) | 17.1027 | 51.8287 | 74.1051 | 78.6902 | 83.2766 |
Note: challenge test set COMET values are reported by the organizers on a 0–100 scale, while devtest COMET values use the standard 0–1 scale.
NERDS-NITS at WMT 2026: Pivot-Based and Direct Neural Machine Translation for Arabic–Asian Low-Resource Language Pairs
We thank the WMT 2026 shared task organizers for providing the datasets and evaluation infrastructure used in this work.