hard_3gram_pile_2layer / config.json
Lanni-ni's picture
Upload SlidingWindowForCausalLM
dd1f0c9 verified
{
"architectures": [
"SlidingWindowForCausalLM"
],
"attention_bias": false,
"auto_map": {
"AutoConfig": "configuration_sliding_window.SlidingWindowConfig",
"AutoModelForCausalLM": "modeling_sliding_window.SlidingWindowForCausalLM"
},
"bos_token_id": null,
"dtype": "float32",
"elementwise_affine": true,
"eos_token_id": null,
"fuse_cross_entropy": true,
"fuse_norm": true,
"hidden_act": "swish",
"hidden_ratio": 4,
"hidden_size": 256,
"initializer_range": 0.02,
"intermediate_size": 1024,
"max_position_embeddings": null,
"model_type": "sliding_window",
"norm_eps": 1e-06,
"num_heads": 4,
"num_hidden_layers": 2,
"num_kv_heads": null,
"qk_norm": false,
"qk_norm_share_param_across_head": false,
"transformers_version": "4.57.1",
"use_cache": true,
"use_k_shift": false,
"use_rope": false,
"use_v_shift": false,
"vocab_size": 50277,
"window_size": 2
}