Downloads · 30 days
22
18% of all-time downloads
KasuleTrevor/cdli-qwen3-asr-lg-atypical-stage3-1p7b-typical-base-spec-cosine
cdli-qwen3-asr-lg-atypical-stage3-1p7b-typical-base-spec-cosine is a automatic speech recognition model from KasuleTrevor. Use it when you need speech turned into text.
This repository contains a Qwen3-ASR checkpoint fine-tuned for Luganda atypical speech from the CDLI Luganda non-standard speech dataset, using a SpecAugment-enabled cosine-scheduler run.
Downloads · 30 days
22
18% of all-time downloads
All-time downloads
122
Public
Parameters
2B
4.1 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors4.1 GB · 100%
From the Hugging Face model README
This repository contains a Qwen3-ASR checkpoint fine-tuned for Luganda atypical speech from the CDLI Luganda non-standard speech dataset, using a SpecAugment-enabled cosine-scheduler run.
KasuleTrevor/cdli-qwen3-asr-lg-typical-1p7b-base-finetunecdli/ugandan_luganda_nonstandard_speech_v1.0lg)qwen3-asr-luganda-atypical-cleaned-finetune-spec-copy.ipynb5e-5cosine584250Notebook outputs show that validation checkpoint comparison selected
checkpoint-500 by normalized corpus WER.
| Checkpoint | Validation WER | Validation CER | Eval loss |
|---|---|---|---|
checkpoint-500 | 0.635647 | 0.321327 | 0.515153 |
| Checkpoint | Test WER | Test CER | Avg WER capped | Avg CER capped |
|---|---|---|---|---|
checkpoint-500 | 0.515667 | 0.235181 | 0.438203 | 0.188558 |
These per-speaker aggregates come from the notebook-displayed grouped analysis
for checkpoint-500.
| Speaker | Severity | Disorder | Etiology | N | Mean WER | Mean CER | Median WER | Median CER |
|---|---|---|---|---|---|---|---|---|
UG057 | Severe (frequent breakdowns) | Voice disorder | Acquired hearing impairment | 105 | 0.6408 | 0.3398 | 0.7297 | 0.3731 |
UG042 | Severe (frequent breakdowns) | Stuttering (Disfluency Disorders) | Developmental | 85 | 0.5421 | 0.2372 | 0.5000 | 0.1667 |
UG052 | Moderate (requires effort to understand) | Stuttering (Disfluency Disorders) | Developmental | 79 | 0.4995 | 0.2304 | 0.4286 | 0.1571 |
UG058 | Severe (frequent breakdowns) | Stuttering (Disfluency Disorders) | Developmental | 125 | 0.4928 | 0.2662 | 0.5000 | 0.2622 |
UG021 | Moderate (requires effort to understand) | Articulation Disorders | Structural presence of akloglosia, simply tong... | 91 | 0.4294 | 0.1427 | 0.4242 | 0.1292 |
UG022 | Mild (easily understood with minimal effort) | Articulation Disorders | Developmental | 117 | 0.4102 | 0.1581 | 0.3750 | 0.0972 |
UG036 | Moderate (requires effort to understand) | Dysarthria | Cerebral Palsy | 177 | 0.3925 | 0.1442 | 0.3636 | 0.1087 |
UG001 | Mild (easily understood with minimal effort) | Dysarthria | Cerebral palsy - cerebral malaria | 99 | 0.3404 | 0.1379 | 0.3333 | 0.1111 |
UG014 | Mild (easily understood with minimal effort) | Stuttering (Disfluency Disorders) | Idiopathic | 149 | 0.3046 | 0.1052 | 0.2500 | 0.0492 |
The model card is reconstructed from executed notebook outputs and extracted
local summaries. Full per-utterance prediction files were referenced in notebook
outputs as remote /workspace/... artifact paths, but were not embedded in the
notebook JSON available locally.