Skip to content
Audio

vaakku-whisper-small

LoRA-адаптер для индийской речи

1 min read

Это LoRA-адаптер для распознавания многоязычной речи на базе openai/whisper-small от автора rudra11. Поддерживает тамильский, телугу, хинди, каннада и малаялам; обучен через Hugging Face PEFT на google/fleurs.

Ключевые особенности

  • Пять языков: ta, te, hi, kn, ml.
  • LoRA-адаптер, а не полный fine-tune: rank 16, alpha 32, dropout 0.05, target-модули q_proj и v_proj.
  • База — openai/whisper-small; обучение в bf16.
  • Оценка на google/fleurs: micro WER 0.3553, micro CER 0.2966; macro WER 0.3189, macro CER 0.2460.
  • Репозиторий содержит PEFT-адаптер и processor, а не слитые веса Whisper.

Пример

from peft import PeftModel
from transformers import WhisperForConditionalGeneration, WhisperProcessor

base = WhisperForConditionalGeneration.from_pretrained('openai/whisper-small')
model = PeftModel.from_pretrained(base, 'rudra11/vaakku-whisper-small')
processor = WhisperProcessor.from_pretrained('rudra11/vaakku-whisper-small')
ПодробнееЛучше всего по WER: каннада — 0.2153, хинди — 0.2430, тамильский — 0.2366; хуже всего малаялам — WER 0.6046.
Micro WER 0.3553 и micro CER 0.2966; macro WER 0.3189 и macro CER 0.2460.
Требования к VRAM/железу в карточке не указаны. Это исследовательский/демонстрационный адаптер, не production и не SOTA; языки вне пяти указанных не поддерживаются.
Discuss Telegram 0 16

Will it run?

Working out the memory needs

The bot is still collecting this model's weight sizes; the estimate will appear here in a few minutes. Meanwhile you can rent a GPU with headroom.

Massed ComputeOpen Massed Compute

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # rudra11/vaakku-whisper-small
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download rudra11/vaakku-whisper-small --local-dir ~/models/vaakku-whisper-small
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to vaakku-whisper-small
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

New posts about versions of vaakku-whisper-small (quants, finetunes, LoRAs) will show up under the bell.

No known relatives of this model yet. All derivatives on Hugging Face

About the model

rudra11/vaakku-whisper-small
Author
rudra11
Category
Audio
Released
Posted

On NeyroLearn

Hugging Face

160

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments