Hugging Face's logo

kilooli
/
EXAONE-3.5-2.4B-Instruct_gptq-4bit-wikitext2

Text Generation
Transformers
Safetensors
gemma3_text
conversational
text-generation-inference
4-bit precision
gptq
Model card Files
xet
Community
kilooli commited on
Commit
888a164
·
verified ·
1 Parent(s): 5cca65e

Upload Gemma3ForCausalLM

Browse files
Files changed (3) hide show
  1. config.json +53 -26
  2. generation_config.json +9 -3
  3. model.safetensors +2 -2
config.json CHANGED
@@ -1,27 +1,56 @@
1
  {
2
- "activation_function": "silu",
3
  "architectures": [
4
- "ExaoneForCausalLM"
5
  ],
 
6
  "attention_dropout": 0.0,
7
- "auto_map": {
8
- "AutoConfig": "configuration_exaone.ExaoneConfig",
9
- "AutoModelForCausalLM": "modeling_exaone.ExaoneForCausalLM",
10
- "AutoModelForSequenceClassification": "modeling_exaone.ExaoneForSequenceClassification"
11
- },
12
- "bos_token_id": 1,
13
- "embed_dropout": 0.0,
14
- "eos_token_id": 361,
15
- "head_dim": 80,
16
- "hidden_size": 2560,
 
17
  "initializer_range": 0.02,
18
- "intermediate_size": 7168,
19
- "layer_norm_epsilon": 1e-05,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
20
  "max_position_embeddings": 32768,
21
- "model_type": "exaone",
22
- "num_attention_heads": 32,
23
- "num_key_value_heads": 8,
24
- "num_layers": 30,
25
  "pad_token_id": 0,
26
  "quantization_config": {
27
  "backend": null,
@@ -55,16 +84,14 @@
55
  "use_cuda_fp16": false,
56
  "use_exllama": true
57
  },
58
- "rope_scaling": {
59
- "factor": 8.0,
60
- "high_freq_factor": 4.0,
61
- "low_freq_factor": 1.0,
62
- "original_max_position_embeddings": 8192,
63
- "rope_type": "llama3"
64
- },
65
  "rope_theta": 1000000,
 
66
  "torch_dtype": "float16",
67
  "transformers_version": "4.55.2",
68
  "use_cache": true,
69
- "vocab_size": 102400
70
  }
 
1
  {
2
+ "_sliding_window_pattern": 6,
3
  "architectures": [
4
+ "Gemma3ForCausalLM"
5
  ],
6
+ "attention_bias": false,
7
  "attention_dropout": 0.0,
8
+ "attn_logit_softcapping": null,
9
+ "bos_token_id": 2,
10
+ "cache_implementation": "hybrid",
11
+ "eos_token_id": [
12
+ 1,
13
+ 106
14
+ ],
15
+ "final_logit_softcapping": null,
16
+ "head_dim": 256,
17
+ "hidden_activation": "gelu_pytorch_tanh",
18
+ "hidden_size": 1152,
19
  "initializer_range": 0.02,
20
+ "intermediate_size": 6912,
21
+ "layer_types": [
22
+ "sliding_attention",
23
+ "sliding_attention",
24
+ "sliding_attention",
25
+ "sliding_attention",
26
+ "sliding_attention",
27
+ "full_attention",
28
+ "sliding_attention",
29
+ "sliding_attention",
30
+ "sliding_attention",
31
+ "sliding_attention",
32
+ "sliding_attention",
33
+ "full_attention",
34
+ "sliding_attention",
35
+ "sliding_attention",
36
+ "sliding_attention",
37
+ "sliding_attention",
38
+ "sliding_attention",
39
+ "full_attention",
40
+ "sliding_attention",
41
+ "sliding_attention",
42
+ "sliding_attention",
43
+ "sliding_attention",
44
+ "sliding_attention",
45
+ "full_attention",
46
+ "sliding_attention",
47
+ "sliding_attention"
48
+ ],
49
  "max_position_embeddings": 32768,
50
+ "model_type": "gemma3_text",
51
+ "num_attention_heads": 4,
52
+ "num_hidden_layers": 26,
53
+ "num_key_value_heads": 1,
54
  "pad_token_id": 0,
55
  "quantization_config": {
56
  "backend": null,
 
84
  "use_cuda_fp16": false,
85
  "use_exllama": true
86
  },
87
+ "query_pre_attn_scalar": 256,
88
+ "rms_norm_eps": 1e-06,
89
+ "rope_local_base_freq": 10000,
90
+ "rope_scaling": null,
 
 
 
91
  "rope_theta": 1000000,
92
+ "sliding_window": 512,
93
  "torch_dtype": "float16",
94
  "transformers_version": "4.55.2",
95
  "use_cache": true,
96
+ "vocab_size": 262144
97
  }
generation_config.json CHANGED
@@ -1,7 +1,13 @@
1
  {
2
- "_from_model_config": true,
3
- "bos_token_id": 1,
4
- "eos_token_id": 361,
 
 
 
 
5
  "pad_token_id": 0,
 
 
6
  "transformers_version": "4.55.2"
7
  }
 
1
  {
2
+ "bos_token_id": 2,
3
+ "cache_implementation": "hybrid",
4
+ "do_sample": true,
5
+ "eos_token_id": [
6
+ 1,
7
+ 106
8
+ ],
9
  "pad_token_id": 0,
10
+ "top_k": 64,
11
+ "top_p": 0.95,
12
  "transformers_version": "4.55.2"
13
  }
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e200e0a191dc47b1e5fcfd45ebabe4e7e36a3959d34426e14b2af84f999e4a86
3
- size 1640775056
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4c9f4a1eb9490b0afb755cfa3d96456278e5de7fc3d2e0848f49731a5ad1c441
3
+ size 968279216