Pretrain from scratch 4096 context length on 90B tokens Malaysian text, https://huggingface.co/papers/2401.14680

Mesolitica
company
AI & ML interests
We develop Multimodality Artificial Intelligence for South East Asia.
Recent Activity
View all activity
Collections
24
spaces
4
models
241

mesolitica/Malaysian-F5-TTS-v2
Updated
•
1

mesolitica/malay-parler-tts-mini-v1
Text2Text Generation
•
Updated
•
119
•
1

mesolitica/nanot5-small-malaysian-translation-v2
Translation
•
Updated
•
142
•
1

mesolitica/nanot5-base-malaysian-translation-v2
Translation
•
Updated
•
200
•
1

mesolitica/Malaysian-F5-TTS
Updated

mesolitica/Llama-3.2-1B-Malaysian-Reasoning
Updated
•
16

mesolitica/malaysian-Llama-3.2-1B-Instruct
Updated
•
61

mesolitica/Llama-3.2-3B-Malaysian-Reasoning
Updated
•
69

mesolitica/malaysian-Llama-3.2-3B-Instruct
Updated
•
67
•
2

mesolitica/malaysian-vocos-mel-24khz
Updated
•
19
datasets
206
mesolitica/Malaysian-STT-Whisper
Viewer
•
Updated
•
4.85M
•
1.6k
•
2
mesolitica/pseudolabel-malaysian-youtube-whisper-large-v3-timestamp
Preview
•
Updated
•
1.13k
mesolitica/TTS
Viewer
•
Updated
•
646k
•
2.59k
mesolitica/pseudolabel-science-large-v3-timestamp
Updated
•
210
mesolitica/Malaysian-Voice-Conversion
Viewer
•
Updated
•
3.74M
•
698
mesolitica/Malaysian-SFT
Preview
•
Updated
•
1.65k
•
1
mesolitica/Extra-Emilia
Updated
•
362
mesolitica/Malaysian-Emilia
Updated
•
1.1k
•
2
mesolitica/pseudolabel-tamil-large-v3-timestamp
Updated
•
351
mesolitica/pseudolabel-mandarin-large-v3-timestamp
Updated
•
294