Optimizing Pretraining Data Mixes with LLM-Estimated Utility
•
2
Will Held PRO
WillHeld
AI & ML interests
Machine Learning and Natural Language Processing for low-resource languages and language variants
Recent Activity
upvoted
an
article
1 day ago
Optimizing Pretraining Data Mixes with LLM-Estimated Utility
updated
a model
about 1 month ago
WillHeld/DiVA-llama-3-v0-8b
updated
a Space
about 1 month ago
WillHeld/diva-audio-chat
Articles
Organizations
models
12
WillHeld/DiVA-llama-3-v0-8b
Feature Extraction
•
Updated
•
533
•
30
WillHeld/debug_llama
Text Generation
•
Updated
•
191
WillHeld/DiVA-llama-3.2-1b
Updated
•
71
WillHeld/DiVA-llama-3.2-3B
Updated
•
25
WillHeld/DiVA-llama-3-distill-only-8b
Updated
•
5
WillHeld/DiVA-llama-3-token-align-8b
Updated
•
3
WillHeld/llama3-via-v0
Updated
•
16
WillHeld/via-llama
Updated
•
19
WillHeld/via-7b
Updated
•
3
WillHeld/levanter-whisper-tiny-librispeech
Automatic Speech Recognition
•
Updated
•
3
•
1
datasets
29
WillHeld/tst
Viewer
•
Updated
•
10
•
49
WillHeld/emotive_formatted_audio
Viewer
•
Updated
•
13.1k
•
168
•
1
WillHeld/CommonVoice_Shuffled_Processed
Viewer
•
Updated
•
1.12M
•
34
WillHeld/PeoplesSpeech_Shuffled_Processed
Viewer
•
Updated
•
1.52M
•
33
WillHeld/IntrinsicSaSEBenchmarks
Viewer
•
Updated
•
2.19k
•
50
•
1
WillHeld/MLS_Full_En
Viewer
•
Updated
•
10.8M
•
34
•
1
WillHeld/HeySQuAD_distill
Viewer
•
Updated
•
148k
•
470
•
1
WillHeld/librispeech_parquet
Viewer
•
Updated
•
568k
•
225
WillHeld/india_accent_cv
Viewer
•
Updated
•
91.5k
•
100
WillHeld/us_accent_cv
Viewer
•
Updated
•
49.1k
•
79