Hugging Face
Models
Datasets
Spaces
Buckets
Docs
Enterprise
Pricing
Website
Tasks
HuggingChat
Collections
Languages
Organizations
Community
Blog
Posts
Daily Papers
Hardware
Learn
Discord
Forum
GitHub
Solutions
Team & Enterprise
Hugging Face PRO
Enterprise Support
Inference Providers
Inference Endpoints
Storage Buckets
Log In
Sign Up
prometheus04
/
matilda-mini
Like
0
matilda
custom_code
Model card
Files
Files and versions
xet
Community
Copy to bucket
main
matilda-mini
/
src
/
matilda
45.1 kB
Ctrl+K
Ctrl+K
2 contributors
History:
6 commits
prometheus04
Fix RoPE dtype cast for bfloat16 inference
9b5174f
verified
5 months ago
__init__.py
Safe
151 Bytes
Matilda-Mini phases 1-5 + runbook
5 months ago
checkpoint.py
Safe
4.05 kB
GPU-session fixes (RNG cpu, shard filter, cu124, 3090 config)
5 months ago
config.py
Safe
2.35 kB
Matilda-Mini phases 1-5 + runbook
5 months ago
data.py
Safe
8.05 kB
second review fixes
5 months ago
model.py
Safe
8.24 kB
Fix RoPE dtype cast for bfloat16 inference
5 months ago
monitor.py
Safe
2.91 kB
Matilda-Mini phases 1-5 + runbook
5 months ago
optim.py
Safe
7.44 kB
second review fixes
5 months ago
train.py
Safe
11.9 kB
Add trained checkpoint: 3B tokens, loss=3.16, MFU=31.5%
5 months ago
