Hugging Face's logo

Kukedlc
/
fusion_model_2

Text Generation
Transformers
Safetensors
gemma2
conversational
text-generation-inference
Inference Endpoints
4-bit precision
bitsandbytes
Model card Files Community
Kukedlc commited on
Commit
dea862b
•
1 Parent(s): 10bf5cd

Update config.json

Browse files
Files changed (1) hide show
  1. config.json +51 -48
config.json CHANGED
@@ -1,48 +1,51 @@
1
- {
2
- "_name_or_path": "fusion_model_2",
3
- "architectures": [
4
- "Gemma2ForCausalLM"
5
- ],
6
- "attention_bias": false,
7
- "attention_dropout": 0.0,
8
- "attn_logit_softcapping": 50.0,
9
- "bos_token_id": 2,
10
- "cache_implementation": "hybrid",
11
- "eos_token_id": 1,
12
- "final_logit_softcapping": 30.0,
13
- "head_dim": 256,
14
- "hidden_act": "gelu_pytorch_tanh",
15
- "hidden_activation": "gelu_pytorch_tanh",
16
- "hidden_size": 2304,
17
- "initializer_range": 0.02,
18
- "intermediate_size": 9216,
19
- "max_position_embeddings": 8192,
20
- "model_type": "gemma2",
21
- "num_attention_heads": 8,
22
- "num_hidden_layers": 26,
23
- "num_key_value_heads": 4,
24
- "pad_token_id": 0,
25
- "quantization_config": {
26
- "_load_in_4bit": true,
27
- "_load_in_8bit": false,
28
- "bnb_4bit_compute_dtype": "float32",
29
- "bnb_4bit_quant_storage": "uint8",
30
- "bnb_4bit_quant_type": "fp4",
31
- "bnb_4bit_use_double_quant": false,
32
- "llm_int8_enable_fp32_cpu_offload": false,
33
- "llm_int8_has_fp16_weight": false,
34
- "llm_int8_skip_modules": null,
35
- "llm_int8_threshold": 6.0,
36
- "load_in_4bit": true,
37
- "load_in_8bit": false,
38
- "quant_method": "bitsandbytes"
39
- },
40
- "query_pre_attn_scalar": 256,
41
- "rms_norm_eps": 1e-06,
42
- "rope_theta": 10000.0,
43
- "sliding_window": 4096,
44
- "torch_dtype": "bfloat16",
45
- "transformers_version": "4.45.0.dev0",
46
- "use_cache": true,
47
- "vocab_size": 256000
48
- }
 
 
 
 
1
+ {
2
+ "_name_or_path": "fusion_model_2",
3
+ "architectures": [
4
+ "Gemma2ForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "attn_logit_softcapping": 50.0,
9
+ "bos_token_id": 2,
10
+ "cache_implementation": "hybrid",
11
+ "eos_token_id": [
12
+ 1,
13
+ 107
14
+ ],
15
+ "final_logit_softcapping": 30.0,
16
+ "head_dim": 256,
17
+ "hidden_act": "gelu_pytorch_tanh",
18
+ "hidden_activation": "gelu_pytorch_tanh",
19
+ "hidden_size": 2304,
20
+ "initializer_range": 0.02,
21
+ "intermediate_size": 9216,
22
+ "max_position_embeddings": 8192,
23
+ "model_type": "gemma2",
24
+ "num_attention_heads": 8,
25
+ "num_hidden_layers": 26,
26
+ "num_key_value_heads": 4,
27
+ "pad_token_id": 0,
28
+ "quantization_config": {
29
+ "_load_in_4bit": true,
30
+ "_load_in_8bit": false,
31
+ "bnb_4bit_compute_dtype": "float32",
32
+ "bnb_4bit_quant_storage": "uint8",
33
+ "bnb_4bit_quant_type": "fp4",
34
+ "bnb_4bit_use_double_quant": false,
35
+ "llm_int8_enable_fp32_cpu_offload": false,
36
+ "llm_int8_has_fp16_weight": false,
37
+ "llm_int8_skip_modules": null,
38
+ "llm_int8_threshold": 6.0,
39
+ "load_in_4bit": true,
40
+ "load_in_8bit": false,
41
+ "quant_method": "bitsandbytes"
42
+ },
43
+ "query_pre_attn_scalar": 256,
44
+ "rms_norm_eps": 1e-06,
45
+ "rope_theta": 10000.0,
46
+ "sliding_window": 4096,
47
+ "torch_dtype": "bfloat16",
48
+ "transformers_version": "4.45.0.dev0",
49
+ "use_cache": true,
50
+ "vocab_size": 256000
51
+ }