PyTorch
atomslm
language-model
shared-weights
AtomSLM-1.2M / config.json
Sqersters's picture
Upload AtomSLM checkpoint
e3d130b verified
Raw
History Blame Contribute Delete
232 Bytes
{
"vocab_size": 4096,
"d_model": 192,
"n_layers": 10,
"ffn_mult": 2.0,
"dropout": 0.1,
"max_seq_len": 256,
"tie_weights": true,
"architecture": "AtomSLM",
"val_loss": 3.1145258569717407,
"step": 5000
}