Hugging Face's logo

RichardErkhov
/
jpacifico_-_Chocolatine-3B-Instruct-DPO-v1.2-4bits

Safetensors
phi3
custom_code
4-bit precision
bitsandbytes
Model card Files Community
RichardErkhov commited on
Commit
69fab34
•
1 Parent(s): dc44273

uploaded model

Browse files
Files changed (1) hide show
  1. config.json +153 -0
config.json ADDED
@@ -0,0 +1,153 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "Chocolatine-3B-Instruct-DPO-v1.2",
3
+ "architectures": [
4
+ "Phi3ForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "auto_map": {
9
+ "AutoConfig": "microsoft/Phi-3.5-mini-instruct--configuration_phi3.Phi3Config",
10
+ "AutoModelForCausalLM": "microsoft/Phi-3.5-mini-instruct--modeling_phi3.Phi3ForCausalLM"
11
+ },
12
+ "bos_token_id": 1,
13
+ "embd_pdrop": 0.0,
14
+ "eos_token_id": 32000,
15
+ "hidden_act": "silu",
16
+ "hidden_size": 3072,
17
+ "initializer_range": 0.02,
18
+ "intermediate_size": 8192,
19
+ "max_position_embeddings": 131072,
20
+ "model_type": "phi3",
21
+ "num_attention_heads": 32,
22
+ "num_hidden_layers": 32,
23
+ "num_key_value_heads": 32,
24
+ "original_max_position_embeddings": 4096,
25
+ "pad_token_id": 32000,
26
+ "quantization_config": {
27
+ "_load_in_4bit": true,
28
+ "_load_in_8bit": false,
29
+ "bnb_4bit_compute_dtype": "float32",
30
+ "bnb_4bit_quant_storage": "uint8",
31
+ "bnb_4bit_quant_type": "fp4",
32
+ "bnb_4bit_use_double_quant": false,
33
+ "llm_int8_enable_fp32_cpu_offload": false,
34
+ "llm_int8_has_fp16_weight": false,
35
+ "llm_int8_skip_modules": null,
36
+ "llm_int8_threshold": 6.0,
37
+ "load_in_4bit": true,
38
+ "load_in_8bit": false,
39
+ "quant_method": "bitsandbytes"
40
+ },
41
+ "resid_pdrop": 0.0,
42
+ "rms_norm_eps": 1e-05,
43
+ "rope_scaling": {
44
+ "long_factor": [
45
+ 1.0800000429153442,
46
+ 1.1100000143051147,
47
+ 1.1399999856948853,
48
+ 1.340000033378601,
49
+ 1.5899999141693115,
50
+ 1.600000023841858,
51
+ 1.6200000047683716,
52
+ 2.620000123977661,
53
+ 3.2300000190734863,
54
+ 3.2300000190734863,
55
+ 4.789999961853027,
56
+ 7.400000095367432,
57
+ 7.700000286102295,
58
+ 9.09000015258789,
59
+ 12.199999809265137,
60
+ 17.670000076293945,
61
+ 24.46000099182129,
62
+ 28.57000160217285,
63
+ 30.420001983642578,
64
+ 30.840002059936523,
65
+ 32.590003967285156,
66
+ 32.93000411987305,
67
+ 42.320003509521484,
68
+ 44.96000289916992,
69
+ 50.340003967285156,
70
+ 50.45000457763672,
71
+ 57.55000305175781,
72
+ 57.93000411987305,
73
+ 58.21000289916992,
74
+ 60.1400032043457,
75
+ 62.61000442504883,
76
+ 62.62000274658203,
77
+ 62.71000289916992,
78
+ 63.1400032043457,
79
+ 63.1400032043457,
80
+ 63.77000427246094,
81
+ 63.93000411987305,
82
+ 63.96000289916992,
83
+ 63.970001220703125,
84
+ 64.02999877929688,
85
+ 64.06999969482422,
86
+ 64.08000183105469,
87
+ 64.12000274658203,
88
+ 64.41000366210938,
89
+ 64.4800033569336,
90
+ 64.51000213623047,
91
+ 64.52999877929688,
92
+ 64.83999633789062
93
+ ],
94
+ "short_factor": [
95
+ 1.0,
96
+ 1.0199999809265137,
97
+ 1.0299999713897705,
98
+ 1.0299999713897705,
99
+ 1.0499999523162842,
100
+ 1.0499999523162842,
101
+ 1.0499999523162842,
102
+ 1.0499999523162842,
103
+ 1.0499999523162842,
104
+ 1.0699999332427979,
105
+ 1.0999999046325684,
106
+ 1.1099998950958252,
107
+ 1.1599998474121094,
108
+ 1.1599998474121094,
109
+ 1.1699998378753662,
110
+ 1.2899998426437378,
111
+ 1.339999794960022,
112
+ 1.679999828338623,
113
+ 1.7899998426437378,
114
+ 1.8199998140335083,
115
+ 1.8499997854232788,
116
+ 1.8799997568130493,
117
+ 1.9099997282028198,
118
+ 1.9399996995925903,
119
+ 1.9899996519088745,
120
+ 2.0199997425079346,
121
+ 2.0199997425079346,
122
+ 2.0199997425079346,
123
+ 2.0199997425079346,
124
+ 2.0199997425079346,
125
+ 2.0199997425079346,
126
+ 2.0299997329711914,
127
+ 2.0299997329711914,
128
+ 2.0299997329711914,
129
+ 2.0299997329711914,
130
+ 2.0299997329711914,
131
+ 2.0299997329711914,
132
+ 2.0299997329711914,
133
+ 2.0299997329711914,
134
+ 2.0299997329711914,
135
+ 2.0799996852874756,
136
+ 2.0899996757507324,
137
+ 2.189999580383301,
138
+ 2.2199995517730713,
139
+ 2.5899994373321533,
140
+ 2.729999542236328,
141
+ 2.749999523162842,
142
+ 2.8399994373321533
143
+ ],
144
+ "type": "longrope"
145
+ },
146
+ "rope_theta": 10000.0,
147
+ "sliding_window": 262144,
148
+ "tie_word_embeddings": false,
149
+ "torch_dtype": "float16",
150
+ "transformers_version": "4.44.2",
151
+ "use_cache": true,
152
+ "vocab_size": 32064
153
+ }