GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
titaniumbones
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Run this model inference with full control and performance in your environment.
Model Details
Model Provider
titaniumbones
Model Tree
openai/whisper-small
Input Modalities
Output Modalities
Supported Functionality
| Model | Raw WER | Normalized WER | Test clips |
|---|---|---|---|
| whisper-base fine-tuned | 105.7% | 94.1% | 300 |
| whisper-base zero-shot | 107.6% | 134.7% | 300 |
| whisper-small fine-tuned | 46.7% | 38.9% | 300 |
| whisper-small zero-shot | 110.1% | 100.6% | 300 |
Zero-shot Whisper barely functions on Maltese (≈1 h of it in the
original training mix); the WER delta above is what a few hours of
community-donated Common Voice speech and a laptop fine-tune buy. WER
computed on the official Common Voice test split (speaker-disjoint from
training data — note the training split draws on only 6
speakers, which limits generalization; see Limitations). "Normalized"
applies BasicTextNormalizer (lowercase, strip punctuation) to both
reference and hypothesis.
mozilla-foundation/common_voice_* datasets on this
Hub are empty shells).Standard Hugging Face Seq2SeqTrainer recipe
(fine-tune-whisper):
batch 4 × grad-accum 4, lr 1e-05, warmup 50,
1000 steps, fp32 (MPS), gradient checkpointing, greedy decoding,
best-checkpoint-by-WER selection.
python
from transformers import pipelineasr = pipeline("automatic-speech-recognition", model="titaniumbones/whisper-small-mt-commonvoice")print(asr("clip.mp3", generate_kwargs={"language": "maltese"}))