KasuleTrevor
/

cdli-qwen3-asr-en-finetune-salt-15

Automatic Speech Recognition

text-generation

atypical-speech

Model card Files Files and versions

CDLI Qwen3-ASR English Atypical Speech Fine-tune

This repo contains the latest training checkpoint only: checkpoint-13860.

Training Setup

Base model: Qwen/Qwen3-ASR-1.7B
Dataset: cdli/ugandan_english_nonstandard_speech_v1.0
Task: Automatic Speech Recognition
Language: English

Notes

This upload excludes training-state files to keep the repo clean.
The checkpoint folder already includes tokenizer/processor files needed for inference.
WER/CER are computed and uploaded later by the evaluation/results cells.
Uploaded: 2026-04-13T20:05:41.801455+00:00

Downloads last month: 2

Safetensors

Model size

2B params

Tensor type

BF16

·

Model tree for KasuleTrevor/cdli-qwen3-asr-en-finetune-salt-15

Base model

Qwen/Qwen3-ASR-1.7B

Finetuned

(68)

this model

Dataset used to train KasuleTrevor/cdli-qwen3-asr-en-finetune-salt-15

Collection including KasuleTrevor/cdli-qwen3-asr-en-finetune-salt-15

CDLI

This is a collection of models used for the CDLI ASR challenge for atypical speech in Uganda on Ugandan English and Luganda. • 26 items • Updated May 16