trumancai/Revela-135M

Model Summary

Revela-135M is a compact (135 M-parameter) variant of the Revela dense-retriever.
It is ideal for resource-constrained environments while maintaining strong general-domain retrieval quality.
Training used the same 320 K Wikipedia batches with in-batch attention.

• Repository: TRUMANCFY/Revela
• Training Dataset: trumancai/revela_training_corpus

Other Links

Binary	Description
trumancai/Revela-1b	1 B-parameter variant (LLaMA-3.2-1B backbone).
trumancai/Revela-500M	500 M-parameter variant (Qwen2.5-0.5B backbone).
trumancai/Revela-135M	← current repo
trumancai/Revela-code-1b	1 B-parameter code-retriever.
trumancai/Revela-code-500M	500 M-parameter code-retriever.
trumancai/Revela-code-135M	135 M-parameter code-retriever.
trumancai/revela_training_corpus	Wikipedia training corpus.
trumancai/revela_code_training_corpus	Code training corpus.

Usage

from mteb.model_meta import ModelMeta
from mteb.models.repllama_models import RepLLaMAWrapper, _loader
import mteb, torch

revela_smol_135m = ModelMeta(
    loader=_loader(
        RepLLaMAWrapper,
        base_model_name_or_path="HuggingFaceTB/SmolLM2-135M",
        peft_model_name_or_path="trumancai/Revela-135M",
        device_map="auto",
        torch_dtype=torch.bfloat16,
    ),
    name="trumancai/Revela-135M",
    languages=["eng_Latn"],
    open_source=True,
    revision="c84848e2708dee28e9a58edaed78867537b489e3",
    release_date="2024-04-13",
)
model = revela_smol_135m.loader()

mteb.MTEB(tasks=["SciFact", "NFCorpus"]).run(model=model, output_folder="results/Revela-135M")

License

Citation