alessiodevoto's picture
Push model using huggingface_hub.
1cfae81 verified
{
"dataset_name": "kmfoda/booksum",
"head_dim": 128,
"model_name": "Qwen/Qwen3-4B-Instruct-2507",
"n_sink": 4,
"num_heads": 32,
"num_layers": 36,
"num_samples": 100,
"sample_seq_len": 1000
}