Safetensors
llama
falcon3
4-bit precision
awq
File size: 869 Bytes
fe9695a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
{
  "architectures": [
    "LlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "eos_token_id": 11,
  "head_dim": 256,
  "hidden_act": "silu",
  "hidden_size": 3072,
  "initializer_range": 0.02,
  "intermediate_size": 9216,
  "max_position_embeddings": 32768,
  "mlp_bias": false,
  "model_type": "llama",
  "num_attention_heads": 12,
  "num_hidden_layers": 22,
  "num_key_value_heads": 4,
  "output_hidden_states": true,
  "pretraining_tp": 1,
  "quantization_config": {
    "bits": 4,
    "group_size": 128,
    "modules_to_not_convert": null,
    "quant_method": "awq",
    "version": "gemm",
    "zero_point": true
  },
  "rms_norm_eps": 1e-06,
  "rope_scaling": null,
  "rope_theta": 1000042,
  "tie_word_embeddings": false,
  "torch_dtype": "float16",
  "transformers_version": "4.47.0",
  "use_cache": false,
  "vocab_size": 131072
}