GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
sinhala-nlp
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
sinhala-nlp
Model Tree
Qwen/Qwen3.5-4B
Input Modalities
Output Modalities
Supported Functionality
python
from transformers import AutoTokenizer, AutoModelForCausalLMfrom peft import PeftModeltok = AutoTokenizer.from_pretrained("Qwen/Qwen3.5-4B")model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3.5-4B", dtype="auto", device_map="auto")model = PeftModel.from_pretrained(model, "sinhala-nlp/Qwen3.5-4B-NSINA-Headlines-en")
Prompts use the base model's chat template with thinking disabled, and the assistant
response begins with the Headline: prefix. Articles are trimmed to
2500 characters and then budgeted to fit max_seq_len from the lead.
| Training articles | 8000 |
| Instruction language | en |
| Epochs | 1.0 |
| Effective batch size | 16 |
| Learning rate | 0.0002 |
| Max sequence length | 2560 |
| LoRA r / alpha / dropout | 16 / 32 / 0.05 |
| Thinking during training | False |
First 1000 instances of the NSINA-Headlines test split, ROUGE F1 x100 with
whitespace tokenization (the default rouge_score tokenizer strips non-ASCII and
zeroes out every Sinhala score):
| Metric | Score |
|---|---|
| ROUGE-1 | 28.80 |
| ROUGE-2 | 13.36 |
| ROUGE-L | 28.13 |
This adapter inherits the licence of the base model; check the base model card before redistributing. The training data is NSINA-Headlines, derived from scraped Sri Lankan news content -- verify its terms on the dataset card, as they may be more restrictive than the base model's licence.