Hugging Face's logo

TheBloke
/
Hermes-LLongMA-2-13B-8K-AWQ

Text Generation
Transformers
Safetensors
llama
custom_code
text-generation-inference
4-bit precision
awq
Model card Files Community
1
TheBloke commited on
Commit
c290814
•
1 Parent(s): 13e384b

Update for Transformers AWQ support

Browse files
Files changed (1) hide show
  1. config.json +45 -38
config.json CHANGED
@@ -1,40 +1,47 @@
1
  {
2
- "_name_or_path": "conceptofmind/LLongMA-2-13b",
3
- "architectures": [
4
- "LlamaForCausalLM"
5
- ],
6
- "auto_map": {
7
- "AutoConfig": "conceptofmind/LLongMA-2-13b--configuration_llama.LlamaConfig",
8
- "AutoModel": "conceptofmind/LLongMA-2-13b--modeling_llama.LlamaModel",
9
- "AutoModelForCausalLM": "conceptofmind/LLongMA-2-13b--modeling_llama.LlamaForCausalLM",
10
- "AutoModelForSequenceClassification": "conceptofmind/LLongMA-2-13b--modeling_llama.LlamaForSequenceClassification"
11
- },
12
- "bos_token_id": 1,
13
- "eos_token_id": 2,
14
- "hidden_act": "silu",
15
- "hidden_size": 5120,
16
- "initializer_range": 0.02,
17
- "intermediate_size": 13824,
18
- "max_position_embeddings": 8192,
19
- "model_type": "llama",
20
- "ntk_alpha": null,
21
- "num_attention_heads": 40,
22
- "num_hidden_layers": 40,
23
- "num_key_value_heads": 40,
24
- "pad_token_id": 0,
25
- "part_ntk_scale": null,
26
- "position_interpolation_scale": 1,
27
- "pretraining_tp": 1,
28
- "rms_norm_eps": 1e-05,
29
- "rope_scaling": {
30
- "factor": 2.0,
31
- "type": "linear"
32
- },
33
- "tie_word_embeddings": false,
34
- "torch_dtype": "bfloat16",
35
- "transformer_engine": null,
36
- "transformers_version": "4.32.0.dev0",
37
- "use_cache": true,
38
- "use_xpos": false,
39
- "vocab_size": 32000
 
 
 
 
 
 
 
40
  }
 
1
  {
2
+ "_name_or_path": "conceptofmind/LLongMA-2-13b",
3
+ "architectures": [
4
+ "LlamaForCausalLM"
5
+ ],
6
+ "auto_map": {
7
+ "AutoConfig": "conceptofmind/LLongMA-2-13b--configuration_llama.LlamaConfig",
8
+ "AutoModel": "conceptofmind/LLongMA-2-13b--modeling_llama.LlamaModel",
9
+ "AutoModelForCausalLM": "conceptofmind/LLongMA-2-13b--modeling_llama.LlamaForCausalLM",
10
+ "AutoModelForSequenceClassification": "conceptofmind/LLongMA-2-13b--modeling_llama.LlamaForSequenceClassification"
11
+ },
12
+ "bos_token_id": 1,
13
+ "eos_token_id": 2,
14
+ "hidden_act": "silu",
15
+ "hidden_size": 5120,
16
+ "initializer_range": 0.02,
17
+ "intermediate_size": 13824,
18
+ "max_position_embeddings": 8192,
19
+ "model_type": "llama",
20
+ "ntk_alpha": null,
21
+ "num_attention_heads": 40,
22
+ "num_hidden_layers": 40,
23
+ "num_key_value_heads": 40,
24
+ "pad_token_id": 0,
25
+ "part_ntk_scale": null,
26
+ "position_interpolation_scale": 1,
27
+ "pretraining_tp": 1,
28
+ "rms_norm_eps": 1e-05,
29
+ "rope_scaling": {
30
+ "factor": 2.0,
31
+ "type": "linear"
32
+ },
33
+ "tie_word_embeddings": false,
34
+ "torch_dtype": "bfloat16",
35
+ "transformer_engine": null,
36
+ "transformers_version": "4.32.0.dev0",
37
+ "use_cache": true,
38
+ "use_xpos": false,
39
+ "vocab_size": 32000,
40
+ "quantization_config": {
41
+ "quant_method": "awq",
42
+ "zero_point": true,
43
+ "group_size": 128,
44
+ "bits": 4,
45
+ "version": "gemm"
46
+ }
47
  }