hoffmannlu's picture
Add predict.py, README.md, config.json, training_args.json, model.safetensors
6125e45 verified
Raw History Blame Contribute Delete
452 Bytes
{
"architectures": [
"CustomResearchModel"
],
"architecture": "tiny_transformer",
"model_type": "tiny_transformer",
"hidden_size": 384,
"num_hidden_layers": 6,
"num_attention_heads": 8,
"intermediate_size": 1536,
"hidden_act": "swish",
"max_position_embeddings": 256,
"layer_norm_eps": 1e-12,
"checkpoint_status": "initialization-only",
"notes": "Untrained checkpoint for smoke tests; no benchmark claim."
}