Downloads · 30 days
20
11% of all-time downloads
Mieaz/gpt15m-chat
gpt15m-chat is a text generation model from Mieaz. Use it when you need the model to write or continue text. The card lists the license as mit.
12.4M-param GPT-2-style model (6 layers, 384 hidden, 6 heads, tied embeddings, vocab 4095). Trained for 1 epoch on 300M tokens: 200M TinyStories + 100M ultrachat dialogues. Final held-out validation loss: 1.75 (TinySt…
Downloads · 30 days
20
11% of all-time downloads
All-time downloads
187
Public
Repo size
333 MB
Likes
0
Public
Click a slice to open those files.
.pt56 MB · 52%
From the Hugging Face model README
12.4M-param GPT-2-style model (6 layers, 384 hidden, 6 heads, tied embeddings, vocab 4095). Trained for 1 epoch on 300M tokens: 200M TinyStories + 100M ultrachat dialogues. Final held-out validation loss: 1.75 (TinyStories-valid).
Use with Transformers.js in the browser (ONNX) or PyTorch (model.pt).