Text Generation
Transformers
Safetensors
English
llama
text-generation-inference
4-bit precision
awq
File size: 729 Bytes
f40a50a
66344c9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f40a50a
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
{
  "_name_or_path": "/workspace/process/tulu-30b/source",
  "architectures": [
    "LlamaForCausalLM"
  ],
  "bos_token_id": 1,
  "eos_token_id": 2,
  "hidden_act": "silu",
  "hidden_size": 6656,
  "initializer_range": 0.02,
  "intermediate_size": 17920,
  "max_position_embeddings": 2048,
  "model_type": "llama",
  "num_attention_heads": 52,
  "num_hidden_layers": 60,
  "pad_token_id": 0,
  "rms_norm_eps": 1e-06,
  "tie_word_embeddings": false,
  "torch_dtype": "float16",
  "transformers_version": "4.30.0.dev0",
  "use_cache": true,
  "vocab_size": 32001,
  "pretraining_tp": 1,
  "quantization_config": {
    "quant_method": "awq",
    "zero_point": true,
    "group_size": 128,
    "bits": 4,
    "version": "gemm"
  }
}