microgpt-deva / config.json
ss-76's picture
Revised upload of MicroGPT-Deva model with 3 epoch training
1668709 verified
{"batch_size": 32, "block_size": 512, "dropout": 0.0, "lr": 0.0003, "n_embd": 512, "n_head": 8, "n_layer": 8, "num_epochs": 3, "resume_path": "model.pth", "vocab_size": 12000}