Hugging Face's logo

microsoft
/
MediPhi

Text Generation
Transformers
Safetensors
English
phi3
Merge
mergekit
medical
clinical
conversational
text-generation-inference
Model card Files Community
jpcorb20 commited on
Commit
98d18b3
·
verified ·
1 Parent(s): 493ec0b

Update config.json

Browse files
Files changed (1) hide show
  1. config.json +133 -138
config.json CHANGED
@@ -1,138 +1,133 @@
1
- {
2
- "_name_or_path": "./Phi-3.5-mini-instruct/",
3
- "architectures": [
4
- "Phi3ForCausalLM"
5
- ],
6
- "attention_bias": false,
7
- "attention_dropout": 0.0,
8
- "auto_map": {
9
- "AutoConfig": "configuration_phi3.Phi3Config",
10
- "AutoModelForCausalLM": "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
- "resid_pdrop": 0.0,
27
- "rms_norm_eps": 1e-05,
28
- "rope_scaling": {
29
- "long_factor": [
30
- 1.0800000429153442,
31
- 1.1100000143051147,
32
- 1.1399999856948853,
33
- 1.340000033378601,
34
- 1.5899999141693115,
35
- 1.600000023841858,
36
- 1.6200000047683716,
37
- 2.620000123977661,
38
- 3.2300000190734863,
39
- 3.2300000190734863,
40
- 4.789999961853027,
41
- 7.400000095367432,
42
- 7.700000286102295,
43
- 9.09000015258789,
44
- 12.199999809265137,
45
- 17.670000076293945,
46
- 24.46000099182129,
47
- 28.57000160217285,
48
- 30.420001983642578,
49
- 30.840002059936523,
50
- 32.590003967285156,
51
- 32.93000411987305,
52
- 42.320003509521484,
53
- 44.96000289916992,
54
- 50.340003967285156,
55
- 50.45000457763672,
56
- 57.55000305175781,
57
- 57.93000411987305,
58
- 58.21000289916992,
59
- 60.1400032043457,
60
- 62.61000442504883,
61
- 62.62000274658203,
62
- 62.71000289916992,
63
- 63.1400032043457,
64
- 63.1400032043457,
65
- 63.77000427246094,
66
- 63.93000411987305,
67
- 63.96000289916992,
68
- 63.970001220703125,
69
- 64.02999877929688,
70
- 64.06999969482422,
71
- 64.08000183105469,
72
- 64.12000274658203,
73
- 64.41000366210938,
74
- 64.4800033569336,
75
- 64.51000213623047,
76
- 64.52999877929688,
77
- 64.83999633789062
78
- ],
79
- "short_factor": [
80
- 1.0,
81
- 1.0199999809265137,
82
- 1.0299999713897705,
83
- 1.0299999713897705,
84
- 1.0499999523162842,
85
- 1.0499999523162842,
86
- 1.0499999523162842,
87
- 1.0499999523162842,
88
- 1.0499999523162842,
89
- 1.0699999332427979,
90
- 1.0999999046325684,
91
- 1.1099998950958252,
92
- 1.1599998474121094,
93
- 1.1599998474121094,
94
- 1.1699998378753662,
95
- 1.2899998426437378,
96
- 1.339999794960022,
97
- 1.679999828338623,
98
- 1.7899998426437378,
99
- 1.8199998140335083,
100
- 1.8499997854232788,
101
- 1.8799997568130493,
102
- 1.9099997282028198,
103
- 1.9399996995925903,
104
- 1.9899996519088745,
105
- 2.0199997425079346,
106
- 2.0199997425079346,
107
- 2.0199997425079346,
108
- 2.0199997425079346,
109
- 2.0199997425079346,
110
- 2.0199997425079346,
111
- 2.0299997329711914,
112
- 2.0299997329711914,
113
- 2.0299997329711914,
114
- 2.0299997329711914,
115
- 2.0299997329711914,
116
- 2.0299997329711914,
117
- 2.0299997329711914,
118
- 2.0299997329711914,
119
- 2.0299997329711914,
120
- 2.0799996852874756,
121
- 2.0899996757507324,
122
- 2.189999580383301,
123
- 2.2199995517730713,
124
- 2.5899994373321533,
125
- 2.729999542236328,
126
- 2.749999523162842,
127
- 2.8399994373321533
128
- ],
129
- "type": "longrope"
130
- },
131
- "rope_theta": 10000.0,
132
- "sliding_window": 262144,
133
- "tie_word_embeddings": false,
134
- "torch_dtype": "bfloat16",
135
- "transformers_version": "4.47.0",
136
- "use_cache": true,
137
- "vocab_size": 32064
138
- }
 
1
+ {
2
+ "architectures": [
3
+ "Phi3ForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 1,
8
+ "embd_pdrop": 0.0,
9
+ "eos_token_id": 32000,
10
+ "hidden_act": "silu",
11
+ "hidden_size": 3072,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 8192,
14
+ "max_position_embeddings": 131072,
15
+ "model_type": "phi3",
16
+ "num_attention_heads": 32,
17
+ "num_hidden_layers": 32,
18
+ "num_key_value_heads": 32,
19
+ "original_max_position_embeddings": 4096,
20
+ "pad_token_id": 32000,
21
+ "resid_pdrop": 0.0,
22
+ "rms_norm_eps": 1e-05,
23
+ "rope_scaling": {
24
+ "long_factor": [
25
+ 1.0800000429153442,
26
+ 1.1100000143051147,
27
+ 1.1399999856948853,
28
+ 1.340000033378601,
29
+ 1.5899999141693115,
30
+ 1.600000023841858,
31
+ 1.6200000047683716,
32
+ 2.620000123977661,
33
+ 3.2300000190734863,
34
+ 3.2300000190734863,
35
+ 4.789999961853027,
36
+ 7.400000095367432,
37
+ 7.700000286102295,
38
+ 9.09000015258789,
39
+ 12.199999809265137,
40
+ 17.670000076293945,
41
+ 24.46000099182129,
42
+ 28.57000160217285,
43
+ 30.420001983642578,
44
+ 30.840002059936523,
45
+ 32.590003967285156,
46
+ 32.93000411987305,
47
+ 42.320003509521484,
48
+ 44.96000289916992,
49
+ 50.340003967285156,
50
+ 50.45000457763672,
51
+ 57.55000305175781,
52
+ 57.93000411987305,
53
+ 58.21000289916992,
54
+ 60.1400032043457,
55
+ 62.61000442504883,
56
+ 62.62000274658203,
57
+ 62.71000289916992,
58
+ 63.1400032043457,
59
+ 63.1400032043457,
60
+ 63.77000427246094,
61
+ 63.93000411987305,
62
+ 63.96000289916992,
63
+ 63.970001220703125,
64
+ 64.02999877929688,
65
+ 64.06999969482422,
66
+ 64.08000183105469,
67
+ 64.12000274658203,
68
+ 64.41000366210938,
69
+ 64.4800033569336,
70
+ 64.51000213623047,
71
+ 64.52999877929688,
72
+ 64.83999633789062
73
+ ],
74
+ "short_factor": [
75
+ 1.0,
76
+ 1.0199999809265137,
77
+ 1.0299999713897705,
78
+ 1.0299999713897705,
79
+ 1.0499999523162842,
80
+ 1.0499999523162842,
81
+ 1.0499999523162842,
82
+ 1.0499999523162842,
83
+ 1.0499999523162842,
84
+ 1.0699999332427979,
85
+ 1.0999999046325684,
86
+ 1.1099998950958252,
87
+ 1.1599998474121094,
88
+ 1.1599998474121094,
89
+ 1.1699998378753662,
90
+ 1.2899998426437378,
91
+ 1.339999794960022,
92
+ 1.679999828338623,
93
+ 1.7899998426437378,
94
+ 1.8199998140335083,
95
+ 1.8499997854232788,
96
+ 1.8799997568130493,
97
+ 1.9099997282028198,
98
+ 1.9399996995925903,
99
+ 1.9899996519088745,
100
+ 2.0199997425079346,
101
+ 2.0199997425079346,
102
+ 2.0199997425079346,
103
+ 2.0199997425079346,
104
+ 2.0199997425079346,
105
+ 2.0199997425079346,
106
+ 2.0299997329711914,
107
+ 2.0299997329711914,
108
+ 2.0299997329711914,
109
+ 2.0299997329711914,
110
+ 2.0299997329711914,
111
+ 2.0299997329711914,
112
+ 2.0299997329711914,
113
+ 2.0299997329711914,
114
+ 2.0299997329711914,
115
+ 2.0799996852874756,
116
+ 2.0899996757507324,
117
+ 2.189999580383301,
118
+ 2.2199995517730713,
119
+ 2.5899994373321533,
120
+ 2.729999542236328,
121
+ 2.749999523162842,
122
+ 2.8399994373321533
123
+ ],
124
+ "type": "longrope"
125
+ },
126
+ "rope_theta": 10000.0,
127
+ "sliding_window": 262144,
128
+ "tie_word_embeddings": false,
129
+ "torch_dtype": "bfloat16",
130
+ "transformers_version": "4.47.0",
131
+ "use_cache": true,
132
+ "vocab_size": 32064
133
+ }