# whisper-small-malayalam by Sajil C K: Open-Weight Model
Source: https://savrn.com/models/whisper-small-malayalam
Markdown alternate of the page above; the site index is https://savrn.com/llms.txt

---

## Runs On

What it takes to serve whisper-small-malayalam (242M 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 | 0.5 GB | 0.6 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 | 0.2 GB | 0.3 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 | 0.1 GB | 0.1 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 9, 2026.

[whisper-small-malayalam on every accelerator the SAVRN Index prices, at every precision](https://savrn.com/models/whisper-small-malayalam/gpus)

## Model Card

By Sajil C K, published under apache-2.0, revision dbd72416303a.

Fine-tuned version of [openai/whisper-small](https://savrn.com/models/whisper-small) on a multi-corpus Malayalam speech dataset.

### Model Description

- Base model: openai/whisper-small (244M parameters)
- Language: Malayalam (ml)
- Task: Automatic Speech Recognition (transcription)
- Training steps: 3500
- Best WER: 37.64% on CommonVoice 25 Malayalam test set
- Leakage-free multi-source test: 47.8% WER / 13.6% CER (231 clips across 6 sources, none with a transcript in the training split — see Multi-source evaluation)
- CPU speed (Transformers, FP32, 4 vCPU): RTF 1.96, i.e. slower than real time. For CPU deployment use the whisper.cpp builds

### Training Data

The model was trained on an aggregated corpus of 5 Malayalam speech datasets, combined and published as [sajilck/malayalam-asr-corpus](https://huggingface.co/datasets/sajilck/malayalam-asr-corpus).

| Corpus | Source | Domain | Access |
| --- | --- | --- | --- |
| [IMaSC](https://huggingface.co/datasets/thennal/imasc) | thennal/imasc | TTS / Read speech | HuggingFace |
| [SMC Malayalam Speech Corpus](https://github.com/smcproject/malayalam-speech-corpus) | sajilck/smc-malayalam-speech-corpus | Read speech | Kaggle |
| [IndicTTS Malayalam](https://www.iitm.ac.in/donlab/indictts/) | kavyamanohar/indic-tts-malayalam-speech-corpus | TTS / Read speech | Kaggle |
| [OpenSLR 63](https://openslr.org/63/) | sajilck/openslr63 | Crowdsourced | Kaggle |
| [CommonVoice 25 Malayalam](https://commonvoice.mozilla.org/) | sajilck/common-voice-malayalam | Crowdsourced | Kaggle |

Total: ~86,000 samples across TTS-recorded, read speech, and crowdsourced domains.

### Benchmark Results

[Read the full model card (2,020 words)](https://savrn.com/models/whisper-small-malayalam/card)

## Configuration

Architecture

WhisperForConditionalGeneration

Layers

12

Vocabulary size

51,865

Model type

whisper

## Identity and Version

Repository

sajilck/whisper-small-malayalam

Publisher

Sajil C K

Task

Speech recognition

Modality

Audio

Library

Not stated by the source

Parameters

242M parameters

Languages

ml

Revision

dbd72416303aa36b0fecf1d87bee6b0c9b20030f

First published

2026-06-11

Last updated

2026-10-09

## Files and Weights

23 files, 4.8 GB in total. The weights are 11 files totalling 4.8 GB in bin, pt, pth, safetensors.

Weights11 files · 4.8 GB

Configuration8 files · 16.7 KB

Tokenizer2 files · 3.9 MB

Documentation1 file · 17.0 KB

Repository1 file · 1.5 KB

Every file

| File | Type | Size | SHA-256 |
| --- | --- | --- | --- |
| ggml/ggml-model-f16.bin | Weights | 487.6 MB | 3d95f02e782f |
| ggml/ggml-model-q5_0.bin | Weights | 175.2 MB | 38c42ff640f2 |
| ggml/ggml-model-q8_0.bin | Weights | 264.5 MB | 81d784cbf858 |
| last-checkpoint/model.safetensors | Weights | 967.0 MB | 4bf4160946b6 |
| last-checkpoint/optimizer.pt | Weights | 1.9 GB | b5fd9c42629a |
| last-checkpoint/rng_state.pth | Weights | 14.6 KB | 76462a3c6db6 |
| last-checkpoint/scaler.pt | Weights | 1.4 KB | 0ff58b41c367 |
| last-checkpoint/scheduler.pt | Weights | 1.5 KB | b93466c1593a |
| last-checkpoint/training_args.bin | Weights | 5.4 KB | 0a357fe12905 |
| model.safetensors | Weights | 967.0 MB | de719523b46e |
| training_args.bin | Weights | 5.4 KB | 0a357fe12905 |
| config.json | Configuration | 1.3 KB | — |
| generation_config.json | Configuration | 3.8 KB | — |
| last-checkpoint/config.json | Configuration | 1.3 KB | — |
| last-checkpoint/generation_config.json | Configuration | 4.6 KB | — |
| last-checkpoint/preprocessor_config.json | Configuration | 315 B | — |
| last-checkpoint/trainer_state.json | Configuration | 4.7 KB | — |
| preprocessor_config.json | Configuration | 315 B | — |
| processor_config.json | Configuration | 409 B | — |
| README.md | Documentation | 17.0 KB | — |
| .gitattributes | Repository | 1.5 KB | — |
| tokenizer.json | Tokenizer | 3.9 MB | — |
| tokenizer_config.json | Tokenizer | 2.1 KB | — |

## License and Download

License

apache-2.0

Access

Open weights, no gate

Download size

4.8 GB

[Download from Sajil C K](https://huggingface.co/sajilck/whisper-small-malayalam)

Released by Sajil C K through its official repository on Hugging Face. [Read the license](https://www.apache.org/licenses/LICENSE-2.0).

## Built From

- Derived from [openai/whisper-small](https://savrn.com/models/whisper-small)
- Trained on (disclosed) sajilck/malayalam-asr-corpus

## 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 |
| --- | --- | --- | --- | --- | --- |
| Common Voice 25 (Malayalam) | Configuration mlTask Automatic Speech RecognitionMetric WERComparison conditions not established | 37.64 | [sajilck](https://huggingface.co/sajilck/whisper-small-malayalam) Publisher reported | Evaluated revision not stated | — |
| malayalam-asr-corpus test, leakage-filtered (231-clip source-balanced sample) | Task Automatic Speech RecognitionMetric CER (normalised)Comparison conditions not established | 13.56 | [sajilck](https://huggingface.co/sajilck/whisper-small-malayalam) Publisher reported | Evaluated revision not stated | — |
| malayalam-asr-corpus test, leakage-filtered (231-clip source-balanced sample) | Task Automatic Speech RecognitionMetric WER (normalised)Comparison conditions not established | 47.82 | [sajilck](https://huggingface.co/sajilck/whisper-small-malayalam) Publisher reported | Evaluated revision not stated | — |

## Memory Requirements

| Precision | Weights in memory |
| --- | --- |
| As published | 4.8 GB |
| 16-bit | 0.5 GB |
| 8-bit | 0.2 GB |
| 4-bit | 0.1 GB |

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

## Questions About whisper-small-malayalam

### How much GPU memory does whisper-small-malayalam need?

About 0.6 GB at 16-bit and 0.1 GB at 4-bit: the weights (242M parameters) plus a working margin. A long context needs more.

### What is the cheapest GPU to run whisper-small-malayalam 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 whisper-small-malayalam commercially?

Yes. whisper-small-malayalam is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.

## Similar Models

Model · Speech recognition

### [whisper-small](https://savrn.com/models/whisper-small)

[OpenAI](https://savrn.com/model-publishers/openai)

Whisper is a pre-trained model for automatic speech recognition (ASR) and speech translation. Trained on 680k hours of labelled data, Whisper models demonstrate a strong ability to generalise to many datasets and domains without the need for fine-tuning. Whisper was proposed in the paper Robust Speech Recognition via Large-Scale Weak Supervision by Alec Radford et al from OpenAI. The original code repository can be found here. Disclaimer: Content for this model card has partly been written by the Hugging Face team, and parts of it were copied and pasted from the original model card. Whisper is a Transformer based encoder-decoder model, also referred to as a sequence-to-sequence model. It…

Open weights apache-2.0 242M parameters transformers

[View model](https://savrn.com/models/whisper-small)

Model · Speech recognition

### [whisper-small-ha-merged](https://savrn.com/models/whisper-small-ha-merged)

[Bello Abdullahi](https://savrn.com/model-publishers/therealbee)

This is the model card of a transformers model that has been pushed on the Hub. Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. Use the code below to get started with the model. Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).

Open weights 242M parameters transformers

[View model](https://savrn.com/models/whisper-small-ha-merged)

Model · Speech recognition

### [whisper-small-tibetan](https://savrn.com/models/whisper-small-tibetan)

[Ngodup](https://savrn.com/model-publishers/tenzin)

This model is a fine-tuned version of openai/whisper-small on the generator dataset. It achieves the following results on the evaluation set: The following hyperparameters were used during training: - learningrate: 1e-05 - trainbatchsize: 8 - evalbatchsize: 8 - gradientaccumulationsteps: 2 - totaltrainbatchsize: 16 - lrschedulertype: linear - lrschedulerwarmupsteps: 300 - numepochs: 3.0 - Transformers 4.57.6 - Pytorch 2.2.1+cu121 - Datasets 5.0.1 - Tokenizers 0.22.2

Open weights apache-2.0 242M parameters transformers

[View model](https://savrn.com/models/whisper-small-tibetan)

Model · Speech recognition

### [results](https://savrn.com/models/results)

[Sum](https://savrn.com/model-publishers/summi1125)

This model is a fine-tuned version of openai/whisper-small on an unknown dataset. The following hyperparameters were used during training: - learningrate: 1e-05 - trainbatchsize: 8 - evalbatchsize: 8 - lrschedulertype: linear - numepochs: 3 - 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 242M parameters transformers

[View model](https://savrn.com/models/results)

Model · Speech recognition

### [mms-300m-1130-forced-aligner](https://savrn.com/models/mms-300m-1130-forced-aligner)

[Mahmoud Ashraf](https://savrn.com/model-publishers/mahmoudashraf)

This Python package provides an efficient way to perform forced alignment between text and audio using Hugging Face's pretrained models. it also features an improved implementation to use much less memory than TorchAudio forced alignment API. The model checkpoint uploaded here is a conversion from torchaudio to HF Transformers for the MMS-300M checkpoint trained on forced alignment dataset

Open weights cc-by-nc-4.0 315M parameters transformers

[View model](https://savrn.com/models/mms-300m-1130-forced-aligner)

Model · Speech recognition

### [wav2vec2-xls-r-300m-hebrew](https://savrn.com/models/wav2vec2-xls-r-300m-hebrew)

[Vladimir Gurevich](https://savrn.com/model-publishers/imvladikon)

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the private datasets in 2 stages - firstly was fine-tuned on a small dataset with good samples Then the obtained model was fine-tuned on a large dataset with the small good dataset, with various samples from different sources, and with an unlabeled dataset that was weakly labeled using a previously trained model. (weakly labeled data wasn't used in validation set) on small dataset on large dataset on small dataset on large dataset The following hyperparameters were used during training: - learningrate: 0.0003 - trainbatchsize: 8 - evalbatchsize: 8 - distributedtype: multi-GPU - numdevices: 2 - gradientaccumulationsteps: 4…

Open weights 315M parameters transformers

[View model](https://savrn.com/models/wav2vec2-xls-r-300m-hebrew)

## Sajil C K

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

## Versions

- [dbd72416303a](https://savrn.com/models/whisper-small-malayalam/versions/dbd72416303a) · current 2026-10-09

## Explore More

- [All speech recognition models](https://savrn.com/models/tasks/speech-recognition)
- [All models under apache-2.0](https://savrn.com/models/licenses/apache-2-0)
- [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 2026-10-09.
- [Hugging Face record](https://huggingface.co/sajilck/whisper-small-malayalam)
- [How the hub is built](https://savrn.com/model-hub/methodology)