config.json
691 B · 31 lines · json Raw
1 {
2 "architectures": [
3 "LlamaForCausalLM"
4 ],
5 "attention_bias": false,
6 "attention_dropout": 0.0,
7 "bos_token_id": 0,
8 "eos_token_id": 1,
9 "head_dim": 64,
10 "hidden_act": "silu",
11 "hidden_size": 16,
12 "initializer_range": 0.02,
13 "intermediate_size": 64,
14 "max_position_embeddings": 8192,
15 "mlp_bias": false,
16 "model_type": "llama",
17 "num_attention_heads": 4,
18 "num_hidden_layers": 2,
19 "num_key_value_heads": 4,
20 "pad_token_id": -1,
21 "pretraining_tp": 1,
22 "rms_norm_eps": 1e-06,
23 "rope_scaling": null,
24 "rope_theta": 10000.0,
25 "tie_word_embeddings": false,
26 "torch_dtype": "bfloat16",
27 "transformers_version": "4.51.3",
28 "use_cache": true,
29 "vocab_size": 32000
30 }
31