GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
SignvrseOfficial
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
SignvrseOfficial
Model Tree
allenai/OLMo-2-0425-1B-Instruct
Input Modalities
Output Modalities
Supported Functionality
| Model | Token F1 | Exact Match |
|---|---|---|
| Teacher (Gemma2-2B v6) | 0.616 | 14.0% |
| This model | 0.618 | 14.3% |
Training tracked on Weights & Biases.
markdown
Translate the following sentence into Kenyan Sign Language (KSL) glosses.<English sentence>
python
import torchfrom peft import PeftModelfrom transformers import AutoModelForCausalLM, AutoTokenizerBASE = "allenai/OLMo-2-0425-1B-Instruct"ADAPTER = "SignvrseOfficial/Glosser_OLMo2_1B_it_v1"tokenizer = AutoTokenizer.from_pretrained(ADAPTER)base = AutoModelForCausalLM.from_pretrained(BASE, torch_dtype=torch.float16, device_map="auto")model = PeftModel.from_pretrained(base, ADAPTER)model.eval()sentence = "That house is ours."messages = [{"role": "user","content": ("Translate the following sentence into Kenyan Sign Language "f"(KSL) glosses.\n\n{sentence}"),}]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=128, do_sample=False)gloss = tokenizer.decode(out[0][inputs["input_ids"].shape[1]:], skip_special_tokens=True)print(gloss)