🔄 In a Training Loop
foss
foss22
AI & ML interests
None yet
Recent Activity
updated a model 5 days ago
foss22/NanoZaya-3B-A35M published a model 6 days ago
foss22/NanoZaya-3B-A35M liked a model 8 days ago
Zyphra/ZUNA1.1Organizations
Agentic 0,07 B LM tasks + complete list of use-cases for Vintage LMs
3
#3 opened 11 days ago
by
foss22
77 million parameters is equal to 0.077 billion, not 0.007
❤️ 1
#2 opened 11 days ago
by
foss22
Per-profession vocab
👍 1
6
#4 opened 20 days ago
by
foss22
RedBook tokenizier. Biodiversity-consetvation inspired
#6 opened 11 days ago
by
foss22
Tokens indicate unbalanced subsets, that do not represent general English
#8 opened 15 days ago
by
foss22
Å|į|Ã|§|©
1
#7 opened 15 days ago
by
foss22
Remove ---- .... === ### dupes from datasets and tokenizer
👍 1
4
#3 opened 19 days ago
by
foss22
Capitalization not normalized
#6 opened 15 days ago
by
foss22
When capital letter matters?
#5 opened 15 days ago
by
foss22
Page numbers dominate 645 num tokens
#4 opened 15 days ago
by
foss22
Collect test inference or no value of collecting?
6
#1 opened 17 days ago
by
foss22
29 of 798 Wikipedia 1900-1990s neologisms
5
#1 opened 27 days ago
by
foss22
Questions in 454k vs 783k without questions
1
#5 opened 19 days ago
by
foss22
Redundancy
1
#4 opened 20 days ago
by
foss22
Under-filtering.
1
#3 opened 20 days ago
by
foss22
Over-filtering problem with current banned.txt
1
#2 opened 20 days ago
by
foss22
F16 NaN in 10 steps workaround
1
#2 opened 20 days ago
by
foss22
Extrapolation of 760 achievable steps/hour findins to 75k-steps based on what?
2
#1 opened 20 days ago
by
foss22
Collocations (easy, the same method) + extra mile: Idioms (hard, invent methods)
4
#3 opened 25 days ago
by
foss22