Runs On
What it takes to serve wav2vec2-xls-r-parlaspeech-hr (315M parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.
Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the SAVRN Index, read Sep 18, 2026.
SAVRN's Notes on wav2vec2-xls-r-parlaspeech-hr
Eight tenths of a gigabyte. That is all this Croatian speech recognizer needs in memory at 16-bit, with the weights at 0.6 GB, so the cheapest setup we list, one MI300X with 192 GB at $1.85 an hour, leaves nearly the whole card idle. Nobody should dedicate that accelerator to a 315M parameter transcription model; share the card. CLASSLA fine-tuned it from facebook/wav2vec2-xls-r-300m on 300 hours of Croatian parliament recordings from ParlaSpeech-HR v1.0, so its home is Croatian speech in a parliamentary register.
The license field on this listing is blank, and that is the first thing to settle before a deployment, since open access to the files is not the same as permission to use them commercially; ask the publisher what terms apply. The weights ship as float32 and the file set totals 5.0 GB, so plan storage on that figure, not the 0.6 GB 16-bit footprint.
SAVRN Research, 2026-09-18
Model Card
This model for Croatian ASR is based on the facebook/wav2vec2-xls-r-300m model and was fine-tuned with 300 hours of recordings and transcripts from the ASR Croatian parliament dataset ParlaSpeech-HR v1.0. If you use this model, please cite the following paper: Nikola Ljubešić, Danijel Koržinek, Peter Rupnik, Ivo-Pavao Jazbec. ParlaSpeech-HR -- a freely available ASR dataset for Croatian bootstrapped from the ParlaMint corpus. http://www.lrec-conf.org/proceedings/lrec2022/workshops/ParlaCLARINIII/pdf/2022.parlaclariniii-1.16.pdf Evaluation is performed on the dev and test portions of the ParlaSpeech-HR v1.0 dataset. There are multiple models available, and in terms of CER and WER, the…
Excerpt from the card by CLASSLA - CLARIN Knowledge Centre for South Slavic Languages.
Configuration
- Architecture
- Wav2Vec2ForCTC
- Layers
- 24
- Hidden size
- 1,024
- Feed-forward size
- 4,096
- Attention heads
- 16
- Vocabulary size
- 50
- Stored precision
- float32
- Model type
- wav2vec2
Identity and Version
- Repository
- classla/wav2vec2-xls-r-parlaspeech-hr
- Publisher
- CLASSLA - CLARIN Knowledge Centre for South Slavic Languages
- Task
- Speech recognition
- Modality
- Audio
- Library
- transformers
- Parameters
- 315M parameters
- Languages
- hr
- Revision
- a6075e0c9e0c49a533dc61876b1eafc6bc78de92
- First published
- 2022-03-02
- Last updated
- 2025-07-02
Files and Weights
18 files, 5.0 GB in total. The weights are 7 files totalling 5.0 GB in bin, pt, pth, safetensors.
Weights7 files · 5.0 GB
Configuration3 files · 10.2 KB
Tokenizer2 files · 583 B
Documentation1 file · 3.2 KB
Other4 files · 627.3 KB
Repository1 file · 1.2 KB
Every file
License and Download
- License
- Not stated by the source
- Access
- Open weights, no gate
- Download size
- 5.0 GB
Built From
- Trained on (disclosed)
parlaspeech-hr
Memory Requirements
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About wav2vec2-xls-r-parlaspeech-hr
How much GPU memory does wav2vec2-xls-r-parlaspeech-hr need?
About 0.8 GB at 16-bit and 0.2 GB at 4-bit: the weights (315M parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run wav2vec2-xls-r-parlaspeech-hr on?
At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.
Similar Models
Model · Speech recognition
Chem
IPA phoneme recognition for Kölsch (Ripuarian German, Cologne). Fine-tuned from facebook/wav2vec2-xls-r-300m with a CTC head over a 48-symbol IPA inventory. This is also the forced-alignment model. Its frame-level CTC posteriors drive torchaudio.functional.forcedalign to produce Praat TextGrids with word and phone tiers — see notebook 9 in the repository. Trained during the CIF Tandem Fellowship at IfL-Phonetik, University of Cologne, on Alles Kölsch (Bhatt & Lindlar 1998) — 4,670 utterances, 4.5 hours of spontaneous Cologne dialect speech from 105 speakers aged 10–88 across 49 neighbourhoods. Kölsch is Ripuarian German. It has no public speech dataset and no standardised spelling: 94.1 %…
Open weights
cc-by-nc-sa-4.0
315M parameters
transformers
Finetuned version of KBs VoxRex large model using Swedish radio broadcasts, NST and Common Voice data. Evalutation without a language model gives the following: WER for NST + Common Voice test set (2% of total sentences) is 2.5%. WER for Common Voice test set is 8.49% directly and 7.37% with a 4-gram language model. When using this model, make sure that your speech input is sampled at 16kHz. This model has been fine-tuned for 120000 updates on NST + CommonVoice and then for an additional 20000 updates on CommonVoice only. The additional fine-tuning on CommonVoice hurts performance on the NST+CommonVoice test set somewhat and, unsurprisingly, improves it on the CommonVoice test set. It seems…
Open weights
cc0-1.0
315M parameters
transformers
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the commonvoice 8.0 dataset as well as other datasets listed below. It achieves the following results on the evaluation set: The eval.py script results using a LM are: Fine-tuned facebook/wav2vec2-large-xlsr-53 on Czech using the Common Voice dataset. When using this model, make sure that your speech input is sampled at 16kHz. The model can be used directly (without a language model) as follows: The model can be evaluated using the attached eval.py script: The Common Voice 8.0 train and validation datasets were used for training, as well as the following datasets: - Šmídl, Luboš and Pražák, Aleš, 2013, OVM – Otázky…
Open weights
apache-2.0
315M parameters
transformers
A fine-tuned XLS-R 300M CTC model for Urdu automatic speech recognition. It transcribes 16 kHz mono audio and includes an optional 5-gram KenLM decoder. Best reported result: 39.89% WER / 16.70% CER with KenLM decoding on the Urdu Common Voice 8.0 test set. See the Kaggle evaluation notebook for a reproducible example. The repository contains a 5-gram KenLM language model. The Kaggle notebook evaluates a five-sample streaming smoke test from fixie-ai/commonvoice170 (ur, test). Results are reported on the Urdu test split of Mozilla Common Voice 8.0. The language-model row is the model-card score; compare each result only with the same decoding strategy. To reproduce language-model evaluation…
Open weights
apache-2.0
315M parameters
transformers
This model is a finetuned version of facebook/wav2vec2-large-xlsr-53. There was an issue with vocab, seems like there are special characters included and they were not considered during training You could try
Open weights
apache-2.0
315M parameters
transformers
ATTENTION! Metrics (float16) using evaluate library with batchsize=1
Open weights
apache-2.0
315M parameters
transformers
