Hugging Face's logo

ft-lora
/
llama3.2-3b-instruct-finetuned

Text Generation
Transformers
Safetensors
English
llama
text-generation-inference
unsloth
conversational
8-bit precision
Model card Files
xet
 Community
Kajlid commited on
Commit
7b93237
·
verified ·
1 Parent(s): 1286a05

(Trained with Unsloth)

Browse files
Files changed (1) hide show
  1. config.json +36 -53
config.json CHANGED
@@ -1,55 +1,38 @@
1
  {
2
- "architectures": [
3
- "LlamaForCausalLM"
4
- ],
5
- "attention_bias": false,
6
- "attention_dropout": 0.0,
7
- "bos_token_id": 128000,
8
- "torch_dtype": "float16",
9
- "eos_token_id": 128009,
10
- "head_dim": 128,
11
- "hidden_act": "silu",
12
- "hidden_size": 3072,
13
- "initializer_range": 0.02,
14
- "intermediate_size": 8192,
15
- "max_position_embeddings": 131072,
16
- "mlp_bias": false,
17
- "model_type": "llama",
18
- "num_attention_heads": 24,
19
- "num_hidden_layers": 28,
20
- "num_key_value_heads": 8,
21
- "pad_token_id": 128004,
22
- "pretraining_tp": 1,
23
- "quantization_config": {
24
- "bnb_4bit_compute_dtype": "float16",
25
- "bnb_4bit_quant_type": "nf4",
26
- "bnb_4bit_use_double_quant": true,
27
- "llm_int8_enable_fp32_cpu_offload": false,
28
- "llm_int8_has_fp16_weight": false,
29
- "llm_int8_skip_modules": [
30
- "model.layers.1.mlp.gate_proj",
31
- "model.layers.1.mlp.up_proj",
32
- "model.layers.1.mlp.down_proj",
33
- "lm_head"
34
  ],
35
- "llm_int8_threshold": 6.0,
36
- "load_in_4bit": true,
37
- "load_in_8bit": false,
38
- "quant_method": "bitsandbytes"
39
- },
40
- "rms_norm_eps": 1e-05,
41
- "rope_scaling": {
42
- "factor": 32.0,
43
- "high_freq_factor": 4.0,
44
- "low_freq_factor": 1.0,
45
- "original_max_position_embeddings": 8192,
46
- "rope_type": "llama3"
47
- },
48
- "rope_theta": 500000.0,
49
- "tie_word_embeddings": true,
50
- "transformers_version": "4.57.2",
51
- "unsloth_fixed": true,
52
- "unsloth_version": "2025.11.4",
53
- "use_cache": true,
54
- "vocab_size": 128256
55
- }
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  {
2
+ "architectures": [
3
+ "LlamaForCausalLM"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4
  ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 128000,
8
+ "torch_dtype": "float16",
9
+ "eos_token_id": 128009,
10
+ "head_dim": 128,
11
+ "hidden_act": "silu",
12
+ "hidden_size": 3072,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 8192,
15
+ "max_position_embeddings": 131072,
16
+ "mlp_bias": false,
17
+ "model_type": "llama",
18
+ "num_attention_heads": 24,
19
+ "num_hidden_layers": 28,
20
+ "num_key_value_heads": 8,
21
+ "pad_token_id": 128004,
22
+ "pretraining_tp": 1,
23
+ "rms_norm_eps": 1e-05,
24
+ "rope_scaling": {
25
+ "factor": 32.0,
26
+ "high_freq_factor": 4.0,
27
+ "low_freq_factor": 1.0,
28
+ "original_max_position_embeddings": 8192,
29
+ "rope_type": "llama3"
30
+ },
31
+ "rope_theta": 500000.0,
32
+ "tie_word_embeddings": true,
33
+ "transformers_version": "4.57.2",
34
+ "unsloth_fixed": true,
35
+ "unsloth_version": "2025.11.4",
36
+ "use_cache": true,
37
+ "vocab_size": 128256
38
+ }