Models
Datasets
Spaces
Buckets new
Docs
Enterprise
Pricing
- Website
- Community
- Solutions
Log In
Sign Up

Spaces:

Nrighton233j
/

gpt-joy

Sleeping

App Files Files Community

Fetching metadata from the HF Docker repository...

18.9 kB

Ctrl+K

Ctrl+K

1 contributor

History: 9 commits

Nrighton233j

Add tokenizer download route

f19ac4e 2 months ago

Dockerfile

197 Bytes
GPT-Joy: from-scratch GPT training space 3 months ago
README.md

478 Bytes
GPT-Joy: from-scratch GPT training space 3 months ago
app.py

1.39 kB
Add tokenizer download route 2 months ago
data_prep.py

9.91 kB
Fix OOM: stream large datasets with row caps, write corpus incrementally instead of buffering in RAM 3 months ago
model.py

3.44 kB
GPT-Joy: from-scratch GPT training space 3 months ago
requirements.txt

123 Bytes
Push checkpoints to Hub model repo every save 3 months ago
train.py

3.36 kB
Bump MAX_STEPS to 40000 for ~4.5 epoch coverage on larger corpus 3 months ago