Rajit-Anand06 commited on
Commit
87578f3
·
verified ·
1 Parent(s): bee9854

upload config.json

Browse files
Files changed (1) hide show
  1. config.json +77 -0
config.json ADDED
@@ -0,0 +1,77 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "facebook/blenderbot-3B",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "add_bias_logits": false,
6
+ "add_final_layer_norm": true,
7
+ "architectures": [
8
+ "BlenderbotForConditionalGeneration"
9
+ ],
10
+ "attention_dropout": 0.0,
11
+ "bos_token_id": 1,
12
+ "classif_dropout": 0.0,
13
+ "d_model": 2560,
14
+ "decoder_attention_heads": 32,
15
+ "decoder_ffn_dim": 10240,
16
+ "decoder_layerdrop": 0.0,
17
+ "decoder_layers": 24,
18
+ "decoder_start_token_id": 1,
19
+ "dropout": 0.1,
20
+ "encoder_attention_heads": 32,
21
+ "encoder_ffn_dim": 10240,
22
+ "encoder_layerdrop": 0.0,
23
+ "encoder_layers": 2,
24
+ "encoder_no_repeat_ngram_size": 3,
25
+ "eos_token_id": 2,
26
+ "extra_layer_norm": false,
27
+ "extra_pos_embeddings": 0,
28
+ "force_bos_token_to_be_generated": false,
29
+ "forced_eos_token_id": 2,
30
+ "id2label": {
31
+ "0": "LABEL_0",
32
+ "1": "LABEL_1",
33
+ "2": "LABEL_2"
34
+ },
35
+ "init_std": 0.02,
36
+ "is_encoder_decoder": true,
37
+ "label2id": {
38
+ "LABEL_0": 0,
39
+ "LABEL_1": 1,
40
+ "LABEL_2": 2
41
+ },
42
+ "layernorm_variant": "prelayernorm",
43
+ "length_penalty": 0.65,
44
+ "max_length": 60,
45
+ "max_position_embeddings": 128,
46
+ "min_length": 20,
47
+ "model_type": "blenderbot",
48
+ "no_repeat_ngram_size": 3,
49
+ "normalize_before": true,
50
+ "normalize_embedding": false,
51
+ "num_beams": 10,
52
+ "num_hidden_layers": 2,
53
+ "pad_token_id": 0,
54
+ "pipeline_tag": "text2text-generation",
55
+ "quantization_config": {
56
+ "_load_in_4bit": true,
57
+ "_load_in_8bit": false,
58
+ "bnb_4bit_compute_dtype": "bfloat16",
59
+ "bnb_4bit_quant_storage": "uint8",
60
+ "bnb_4bit_quant_type": "nf4",
61
+ "bnb_4bit_use_double_quant": true,
62
+ "llm_int8_enable_fp32_cpu_offload": false,
63
+ "llm_int8_has_fp16_weight": false,
64
+ "llm_int8_skip_modules": null,
65
+ "llm_int8_threshold": 6.0,
66
+ "load_in_4bit": true,
67
+ "load_in_8bit": false,
68
+ "quant_method": "bitsandbytes"
69
+ },
70
+ "scale_embedding": true,
71
+ "static_position_embeddings": false,
72
+ "torch_dtype": "bfloat16",
73
+ "transformers_version": "4.49.0",
74
+ "unk_token_id": 3,
75
+ "use_cache": true,
76
+ "vocab_size": 8008
77
+ }