# wav2vec2-xlsr-53-russian-emotion-recognition by Aniemore
Source: https://savrn.com/models/wav2vec2-xlsr-53-russian-emotion-recognition
Markdown alternate of the page above; the site index is https://savrn.com/llms.txt

---

## Runs On

What it takes to serve wav2vec2-xlsr-53-russian-emotion-recognition (316M parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.

| Precision | Weights | Memory needed | Cheapest setup | Per hour | Also fits |
| --- | --- | --- | --- | --- | --- |
| 16-bit | 0.6 GB | 0.8 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 |
| 8-bit | 0.3 GB | 0.4 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 |
| 4-bit | 0.2 GB | 0.2 GB | 1x [MI300X](https://savrn.com/ai-index/pricing/gpus/mi300x) (192 GB) Vultr | $1.85 | [1x H100](https://savrn.com/ai-index/pricing/gpus/h100) $1.99 · [1x MI325X](https://savrn.com/ai-index/pricing/gpus/mi325x) $2.00 |

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](https://savrn.com/ai-index/pricing/gpus), read Oct 23, 2026.

[wav2vec2-xlsr-53-russian-emotion-recognition on every accelerator the SAVRN Index prices, at every precision](https://savrn.com/models/wav2vec2-xlsr-53-russian-emotion-recognition/gpus)

## Model Card

By Aniemore, published under mit, revision 3a4ab5c8bef1.

Speech emotion recognition for Russian over seven classes: anger, disgust, enthusiasm, fear, happiness, neutral, sadness.

Fine-tuned from [jonatasgrosman/wav2vec2-large-xlsr-53-russian](https://savrn.com/models/wav2vec2-large-xlsr-53-russian) on [Aniemore/resd](https://huggingface.co/datasets/Aniemore/resd).

- Parameters: 316M · weights: 1207 MiB (fp32)
- Input: 16 kHz mono waveform
- Quantized builds: [wav2vec2-xlsr-53-russian-emotion-recognition-quantized](https://huggingface.co/Aniemore/wav2vec2-xlsr-53-russian-emotion-recognition-quantized) — INT8, FP8 and INT4, up to 5.8x smaller on disk at the same score

### Results

| Test set | n | UA | WA | macro-F1 |
| --- | --- | --- | --- | --- |
| RESD test | 280 | 0.7176 | 0.7214 | 0.7206 |
| Dusha podcast test | 12079 | 0.3315 | 0.0804 | 0.0954 |
| CAMEO test | 5187 | 0.2370 | 0.2603 | 0.2248 |

### Usage

[Read the full model card (453 words)](https://savrn.com/models/wav2vec2-xlsr-53-russian-emotion-recognition/card)

## Configuration

Architecture

Wav2Vec2ForSpeechClassification

Layers

24

Hidden size

1,024

Feed-forward size

4,096

Attention heads

16

Vocabulary size

39

Stored precision

float32

Model type

wav2vec2

## Identity and Version

Repository

Aniemore/wav2vec2-xlsr-53-russian-emotion-recognition

Publisher

Aniemore

Task

Audio classification

Modality

Audio

Library

transformers

Parameters

316M parameters

Languages

ru

Revision

3a4ab5c8bef1495ddb8a5ba801234fa37a2954e7

First published

2022-05-22

Last updated

2026-08-01

## Files and Weights

20 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

Configuration7 files · 550.1 KB

Tokenizer2 files · 907 B

Documentation1 file · 5.5 KB

Other2 files · 17.2 KB

Repository1 file · 1.2 KB

Every file

| File | Type | Size | SHA-256 |
| --- | --- | --- | --- |
| model.safetensors | Weights | 1.3 GB | d8c49be01c68 |
| optimizer.pt | Weights | 2.5 GB | 3e255efc4594 |
| pytorch_model.bin | Weights | 1.3 GB | 7b9eef373c63 |
| rng_state.pth | Weights | 14.6 KB | 82ccc90de7cb |
| scaler.pt | Weights | 559 B | c535e3b51d86 |
| scheduler.pt | Weights | 623 B | d5ca2170e70a |
| training_args.bin | Weights | 3.3 KB | 96af3c300ad8 |
| __init__.py | Configuration |  | — |
| config.json | Configuration | 2.9 KB | — |
| preprocessor_config.json | Configuration | 256 B | — |
| special_tokens_map.json | Configuration | 85 B | — |
| trainer_state.json | Configuration | 542.2 KB | — |
| wav2vec2fsr_config.py | Configuration | 183 B | — |
| wav2vec2speechclassification.py | Configuration | 4.4 KB | — |
| README.md | Documentation | 5.5 KB | — |
| assets/confusion.svg | Other | 14.3 KB | — |
| assets/panel.svg | Other | 2.8 KB | — |
| .gitattributes | Repository | 1.2 KB | — |
| tokenizer_config.json | Tokenizer | 520 B | — |
| vocab.json | Tokenizer | 387 B | — |

## License and Download

License

mit

Access

Open weights, no gate

Download size

5.0 GB

[Download from Aniemore](https://huggingface.co/Aniemore/wav2vec2-xlsr-53-russian-emotion-recognition)

Released by Aniemore through its official repository on Hugging Face. [Read the license](https://opensource.org/license/mit).

## Built From

- Derived from [jonatasgrosman/wav2vec2-large-xlsr-53-russian](https://savrn.com/models/wav2vec2-large-xlsr-53-russian)
- Trained on (disclosed) Aniemore/resd
- Trained on (disclosed) Aniemore/resd_annotated

## Evaluations

Each result is shown as reported, with the conditions its reporter stated. None is a SAVRN measurement. A comparison lines two results up only when their configuration, unit and setup are all stated and identical.

| Benchmark | Conditions | Result | Reported by | Revision | Date |
| --- | --- | --- | --- | --- | --- |
| CAMEO test | Task Speech Emotion RecognitionMetric AccuracyComparison conditions not established | 0.2603 | [Aniemore](https://huggingface.co/Aniemore/wav2vec2-xlsr-53-russian-emotion-recognition) Publisher reported | Evaluated revision not stated | — |
| CAMEO test | Task Speech Emotion RecognitionMetric Macro F1Comparison conditions not established | 0.2248 | [Aniemore](https://huggingface.co/Aniemore/wav2vec2-xlsr-53-russian-emotion-recognition) Publisher reported | Evaluated revision not stated | — |
| CAMEO test | Task Speech Emotion RecognitionMetric Unweighted accuracyComparison conditions not established | 0.237 | [Aniemore](https://huggingface.co/Aniemore/wav2vec2-xlsr-53-russian-emotion-recognition) Publisher reported | Evaluated revision not stated | — |
| Dusha podcast test | Task Speech Emotion RecognitionMetric AccuracyComparison conditions not established | 0.0804 | [Aniemore](https://huggingface.co/Aniemore/wav2vec2-xlsr-53-russian-emotion-recognition) Publisher reported | Evaluated revision not stated | — |
| Dusha podcast test | Task Speech Emotion RecognitionMetric Macro F1Comparison conditions not established | 0.0954 | [Aniemore](https://huggingface.co/Aniemore/wav2vec2-xlsr-53-russian-emotion-recognition) Publisher reported | Evaluated revision not stated | — |
| Dusha podcast test | Task Speech Emotion RecognitionMetric Unweighted accuracyComparison conditions not established | 0.3315 | [Aniemore](https://huggingface.co/Aniemore/wav2vec2-xlsr-53-russian-emotion-recognition) Publisher reported | Evaluated revision not stated | — |
| RESD test | Task Speech Emotion RecognitionMetric AccuracyComparison conditions not established | 0.7214 | [Aniemore](https://huggingface.co/Aniemore/wav2vec2-xlsr-53-russian-emotion-recognition) Publisher reported | Evaluated revision not stated | — |
| RESD test | Task Speech Emotion RecognitionMetric Macro F1Comparison conditions not established | 0.7206 | [Aniemore](https://huggingface.co/Aniemore/wav2vec2-xlsr-53-russian-emotion-recognition) Publisher reported | Evaluated revision not stated | — |
| RESD test | Task Speech Emotion RecognitionMetric Unweighted accuracyComparison conditions not established | 0.7176 | [Aniemore](https://huggingface.co/Aniemore/wav2vec2-xlsr-53-russian-emotion-recognition) Publisher reported | Evaluated revision not stated | — |

## Memory Requirements

| Precision | Weights in memory |
| --- | --- |
| As published | 5.0 GB |
| 16-bit | 0.6 GB |
| 8-bit | 0.3 GB |
| 4-bit | 0.2 GB |

Weights only, from the published parameter count; the key-value cache and runtime add to this.

## Compare wav2vec2-xlsr-53-russian-emotion-recognition

- [wav2vec2-large-robust-24-ft-age-gender vs wav2vec2-xlsr-53-russian-emotion-recognition](https://savrn.com/models/compare/wav2vec2-large-robust-24-ft-age-gender-vs-wav2vec2-xlsr-53-russian-emotion-recognition)

## Questions About wav2vec2-xlsr-53-russian-emotion-recognition

### How much GPU memory does wav2vec2-xlsr-53-russian-emotion-recognition need?

About 0.8 GB at 16-bit and 0.2 GB at 4-bit: the weights (316M parameters) plus a working margin. A long context needs more.

### What is the cheapest GPU to run wav2vec2-xlsr-53-russian-emotion-recognition 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.

### Can I use wav2vec2-xlsr-53-russian-emotion-recognition commercially?

Yes. wav2vec2-xlsr-53-russian-emotion-recognition is released under MIT License. The MIT License is a short permissive license. It permits commercial use, modification and redistribution, provided the copyright notice and permission notice are included.

## Similar Models

Model · Audio classification

### [wav2vec2-lg-xlsr-en-speech-emotion-recognition](https://savrn.com/models/wav2vec2-lg-xlsr-en-speech-emotion-recognition)

[Enrique Hernández Calabrés](https://savrn.com/model-publishers/ehcalabres)

The model is a fine-tuned version of jonatasgrosman/wav2vec2-large-xlsr-53-english for a Speech Emotion Recognition (SER) task. The dataset used to fine-tune the original pre-trained model is the RAVDESS dataset. This dataset provides 1440 samples of recordings from actors performing on 8 different emotions in English, which are: It achieves the following results on the evaluation set: The following hyperparameters were used during training: - learningrate: 0.0001 - trainbatchsize: 4 - evalbatchsize: 4 - gradientaccumulationsteps: 2 - totaltrainbatchsize: 8 - lrschedulertype: linear - numepochs: 3 - mixedprecisiontraining: Native AMP Any doubt, contact me on Twitter. - Transformers 4.8.2…

Open weights apache-2.0 316M parameters transformers

[View model](https://savrn.com/models/wav2vec2-lg-xlsr-en-speech-emotion-recognition)

Model · Audio classification

### [wavlm-emotion-russian-resd](https://savrn.com/models/wavlm-emotion-russian-resd)

[Aniemore](https://savrn.com/model-publishers/aniemore)

Speech emotion recognition for Russian over seven classes: anger, disgust, enthusiasm, fear, happiness, neutral, sadness. Fine-tuned from jonatasgrosman/expw2v2truwavlms363 on Aniemore/resd. Audio resampled to 16 kHz mono, clips capped at 12 s, normalized per utterance, padding masked. UA is macro-averaged recall, WA is accuracy, F1 is macro-averaged. All three test sets went through the same harness, so the rows are comparable to each other. The RESD split matches fold 1 of EmoBox bit for bit. The top entry there is WavLM-large at WA 56.47 / UA 55.87 / F1 55.82. These numbers are higher, but the training protocol differs — EmoBox freezes the encoder and trains a probe, this is a full…

Open weights mit 317M parameters transformers

[View model](https://savrn.com/models/wavlm-emotion-russian-resd)

Model · Audio classification

### [audio_cls_1016](https://savrn.com/models/audio-cls-1016-2)

[Ryusunghwan](https://savrn.com/model-publishers/ryusunghwan)

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on an unknown dataset. It achieves the following results on the evaluation set: The following hyperparameters were used during training: - learningrate: 0.0001 - trainbatchsize: 8 - evalbatchsize: 8 - gradientaccumulationsteps: 4 - totaltrainbatchsize: 32 - lrschedulertype: linear - numepochs: 10 - mixedprecisiontraining: Native AMP - Transformers 5.18.0 - Pytorch 2.11.0+cu130 - Datasets 4.8.5 - Tokenizers 0.23.2

Open weights apache-2.0 316M parameters transformers

[View model](https://savrn.com/models/audio-cls-1016-2)

Model · Audio classification

### [audio_cls_1016](https://savrn.com/models/audio-cls-1016-4)

[Ju ung](https://savrn.com/model-publishers/juran123)

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on an unknown dataset. It achieves the following results on the evaluation set: The following hyperparameters were used during training: - learningrate: 0.0001 - trainbatchsize: 8 - evalbatchsize: 8 - gradientaccumulationsteps: 4 - totaltrainbatchsize: 32 - lrschedulertype: linear - numepochs: 10 - mixedprecisiontraining: Native AMP - Transformers 5.18.0 - Pytorch 2.11.0+cu130 - Datasets 4.8.5 - Tokenizers 0.23.2

Open weights apache-2.0 316M parameters transformers

[View model](https://savrn.com/models/audio-cls-1016-4)

Model · Audio classification

### [hubert-large-speech-emotion-recognition-russian-dusha-finetuned](https://savrn.com/models/hubert-large-speech-emotion-recognition-russian-dusha-finetuned)

[Phil](https://savrn.com/model-publishers/xbgoose)

The pre-trained model is this one - facebook/hubert-large-ls960-ft The DUSHA dataset used can be found here Fine-tuned in Google Colab using Pro account with A100 GPU Freezed all layers exept projector, classifier and all 24 HubertEncoderLayerStableLayerNorm layers Used half of the train dataset - 2 epochs - train batch size = 8 - eval batch size = 8 - gradient accumulation steps = 4 - learning rate = 5e-5 without warm up and decay Achieved - accuracy = 0.86 - balanced = 0.76 - macro f1 score = 0.81 on test set, improving accucary and f1 score compared to dataset baseline

Open weights apache-2.0 316M parameters transformers

[View model](https://savrn.com/models/hubert-large-speech-emotion-recognition-russian-dusha-finetuned)

Model · Audio classification

### [wav2vec2-large-xlsr-53-gender-recognition-librispeech](https://savrn.com/models/wav2vec2-large-xlsr-53-gender-recognition-librispeech)

[Alef Iury Siqueira Ferreira](https://savrn.com/model-publishers/alefiury)

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on Librispeech-clean-100 for gender recognition. It achieves the following results on the evaluation set: The Librispeech-clean-100 dataset was used to train the model, with 70% of the data used for training, 10% for validation, and 20% for testing. The following hyperparameters were used during training: - learningrate: 3e-05 - trainbatchsize: 4 - evalbatchsize: 4 - gradientaccumulationsteps: 4 - totaltrainbatchsize: 16 - lrschedulertype: linear - lrschedulerwarmupratio: 0.1 - numepochs: 1 - mixedprecisiontraining: Native AMP - Transformers 4.28.0 - Pytorch 2.0.0+cu118 - Tokenizers 0.13.3

Open weights apache-2.0 316M parameters transformers

[View model](https://savrn.com/models/wav2vec2-large-xlsr-53-gender-recognition-librispeech)

## Aniemore

[All models and datasets](https://savrn.com/model-publishers/aniemore)

## Versions

- [3a4ab5c8bef1](https://savrn.com/models/wav2vec2-xlsr-53-russian-emotion-recognition/versions/3a4ab5c8bef1) · current 2026-09-18

## Explore More

- [All audio classification models](https://savrn.com/models/tasks/audio-classification)
- [All models under mit](https://savrn.com/models/licenses/mit)
- [Model comparisons](https://savrn.com/models/comparisons)
- [The model directory](https://savrn.com/models)
- [Open model prices by host](https://savrn.com/ai-index/pricing/open-models)

## Source

- Repository metadata, read 2026-10-21.
- [Hugging Face record](https://huggingface.co/Aniemore/wav2vec2-xlsr-53-russian-emotion-recognition)
- [How the hub is built](https://savrn.com/model-hub/methodology)