Hugging Face's logo

Xenova
/
OpenELM-270M-Instruct

Text Generation
Transformers.js
ONNX
openelm
conversational
custom_code
Model card Files Community
Xenova HF staff commited on
Commit
965fec0
•
1 Parent(s): 07946dc

Upload folder using huggingface_hub

Browse files
Files changed (10) hide show
  1. config.json +88 -0
  2. generation_config.json +6 -0
  3. onnx/model.onnx +3 -0
  4. onnx/model_bnb4.onnx +3 -0
  5. onnx/model_fp16.onnx +3 -0
  6. onnx/model_int8.onnx +3 -0
  7. onnx/model_q4.onnx +3 -0
  8. onnx/model_quantized.onnx +3 -0
  9. onnx/model_uint8.onnx +3 -0
  10. quantize_config.json +132 -0
config.json ADDED
@@ -0,0 +1,88 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "apple/OpenELM-270M-Instruct",
3
+ "activation_fn_name": "swish",
4
+ "architectures": [
5
+ "OpenELMForCausalLM"
6
+ ],
7
+ "auto_map": {
8
+ "AutoConfig": "apple/OpenELM-270M-Instruct--configuration_openelm.OpenELMConfig",
9
+ "AutoModelForCausalLM": "apple/OpenELM-270M-Instruct--modeling_openelm.OpenELMForCausalLM"
10
+ },
11
+ "bos_token_id": 1,
12
+ "eos_token_id": 2,
13
+ "ffn_dim_divisor": 256,
14
+ "ffn_multipliers": [
15
+ 0.5,
16
+ 0.73,
17
+ 0.97,
18
+ 1.2,
19
+ 1.43,
20
+ 1.67,
21
+ 1.9,
22
+ 2.13,
23
+ 2.37,
24
+ 2.6,
25
+ 2.83,
26
+ 3.07,
27
+ 3.3,
28
+ 3.53,
29
+ 3.77,
30
+ 4.0
31
+ ],
32
+ "ffn_with_glu": true,
33
+ "head_dim": 64,
34
+ "initializer_range": 0.02,
35
+ "max_context_length": 2048,
36
+ "model_dim": 1280,
37
+ "model_type": "openelm",
38
+ "normalization_layer_name": "rms_norm",
39
+ "normalize_qk_projections": true,
40
+ "num_gqa_groups": 4,
41
+ "num_kv_heads": [
42
+ 3,
43
+ 3,
44
+ 3,
45
+ 3,
46
+ 3,
47
+ 4,
48
+ 4,
49
+ 4,
50
+ 4,
51
+ 4,
52
+ 4,
53
+ 4,
54
+ 5,
55
+ 5,
56
+ 5,
57
+ 5
58
+ ],
59
+ "num_query_heads": [
60
+ 12,
61
+ 12,
62
+ 12,
63
+ 12,
64
+ 12,
65
+ 16,
66
+ 16,
67
+ 16,
68
+ 16,
69
+ 16,
70
+ 16,
71
+ 16,
72
+ 20,
73
+ 20,
74
+ 20,
75
+ 20
76
+ ],
77
+ "num_transformer_layers": 16,
78
+ "qkv_multipliers": [
79
+ 0.5,
80
+ 1.0
81
+ ],
82
+ "rope_freq_constant": 10000,
83
+ "rope_max_length": 4096,
84
+ "share_input_output_layers": true,
85
+ "transformers_version": "4.38.2",
86
+ "use_cache": true,
87
+ "vocab_size": 32000
88
+ }
generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "transformers_version": "4.38.2"
6
+ }
onnx/model.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:630b4b8d4aac990e2b169e0928a109bd925e29343da2510f05eedcc28433ad18
3
+ size 1256920099
onnx/model_bnb4.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:220cb57c43b1e6e33c2cd10a56717babb0f223057245899d52d3be15136151c4
3
+ size 323707915
onnx/model_fp16.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:09b3a233192aab4f6837d318009808e3188abba7ae91e5b9c9c776765d764145
3
+ size 631188025
onnx/model_int8.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e05c62d2de6c20bc60d3058a7d93c8924c8a88d7439ab120de03427fd3feaac6
3
+ size 319957458
onnx/model_q4.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2cf69419c4b9fbba0c9fe57bb764f8d205abfc53d1a473759bb3bb6ec8ea94be
3
+ size 340675066
onnx/model_quantized.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e05c62d2de6c20bc60d3058a7d93c8924c8a88d7439ab120de03427fd3feaac6
3
+ size 319957458
onnx/model_uint8.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4d34a694969e060ed90aa0cde42991b12af108c2f08c16cce1e9c27958e256c2
3
+ size 319957487
quantize_config.json ADDED
@@ -0,0 +1,132 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "fp16": {},
3
+ "q8": {
4
+ "per_model_config": {
5
+ "model": {
6
+ "op_types": [
7
+ "Add",
8
+ "And",
9
+ "Cast",
10
+ "Concat",
11
+ "Constant",
12
+ "ConstantOfShape",
13
+ "Div",
14
+ "Equal",
15
+ "Expand",
16
+ "Gather",
17
+ "MatMul",
18
+ "Mul",
19
+ "Neg",
20
+ "Pow",
21
+ "Range",
22
+ "ReduceMean",
23
+ "Reshape",
24
+ "ScatterND",
25
+ "Shape",
26
+ "Sigmoid",
27
+ "Slice",
28
+ "Softmax",
29
+ "Split",
30
+ "Sqrt",
31
+ "Sub",
32
+ "Tile",
33
+ "Transpose",
34
+ "Unsqueeze",
35
+ "Where"
36
+ ],
37
+ "weight_type": "QInt8"
38
+ }
39
+ },
40
+ "per_channel": false,
41
+ "reduce_range": false
42
+ },
43
+ "int8": {
44
+ "per_model_config": {
45
+ "model": {
46
+ "op_types": [
47
+ "Add",
48
+ "And",
49
+ "Cast",
50
+ "Concat",
51
+ "Constant",
52
+ "ConstantOfShape",
53
+ "Div",
54
+ "Equal",
55
+ "Expand",
56
+ "Gather",
57
+ "MatMul",
58
+ "Mul",
59
+ "Neg",
60
+ "Pow",
61
+ "Range",
62
+ "ReduceMean",
63
+ "Reshape",
64
+ "ScatterND",
65
+ "Shape",
66
+ "Sigmoid",
67
+ "Slice",
68
+ "Softmax",
69
+ "Split",
70
+ "Sqrt",
71
+ "Sub",
72
+ "Tile",
73
+ "Transpose",
74
+ "Unsqueeze",
75
+ "Where"
76
+ ],
77
+ "weight_type": "QInt8"
78
+ }
79
+ },
80
+ "per_channel": false,
81
+ "reduce_range": false
82
+ },
83
+ "uint8": {
84
+ "per_model_config": {
85
+ "model": {
86
+ "op_types": [
87
+ "Add",
88
+ "And",
89
+ "Cast",
90
+ "Concat",
91
+ "Constant",
92
+ "ConstantOfShape",
93
+ "Div",
94
+ "Equal",
95
+ "Expand",
96
+ "Gather",
97
+ "MatMul",
98
+ "Mul",
99
+ "Neg",
100
+ "Pow",
101
+ "Range",
102
+ "ReduceMean",
103
+ "Reshape",
104
+ "ScatterND",
105
+ "Shape",
106
+ "Sigmoid",
107
+ "Slice",
108
+ "Softmax",
109
+ "Split",
110
+ "Sqrt",
111
+ "Sub",
112
+ "Tile",
113
+ "Transpose",
114
+ "Unsqueeze",
115
+ "Where"
116
+ ],
117
+ "weight_type": "QUInt8"
118
+ }
119
+ },
120
+ "per_channel": false,
121
+ "reduce_range": false
122
+ },
123
+ "q4": {
124
+ "block_size": 32,
125
+ "is_symmetric": true,
126
+ "accuracy_level": null
127
+ },
128
+ "bnb4": {
129
+ "block_size": 64,
130
+ "quant_type": 1
131
+ }
132
+ }