martinkorelic commited on
Commit
27c5668
·
verified ·
1 Parent(s): 578d880

Upload folder using huggingface_hub

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +5 -0
  2. README.md +40 -0
  3. mobiletransformers_manifest.json +293 -0
  4. optimum/export_report.json +35 -0
  5. optimum/supported_tasks.json +3 -0
  6. shared/tokenizer/chat_template.jinja +279 -0
  7. shared/tokenizer/mobiletransformers_tokenizer_config.json +17 -0
  8. shared/tokenizer/special_tokens_map.json +34 -0
  9. shared/tokenizer/tokenizer.json +3 -0
  10. shared/tokenizer/tokenizer_config.json +0 -0
  11. variants/cpu-int4/checksums.json +93 -0
  12. variants/cpu-int4/inference/chat_template.jinja +279 -0
  13. variants/cpu-int4/inference/config.json +57 -0
  14. variants/cpu-int4/inference/frozen_base.onnx.data +3 -0
  15. variants/cpu-int4/inference/genai_config.json +45 -0
  16. variants/cpu-int4/inference/generation_config.json +12 -0
  17. variants/cpu-int4/inference/merger_lora_fpin_fpout.onnx +3 -0
  18. variants/cpu-int4/inference/model.layers.0.self_attn.q_proj.MatMul.weight.bin +3 -0
  19. variants/cpu-int4/inference/model.layers.0.self_attn.q_proj.MatMul.weight.bin.sha256 +1 -0
  20. variants/cpu-int4/inference/model.layers.0.self_attn.v_proj.MatMul.weight.bin +3 -0
  21. variants/cpu-int4/inference/model.layers.0.self_attn.v_proj.MatMul.weight.bin.sha256 +1 -0
  22. variants/cpu-int4/inference/model.layers.1.self_attn.q_proj.MatMul.weight.bin +3 -0
  23. variants/cpu-int4/inference/model.layers.1.self_attn.q_proj.MatMul.weight.bin.sha256 +1 -0
  24. variants/cpu-int4/inference/model.layers.1.self_attn.v_proj.MatMul.weight.bin +3 -0
  25. variants/cpu-int4/inference/model.layers.1.self_attn.v_proj.MatMul.weight.bin.sha256 +1 -0
  26. variants/cpu-int4/inference/model.layers.10.self_attn.q_proj.MatMul.weight.bin +3 -0
  27. variants/cpu-int4/inference/model.layers.10.self_attn.q_proj.MatMul.weight.bin.sha256 +1 -0
  28. variants/cpu-int4/inference/model.layers.10.self_attn.v_proj.MatMul.weight.bin +3 -0
  29. variants/cpu-int4/inference/model.layers.10.self_attn.v_proj.MatMul.weight.bin.sha256 +1 -0
  30. variants/cpu-int4/inference/model.layers.11.self_attn.q_proj.MatMul.weight.bin +3 -0
  31. variants/cpu-int4/inference/model.layers.11.self_attn.q_proj.MatMul.weight.bin.sha256 +1 -0
  32. variants/cpu-int4/inference/model.layers.11.self_attn.v_proj.MatMul.weight.bin +3 -0
  33. variants/cpu-int4/inference/model.layers.11.self_attn.v_proj.MatMul.weight.bin.sha256 +1 -0
  34. variants/cpu-int4/inference/model.layers.12.self_attn.q_proj.MatMul.weight.bin +3 -0
  35. variants/cpu-int4/inference/model.layers.12.self_attn.q_proj.MatMul.weight.bin.sha256 +1 -0
  36. variants/cpu-int4/inference/model.layers.12.self_attn.v_proj.MatMul.weight.bin +3 -0
  37. variants/cpu-int4/inference/model.layers.12.self_attn.v_proj.MatMul.weight.bin.sha256 +1 -0
  38. variants/cpu-int4/inference/model.layers.13.self_attn.q_proj.MatMul.weight.bin +3 -0
  39. variants/cpu-int4/inference/model.layers.13.self_attn.q_proj.MatMul.weight.bin.sha256 +1 -0
  40. variants/cpu-int4/inference/model.layers.13.self_attn.v_proj.MatMul.weight.bin +3 -0
  41. variants/cpu-int4/inference/model.layers.13.self_attn.v_proj.MatMul.weight.bin.sha256 +1 -0
  42. variants/cpu-int4/inference/model.layers.14.self_attn.q_proj.MatMul.weight.bin +3 -0
  43. variants/cpu-int4/inference/model.layers.14.self_attn.q_proj.MatMul.weight.bin.sha256 +1 -0
  44. variants/cpu-int4/inference/model.layers.14.self_attn.v_proj.MatMul.weight.bin +3 -0
  45. variants/cpu-int4/inference/model.layers.14.self_attn.v_proj.MatMul.weight.bin.sha256 +1 -0
  46. variants/cpu-int4/inference/model.layers.15.self_attn.q_proj.MatMul.weight.bin +3 -0
  47. variants/cpu-int4/inference/model.layers.15.self_attn.q_proj.MatMul.weight.bin.sha256 +1 -0
  48. variants/cpu-int4/inference/model.layers.15.self_attn.v_proj.MatMul.weight.bin +3 -0
  49. variants/cpu-int4/inference/model.layers.15.self_attn.v_proj.MatMul.weight.bin.sha256 +1 -0
  50. variants/cpu-int4/inference/model.layers.16.self_attn.q_proj.MatMul.weight.bin +3 -0
.gitattributes CHANGED
@@ -33,3 +33,8 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ shared/tokenizer/tokenizer.json filter=lfs diff=lfs merge=lfs -text
37
+ variants/cpu-int4/inference/frozen_base.onnx.data filter=lfs diff=lfs merge=lfs -text
38
+ variants/cpu-int4/inference/model.onnx_data filter=lfs diff=lfs merge=lfs -text
39
+ variants/cpu-int4/inference/tokenizer.json filter=lfs diff=lfs merge=lfs -text
40
+ variants/cpu-int4/train/checkpoint filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: google/functiongemma-270m-it
3
+ library_name: mobiletransformers
4
+ pipeline_tag: text-generation
5
+ tags:
6
+ - mobiletransformers
7
+ - onnx
8
+ - on-device
9
+ - android
10
+ - lora
11
+ - int4
12
+ ---
13
+
14
+ # google/functiongemma-270m-it — MobileTransformers package
15
+
16
+ On-device (Android) package exported from **google/functiongemma-270m-it** with MobileTransformers.
17
+
18
+ ## Provenance
19
+ - Base model: `google/functiongemma-270m-it`
20
+ - Selected task: `text-generation-with-past`
21
+ - PEFT methods: lora
22
+ - Quantization: int4
23
+ - Toolchain: optimum-onnx 0.1.0, transformers 4.57.6, ort-training 1.23.0+cpu, ort-genai None
24
+
25
+ ## Licenses
26
+ - Framework: not declared in this package — see the repository
27
+ - Base model weights: see the base model above (this package redistributes an export of those weights, so their terms govern its contents)
28
+
29
+ ## Android runtime
30
+ - Minimum API: 28
31
+ - Recommended device memory (MB): None
32
+ - Required ABIs: any
33
+
34
+ ## Variants
35
+
36
+ | id | EP | quant | engines | features | min API | rec. RAM (MB) |
37
+ | --- | --- | --- | --- | --- | --- | --- |
38
+ | cpu-int4 | cpu | int4 | native | core, inference, train | 28 | None |
39
+
40
+ Default variant: `cpu-int4`.
mobiletransformers_manifest.json ADDED
@@ -0,0 +1,293 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "androidRuntime": {
3
+ "minimumAndroidApi": 28,
4
+ "recommendedDeviceMemoryMb": null,
5
+ "requiredAbis": []
6
+ },
7
+ "architectures": [
8
+ "gemma3_text"
9
+ ],
10
+ "artifactFormatVersion": 1,
11
+ "baseModelId": "google/functiongemma-270m-it",
12
+ "defaultVariant": "cpu-int4",
13
+ "downloadPlan": {
14
+ "cpu-int4": {
15
+ "checksums": [
16
+ "variants/cpu-int4/checksums.json"
17
+ ],
18
+ "core": [
19
+ "mobiletransformers_manifest.json",
20
+ "shared/tokenizer/**",
21
+ "shared/chat_template.jinja",
22
+ "shared/config.json",
23
+ "shared/generation_config.json"
24
+ ],
25
+ "genai": [],
26
+ "inference": [
27
+ "variants/cpu-int4/inference/**"
28
+ ],
29
+ "rag": [],
30
+ "train": [
31
+ "variants/cpu-int4/train/**"
32
+ ]
33
+ }
34
+ },
35
+ "exportedAt": "",
36
+ "fileSizes": {
37
+ "optimum/export_report.json": 822,
38
+ "optimum/supported_tasks.json": 34,
39
+ "shared/tokenizer/chat_template.jinja": 13792,
40
+ "shared/tokenizer/mobiletransformers_tokenizer_config.json": 356,
41
+ "shared/tokenizer/special_tokens_map.json": 706,
42
+ "shared/tokenizer/tokenizer.json": 33384899,
43
+ "shared/tokenizer/tokenizer_config.json": 1155714,
44
+ "variants/cpu-int4/checksums.json": 13424,
45
+ "variants/cpu-int4/inference/chat_template.jinja": 13792,
46
+ "variants/cpu-int4/inference/config.json": 1358,
47
+ "variants/cpu-int4/inference/frozen_base.onnx.data": 1684498944,
48
+ "variants/cpu-int4/inference/genai_config.json": 1044,
49
+ "variants/cpu-int4/inference/generation_config.json": 176,
50
+ "variants/cpu-int4/inference/merger_lora_fpin_fpout.onnx": 811,
51
+ "variants/cpu-int4/inference/model.layers.0.self_attn.q_proj.MatMul.weight.bin": 2621440,
52
+ "variants/cpu-int4/inference/model.layers.0.self_attn.q_proj.MatMul.weight.bin.sha256": 65,
53
+ "variants/cpu-int4/inference/model.layers.0.self_attn.v_proj.MatMul.weight.bin": 655360,
54
+ "variants/cpu-int4/inference/model.layers.0.self_attn.v_proj.MatMul.weight.bin.sha256": 65,
55
+ "variants/cpu-int4/inference/model.layers.1.self_attn.q_proj.MatMul.weight.bin": 2621440,
56
+ "variants/cpu-int4/inference/model.layers.1.self_attn.q_proj.MatMul.weight.bin.sha256": 65,
57
+ "variants/cpu-int4/inference/model.layers.1.self_attn.v_proj.MatMul.weight.bin": 655360,
58
+ "variants/cpu-int4/inference/model.layers.1.self_attn.v_proj.MatMul.weight.bin.sha256": 65,
59
+ "variants/cpu-int4/inference/model.layers.10.self_attn.q_proj.MatMul.weight.bin": 2621440,
60
+ "variants/cpu-int4/inference/model.layers.10.self_attn.q_proj.MatMul.weight.bin.sha256": 65,
61
+ "variants/cpu-int4/inference/model.layers.10.self_attn.v_proj.MatMul.weight.bin": 655360,
62
+ "variants/cpu-int4/inference/model.layers.10.self_attn.v_proj.MatMul.weight.bin.sha256": 65,
63
+ "variants/cpu-int4/inference/model.layers.11.self_attn.q_proj.MatMul.weight.bin": 2621440,
64
+ "variants/cpu-int4/inference/model.layers.11.self_attn.q_proj.MatMul.weight.bin.sha256": 65,
65
+ "variants/cpu-int4/inference/model.layers.11.self_attn.v_proj.MatMul.weight.bin": 655360,
66
+ "variants/cpu-int4/inference/model.layers.11.self_attn.v_proj.MatMul.weight.bin.sha256": 65,
67
+ "variants/cpu-int4/inference/model.layers.12.self_attn.q_proj.MatMul.weight.bin": 2621440,
68
+ "variants/cpu-int4/inference/model.layers.12.self_attn.q_proj.MatMul.weight.bin.sha256": 65,
69
+ "variants/cpu-int4/inference/model.layers.12.self_attn.v_proj.MatMul.weight.bin": 655360,
70
+ "variants/cpu-int4/inference/model.layers.12.self_attn.v_proj.MatMul.weight.bin.sha256": 65,
71
+ "variants/cpu-int4/inference/model.layers.13.self_attn.q_proj.MatMul.weight.bin": 2621440,
72
+ "variants/cpu-int4/inference/model.layers.13.self_attn.q_proj.MatMul.weight.bin.sha256": 65,
73
+ "variants/cpu-int4/inference/model.layers.13.self_attn.v_proj.MatMul.weight.bin": 655360,
74
+ "variants/cpu-int4/inference/model.layers.13.self_attn.v_proj.MatMul.weight.bin.sha256": 65,
75
+ "variants/cpu-int4/inference/model.layers.14.self_attn.q_proj.MatMul.weight.bin": 2621440,
76
+ "variants/cpu-int4/inference/model.layers.14.self_attn.q_proj.MatMul.weight.bin.sha256": 65,
77
+ "variants/cpu-int4/inference/model.layers.14.self_attn.v_proj.MatMul.weight.bin": 655360,
78
+ "variants/cpu-int4/inference/model.layers.14.self_attn.v_proj.MatMul.weight.bin.sha256": 65,
79
+ "variants/cpu-int4/inference/model.layers.15.self_attn.q_proj.MatMul.weight.bin": 2621440,
80
+ "variants/cpu-int4/inference/model.layers.15.self_attn.q_proj.MatMul.weight.bin.sha256": 65,
81
+ "variants/cpu-int4/inference/model.layers.15.self_attn.v_proj.MatMul.weight.bin": 655360,
82
+ "variants/cpu-int4/inference/model.layers.15.self_attn.v_proj.MatMul.weight.bin.sha256": 65,
83
+ "variants/cpu-int4/inference/model.layers.16.self_attn.q_proj.MatMul.weight.bin": 2621440,
84
+ "variants/cpu-int4/inference/model.layers.16.self_attn.q_proj.MatMul.weight.bin.sha256": 65,
85
+ "variants/cpu-int4/inference/model.layers.16.self_attn.v_proj.MatMul.weight.bin": 655360,
86
+ "variants/cpu-int4/inference/model.layers.16.self_attn.v_proj.MatMul.weight.bin.sha256": 65,
87
+ "variants/cpu-int4/inference/model.layers.17.self_attn.q_proj.MatMul.weight.bin": 2621440,
88
+ "variants/cpu-int4/inference/model.layers.17.self_attn.q_proj.MatMul.weight.bin.sha256": 65,
89
+ "variants/cpu-int4/inference/model.layers.17.self_attn.v_proj.MatMul.weight.bin": 655360,
90
+ "variants/cpu-int4/inference/model.layers.17.self_attn.v_proj.MatMul.weight.bin.sha256": 65,
91
+ "variants/cpu-int4/inference/model.layers.2.self_attn.q_proj.MatMul.weight.bin": 2621440,
92
+ "variants/cpu-int4/inference/model.layers.2.self_attn.q_proj.MatMul.weight.bin.sha256": 65,
93
+ "variants/cpu-int4/inference/model.layers.2.self_attn.v_proj.MatMul.weight.bin": 655360,
94
+ "variants/cpu-int4/inference/model.layers.2.self_attn.v_proj.MatMul.weight.bin.sha256": 65,
95
+ "variants/cpu-int4/inference/model.layers.3.self_attn.q_proj.MatMul.weight.bin": 2621440,
96
+ "variants/cpu-int4/inference/model.layers.3.self_attn.q_proj.MatMul.weight.bin.sha256": 65,
97
+ "variants/cpu-int4/inference/model.layers.3.self_attn.v_proj.MatMul.weight.bin": 655360,
98
+ "variants/cpu-int4/inference/model.layers.3.self_attn.v_proj.MatMul.weight.bin.sha256": 65,
99
+ "variants/cpu-int4/inference/model.layers.4.self_attn.q_proj.MatMul.weight.bin": 2621440,
100
+ "variants/cpu-int4/inference/model.layers.4.self_attn.q_proj.MatMul.weight.bin.sha256": 65,
101
+ "variants/cpu-int4/inference/model.layers.4.self_attn.v_proj.MatMul.weight.bin": 655360,
102
+ "variants/cpu-int4/inference/model.layers.4.self_attn.v_proj.MatMul.weight.bin.sha256": 65,
103
+ "variants/cpu-int4/inference/model.layers.5.self_attn.q_proj.MatMul.weight.bin": 2621440,
104
+ "variants/cpu-int4/inference/model.layers.5.self_attn.q_proj.MatMul.weight.bin.sha256": 65,
105
+ "variants/cpu-int4/inference/model.layers.5.self_attn.v_proj.MatMul.weight.bin": 655360,
106
+ "variants/cpu-int4/inference/model.layers.5.self_attn.v_proj.MatMul.weight.bin.sha256": 65,
107
+ "variants/cpu-int4/inference/model.layers.6.self_attn.q_proj.MatMul.weight.bin": 2621440,
108
+ "variants/cpu-int4/inference/model.layers.6.self_attn.q_proj.MatMul.weight.bin.sha256": 65,
109
+ "variants/cpu-int4/inference/model.layers.6.self_attn.v_proj.MatMul.weight.bin": 655360,
110
+ "variants/cpu-int4/inference/model.layers.6.self_attn.v_proj.MatMul.weight.bin.sha256": 65,
111
+ "variants/cpu-int4/inference/model.layers.7.self_attn.q_proj.MatMul.weight.bin": 2621440,
112
+ "variants/cpu-int4/inference/model.layers.7.self_attn.q_proj.MatMul.weight.bin.sha256": 65,
113
+ "variants/cpu-int4/inference/model.layers.7.self_attn.v_proj.MatMul.weight.bin": 655360,
114
+ "variants/cpu-int4/inference/model.layers.7.self_attn.v_proj.MatMul.weight.bin.sha256": 65,
115
+ "variants/cpu-int4/inference/model.layers.8.self_attn.q_proj.MatMul.weight.bin": 2621440,
116
+ "variants/cpu-int4/inference/model.layers.8.self_attn.q_proj.MatMul.weight.bin.sha256": 65,
117
+ "variants/cpu-int4/inference/model.layers.8.self_attn.v_proj.MatMul.weight.bin": 655360,
118
+ "variants/cpu-int4/inference/model.layers.8.self_attn.v_proj.MatMul.weight.bin.sha256": 65,
119
+ "variants/cpu-int4/inference/model.layers.9.self_attn.q_proj.MatMul.weight.bin": 2621440,
120
+ "variants/cpu-int4/inference/model.layers.9.self_attn.q_proj.MatMul.weight.bin.sha256": 65,
121
+ "variants/cpu-int4/inference/model.layers.9.self_attn.v_proj.MatMul.weight.bin": 655360,
122
+ "variants/cpu-int4/inference/model.layers.9.self_attn.v_proj.MatMul.weight.bin.sha256": 65,
123
+ "variants/cpu-int4/inference/model.onnx": 920639,
124
+ "variants/cpu-int4/inference/model.onnx_data": 1743481344,
125
+ "variants/cpu-int4/inference/optimum_config.json": 334,
126
+ "variants/cpu-int4/inference/special_tokens_map.json": 706,
127
+ "variants/cpu-int4/inference/tokenizer.json": 33384899,
128
+ "variants/cpu-int4/inference/tokenizer_config.json": 1155714,
129
+ "variants/cpu-int4/inference/weight_handoff_map.json": 53841,
130
+ "variants/cpu-int4/train/checkpoint": 314028424,
131
+ "variants/cpu-int4/train/eval_model.onnx": 1901828,
132
+ "variants/cpu-int4/train/optimizer_model.onnx": 538,
133
+ "variants/cpu-int4/train/trainable_parameters.json": 7011,
134
+ "variants/cpu-int4/train/training_config.json": 19841,
135
+ "variants/cpu-int4/train/training_model.onnx": 1688100
136
+ },
137
+ "license": {
138
+ "baseModelWeights": null,
139
+ "framework": null,
140
+ "noticeFile": null
141
+ },
142
+ "minReaderVersion": "1.0",
143
+ "mobiletransformersVersion": "0.1.0",
144
+ "onnxRuntimeGenAIVersion": null,
145
+ "onnxRuntimeTrainingVersion": "1.23.0+cpu",
146
+ "optimumOnnxVersion": "0.1.0",
147
+ "peftMethods": [
148
+ "lora"
149
+ ],
150
+ "quantization": [
151
+ "int4"
152
+ ],
153
+ "requiredFiles": [
154
+ "mobiletransformers_manifest.json",
155
+ "shared/tokenizer/tokenizer.json",
156
+ "variants/cpu-int4/inference/model.onnx"
157
+ ],
158
+ "schemaVersion": "1.0",
159
+ "selectedTask": "text-generation-with-past",
160
+ "sha256": {
161
+ "optimum/export_report.json": "edc6a33e6199cbc47846e57f770d37b56822796cda4df3ebfe51eeb23dbec046",
162
+ "optimum/supported_tasks.json": "2591d52d3133d5bd71e384d09e57f48d29e84f4e9a358697595bb4bcb56dcfdf",
163
+ "shared/tokenizer/chat_template.jinja": "db61fb01017bd82401d3ffca4f8e066cd56ff6d38d2a30c4058770c0bf7ab49b",
164
+ "shared/tokenizer/mobiletransformers_tokenizer_config.json": "2b8628a7ad20677fd280311678611f19cf97860938f18ee678083cc0e550c668",
165
+ "shared/tokenizer/special_tokens_map.json": "d39df6f8ddb132a622394f05a603b166ba5f22d485483e30d06c1821ef4ac9b4",
166
+ "shared/tokenizer/tokenizer.json": "b6b09a0b4a803ad453063ca4bb49a784540e8120004e2450e025df2b27d41fb2",
167
+ "shared/tokenizer/tokenizer_config.json": "c33878647bdc69ce85d4b9aeef953815afd8405f8d85001ab0bbb89aa14d2554",
168
+ "variants/cpu-int4/checksums.json": "2fb0f6eee9752e511aab5cea2529e31a219233d36635a9458d2f328d512515be",
169
+ "variants/cpu-int4/inference/chat_template.jinja": "db61fb01017bd82401d3ffca4f8e066cd56ff6d38d2a30c4058770c0bf7ab49b",
170
+ "variants/cpu-int4/inference/config.json": "88a795b85d50b1642e66193935a64c34326fa76efe107f9fceeafd31aa3e2af9",
171
+ "variants/cpu-int4/inference/frozen_base.onnx.data": "9c14b3237169fd6023199017e1949b02bbfe0ef1e80ac10c2d72bba670ebfa82",
172
+ "variants/cpu-int4/inference/genai_config.json": "acac04545d607a84f676049824d2e113f9367179741494a46279551b438c6efc",
173
+ "variants/cpu-int4/inference/generation_config.json": "90f3b346c07b98c000b6de297e805a83f504d6856b01c405a746d805c05444c2",
174
+ "variants/cpu-int4/inference/merger_lora_fpin_fpout.onnx": "953a8ad4704235c19ec06150788107921ba7f2578a5880a32cd59b473544712c",
175
+ "variants/cpu-int4/inference/model.layers.0.self_attn.q_proj.MatMul.weight.bin": "de5441ec8bb19c3a0038e7f3bd3cb25d9b81599d1e6977ceecdc1d894ede2fa1",
176
+ "variants/cpu-int4/inference/model.layers.0.self_attn.q_proj.MatMul.weight.bin.sha256": "b303f006c606265b707027dccaa1c22f9f19b7920a8d51e442cbcfde886d3b65",
177
+ "variants/cpu-int4/inference/model.layers.0.self_attn.v_proj.MatMul.weight.bin": "aced1f8faec6416a6cbee9227400021af9b1f4c0d25feddc54e255f1210575fb",
178
+ "variants/cpu-int4/inference/model.layers.0.self_attn.v_proj.MatMul.weight.bin.sha256": "e41d2d8d47eaea11a011971b193d1ea7f15222a23dc2c2d7342c983d5fad5d58",
179
+ "variants/cpu-int4/inference/model.layers.1.self_attn.q_proj.MatMul.weight.bin": "cc0c039a777068e664ea4916413b822ce497007dc31f22f7e7cc27f9ed0bd820",
180
+ "variants/cpu-int4/inference/model.layers.1.self_attn.q_proj.MatMul.weight.bin.sha256": "1e1651baa468edd1ec0a787271c362646bcf4669d06816750fb9907bda6161ed",
181
+ "variants/cpu-int4/inference/model.layers.1.self_attn.v_proj.MatMul.weight.bin": "e2cde258f00749809d3b5e6a3838dbccb866dfdc228c86e5375a1c3fb3be14b2",
182
+ "variants/cpu-int4/inference/model.layers.1.self_attn.v_proj.MatMul.weight.bin.sha256": "d670153828a24efe77701e8491f5bc182fd88c9f27b924c750b9c20ca4fbf89c",
183
+ "variants/cpu-int4/inference/model.layers.10.self_attn.q_proj.MatMul.weight.bin": "864537f8b063185411edf60cc77bb704f3ddd3d39b6f19f04bd0da2bf1ba5b5e",
184
+ "variants/cpu-int4/inference/model.layers.10.self_attn.q_proj.MatMul.weight.bin.sha256": "286bfb981d94d988671f4572b770728b9a285f400ec4edc6ad83c9ea499d2c40",
185
+ "variants/cpu-int4/inference/model.layers.10.self_attn.v_proj.MatMul.weight.bin": "f8aedd8998857b74a4e93e3182911873cd9ad547f8f0ed8212856e8f0beaef00",
186
+ "variants/cpu-int4/inference/model.layers.10.self_attn.v_proj.MatMul.weight.bin.sha256": "528f124d7e9657be8b57303ec6248483bb0c6dde4c84084575e3c9afd0b8188b",
187
+ "variants/cpu-int4/inference/model.layers.11.self_attn.q_proj.MatMul.weight.bin": "534363c1244562ba5cc4043f27b4da97fa3a48dc2a60772532816b5f11865b91",
188
+ "variants/cpu-int4/inference/model.layers.11.self_attn.q_proj.MatMul.weight.bin.sha256": "19a221e71a3a79d359f1168ea1a3f88003340820f8ecffae5f5472d71be6700c",
189
+ "variants/cpu-int4/inference/model.layers.11.self_attn.v_proj.MatMul.weight.bin": "9b3cc8d43d80eae0f674534447d7845b17a4c779fa74adb1de07254f65de03f0",
190
+ "variants/cpu-int4/inference/model.layers.11.self_attn.v_proj.MatMul.weight.bin.sha256": "643bbdcb8d0381f20eaf58b1d8d9ff7c3f6435ec71d5ee7d4a93d7b60d541a83",
191
+ "variants/cpu-int4/inference/model.layers.12.self_attn.q_proj.MatMul.weight.bin": "5213011869521ddedba25e37944d311b30632ec3b2c061caeb6671d4f4b9596f",
192
+ "variants/cpu-int4/inference/model.layers.12.self_attn.q_proj.MatMul.weight.bin.sha256": "92e8c3fb1b76481b06c2ff2dd3ae3a99ad52477b274511b4ba3944dc9f376858",
193
+ "variants/cpu-int4/inference/model.layers.12.self_attn.v_proj.MatMul.weight.bin": "8dfd86b40b9e613538f07c8297cf3fb54cb3ebf99fe4918c30c8c51bea4da5c0",
194
+ "variants/cpu-int4/inference/model.layers.12.self_attn.v_proj.MatMul.weight.bin.sha256": "ca5d5b8d2723dcb8a31ed895ebd6f72d3111de86981d1f10a57fd64d393a5ceb",
195
+ "variants/cpu-int4/inference/model.layers.13.self_attn.q_proj.MatMul.weight.bin": "3e58dbbd1a83c01943979052519976f15dd2e13b4502fe72edf522f86960d618",
196
+ "variants/cpu-int4/inference/model.layers.13.self_attn.q_proj.MatMul.weight.bin.sha256": "8a75590cded9ed4eb8518b18d4f5694fd4131d8ef225bfff9d148412716fe7c6",
197
+ "variants/cpu-int4/inference/model.layers.13.self_attn.v_proj.MatMul.weight.bin": "21666b649c20492830d5f37db41e2ba54d270228c20a532eea2104fa3c61396e",
198
+ "variants/cpu-int4/inference/model.layers.13.self_attn.v_proj.MatMul.weight.bin.sha256": "327d2d9dd53b71a1662bb5751d554ef719aff537095da4bd8ac2aaaea47de555",
199
+ "variants/cpu-int4/inference/model.layers.14.self_attn.q_proj.MatMul.weight.bin": "ef8d82392a5ac5cf37f65c638a965c7c949111ec783b0c947ecfc8d261e1ad9f",
200
+ "variants/cpu-int4/inference/model.layers.14.self_attn.q_proj.MatMul.weight.bin.sha256": "c183ea62d344b6331a35070bdac7091347be27f629a00a3569acf40f01c87dd2",
201
+ "variants/cpu-int4/inference/model.layers.14.self_attn.v_proj.MatMul.weight.bin": "c1c727d66921e1dceec4142a2a2c36254189234bd90b09db4125927e70b8e458",
202
+ "variants/cpu-int4/inference/model.layers.14.self_attn.v_proj.MatMul.weight.bin.sha256": "75c2d4a4ac3606a197b259bb309ada6a55341a4c3c3e68ab2b5d29985699cd88",
203
+ "variants/cpu-int4/inference/model.layers.15.self_attn.q_proj.MatMul.weight.bin": "31f7849f39f1b1fe0d7ef163aae56279a8f48eac0ec81dceae7dd7fda77dd293",
204
+ "variants/cpu-int4/inference/model.layers.15.self_attn.q_proj.MatMul.weight.bin.sha256": "c314704b435abe1cd3ba672eb1a408e7c6f7c808a215bfb7938d27f5c4583dbe",
205
+ "variants/cpu-int4/inference/model.layers.15.self_attn.v_proj.MatMul.weight.bin": "8f5878be79b76139fc9cc5724e91074c742fffd021b2b7bd6f0406775ec69365",
206
+ "variants/cpu-int4/inference/model.layers.15.self_attn.v_proj.MatMul.weight.bin.sha256": "17e8852305a2dd0be3788629535015c3ba347bd8a84a318c1a2740846ea656d3",
207
+ "variants/cpu-int4/inference/model.layers.16.self_attn.q_proj.MatMul.weight.bin": "cc87b80c44df0b66e627944acd40e87a0b63c77a86d1cfa55085dad5c5610de7",
208
+ "variants/cpu-int4/inference/model.layers.16.self_attn.q_proj.MatMul.weight.bin.sha256": "331eece77b3bc5bea26c224ee395a6c864e2e98667cd6bd0ef8a20c48aeaa27c",
209
+ "variants/cpu-int4/inference/model.layers.16.self_attn.v_proj.MatMul.weight.bin": "0c1f91d46f8f885ca1970f61ada3e69fe7e2a44c4f103beb456e84fb0f39cb13",
210
+ "variants/cpu-int4/inference/model.layers.16.self_attn.v_proj.MatMul.weight.bin.sha256": "c25e491df56cb7902bc53dfe2fdc5f69cc6c94d56a73a1e354920ace0cdedf8c",
211
+ "variants/cpu-int4/inference/model.layers.17.self_attn.q_proj.MatMul.weight.bin": "75aa1f482c3a16420a092f912a073abc981fc584038eead4fc4dad219c4a18df",
212
+ "variants/cpu-int4/inference/model.layers.17.self_attn.q_proj.MatMul.weight.bin.sha256": "5dccc067c68edcfb2aafd0ecfe966965f274afc3455699e6add6a2f4b3c0151f",
213
+ "variants/cpu-int4/inference/model.layers.17.self_attn.v_proj.MatMul.weight.bin": "c331cf9885fc9a470111e2e8e59018a2ffa235ee1441aa005789e321aee80df3",
214
+ "variants/cpu-int4/inference/model.layers.17.self_attn.v_proj.MatMul.weight.bin.sha256": "b3058d47101dd69928aa1b6e78574dd47e8a9eedda003972accac1fcbc50c53f",
215
+ "variants/cpu-int4/inference/model.layers.2.self_attn.q_proj.MatMul.weight.bin": "0aab3b51d26366fda266a9a030b64fa0097a3c9fe0f2ac8b45cd455ed6d00844",
216
+ "variants/cpu-int4/inference/model.layers.2.self_attn.q_proj.MatMul.weight.bin.sha256": "0c44bd5f5b21e9b5292a13d4069c6b313c9b9aa80881c3b469ab0034c0bedd1e",
217
+ "variants/cpu-int4/inference/model.layers.2.self_attn.v_proj.MatMul.weight.bin": "f995a03e11e45e51c9038b610940e2388dea7f89b186c9c32bf0774c7ad56676",
218
+ "variants/cpu-int4/inference/model.layers.2.self_attn.v_proj.MatMul.weight.bin.sha256": "b7f5aa1ce6d2e1be556f65fb6ca619e7591e63dd975534dae239e71290ac5a7b",
219
+ "variants/cpu-int4/inference/model.layers.3.self_attn.q_proj.MatMul.weight.bin": "5f256353f99cd60325cd88345d8095a1d8bf3ec2b2ed8b7dab08368018fd8335",
220
+ "variants/cpu-int4/inference/model.layers.3.self_attn.q_proj.MatMul.weight.bin.sha256": "00990e7e4ec37977e4c3024989b9dedab6f8c94a76755aa59ff21bdde5eae41e",
221
+ "variants/cpu-int4/inference/model.layers.3.self_attn.v_proj.MatMul.weight.bin": "7412ad9549d423c2cfbe4922299f5759b2e30e6839ef89cf99f8880ea8872ee5",
222
+ "variants/cpu-int4/inference/model.layers.3.self_attn.v_proj.MatMul.weight.bin.sha256": "e832cd832df1e7c9249d968196ff8da050eb1f4bb733e6ed1f3df80988037bad",
223
+ "variants/cpu-int4/inference/model.layers.4.self_attn.q_proj.MatMul.weight.bin": "ad0a0d5c5db67a43ea9793d34d5ef09e081772e8e5fba2b771ee95983ecda45a",
224
+ "variants/cpu-int4/inference/model.layers.4.self_attn.q_proj.MatMul.weight.bin.sha256": "6baba6d67fea470291b3da65d1cf9834f6e103b9a6845c8d753c958a2d873b5e",
225
+ "variants/cpu-int4/inference/model.layers.4.self_attn.v_proj.MatMul.weight.bin": "f59967d4272b6bd14e86c0cdc9fed9f1bcb21f67faa8a1d769935f710247f1b3",
226
+ "variants/cpu-int4/inference/model.layers.4.self_attn.v_proj.MatMul.weight.bin.sha256": "294e36d47d7ea2796ec74012dba0d6e672eaade5d8fed5c4fc15885d735fd35a",
227
+ "variants/cpu-int4/inference/model.layers.5.self_attn.q_proj.MatMul.weight.bin": "ff1c5fe943f8959a8b9d09640aec44d881f035f771416ec88ac929e5d4fe6b4b",
228
+ "variants/cpu-int4/inference/model.layers.5.self_attn.q_proj.MatMul.weight.bin.sha256": "aa4e396ee86ecd0b1538e96230fc842659e7ec66c927493682fa914efce3047b",
229
+ "variants/cpu-int4/inference/model.layers.5.self_attn.v_proj.MatMul.weight.bin": "4ee8c2d65e018f46002157c99b0fd025cfc523e03b07ca6fe26da71659f51c70",
230
+ "variants/cpu-int4/inference/model.layers.5.self_attn.v_proj.MatMul.weight.bin.sha256": "31352189895dd2e1d92b88721c0950070cdd2caa8c68c8cae31233825ec5b667",
231
+ "variants/cpu-int4/inference/model.layers.6.self_attn.q_proj.MatMul.weight.bin": "353d2ac53c60dfce33c57e501010ec0dac799d9050617cf6f9f818d37e1a2296",
232
+ "variants/cpu-int4/inference/model.layers.6.self_attn.q_proj.MatMul.weight.bin.sha256": "e869d030595d5ed13ac046acfc08fd9487b3a9ba9e940edbb4724eba90b84167",
233
+ "variants/cpu-int4/inference/model.layers.6.self_attn.v_proj.MatMul.weight.bin": "e5d43600c7a68cc5889563120b46780f107945e26c9385b5fca04c3a22960843",
234
+ "variants/cpu-int4/inference/model.layers.6.self_attn.v_proj.MatMul.weight.bin.sha256": "fd3c65bcc2e5652073fdf480a8be666611a1e20d26662e648e5cdd7fb4e0340c",
235
+ "variants/cpu-int4/inference/model.layers.7.self_attn.q_proj.MatMul.weight.bin": "e9d722fef4ecc4e3dffddd5b1527bcca70af70978eb812df6d7c4452c4c10b90",
236
+ "variants/cpu-int4/inference/model.layers.7.self_attn.q_proj.MatMul.weight.bin.sha256": "5d11e119c112c0f2cd8a46a10856aa4e3a467048928ddd0de0891dec95164b18",
237
+ "variants/cpu-int4/inference/model.layers.7.self_attn.v_proj.MatMul.weight.bin": "d50fe1d3c9c265fc38d169bd63fa727062f145719419564a124a5c06fbe59b04",
238
+ "variants/cpu-int4/inference/model.layers.7.self_attn.v_proj.MatMul.weight.bin.sha256": "bb5995d820468db5e2c676bf0960061c85fff3ccaeebb0c76b89f905b400f0bc",
239
+ "variants/cpu-int4/inference/model.layers.8.self_attn.q_proj.MatMul.weight.bin": "eee1373ed2ee35607cda09ceeef4c92b7ab331a419ffb795c821a58c563e884b",
240
+ "variants/cpu-int4/inference/model.layers.8.self_attn.q_proj.MatMul.weight.bin.sha256": "ebf030acd3a19f40d679c3e8917c15a97e122f4f8076eff7043758fefbef1b4c",
241
+ "variants/cpu-int4/inference/model.layers.8.self_attn.v_proj.MatMul.weight.bin": "3b9e52ee458990d65fd339e4e544d5f379b1600dd4852fc3084a32e96bc00a0b",
242
+ "variants/cpu-int4/inference/model.layers.8.self_attn.v_proj.MatMul.weight.bin.sha256": "6142f9fd4e7ee9f118757cb582ce4d16c3105cb389343ee8aef7dbf4b83b939f",
243
+ "variants/cpu-int4/inference/model.layers.9.self_attn.q_proj.MatMul.weight.bin": "68c5c1b401f009fd33795a986edbef642ed6bb2e8cfdee72f8ef5d9930a608b4",
244
+ "variants/cpu-int4/inference/model.layers.9.self_attn.q_proj.MatMul.weight.bin.sha256": "ec94b226ebc6be2d0eff5dfd48edb112728ec2d7fb4968481f811a87b31a16ae",
245
+ "variants/cpu-int4/inference/model.layers.9.self_attn.v_proj.MatMul.weight.bin": "b029f98f39c8aef6ec3676ae87106eab90d9521ff05ab1c7812963f8a1929555",
246
+ "variants/cpu-int4/inference/model.layers.9.self_attn.v_proj.MatMul.weight.bin.sha256": "9004e7234ef3e8231241955e74f0cfd7880d2ba47a4b6f4f43b3e86759ea7793",
247
+ "variants/cpu-int4/inference/model.onnx": "7816d03ce0be987e340c51ab64916f13a0f70a68630c3a2de82370a32a6fabfa",
248
+ "variants/cpu-int4/inference/model.onnx_data": "984595b520feb718e9dde76cc6dcd2bcc590bb67987c4da01b15a6ea54cade05",
249
+ "variants/cpu-int4/inference/optimum_config.json": "78abd079b5f45b977fad47a63f173701f6681d555c1372a800221831ce811e4f",
250
+ "variants/cpu-int4/inference/special_tokens_map.json": "d39df6f8ddb132a622394f05a603b166ba5f22d485483e30d06c1821ef4ac9b4",
251
+ "variants/cpu-int4/inference/tokenizer.json": "b6b09a0b4a803ad453063ca4bb49a784540e8120004e2450e025df2b27d41fb2",
252
+ "variants/cpu-int4/inference/tokenizer_config.json": "c33878647bdc69ce85d4b9aeef953815afd8405f8d85001ab0bbb89aa14d2554",
253
+ "variants/cpu-int4/inference/weight_handoff_map.json": "2b809307b885be7b6bd2908831ab37e5f18ea6d867b6fa8be2d284138e94581f",
254
+ "variants/cpu-int4/train/checkpoint": "cd9ea126423c721891377e36136904e64aaa4c04db73f82b8f2732aec482b1be",
255
+ "variants/cpu-int4/train/eval_model.onnx": "b66f69dca13c1f8a43a760f9f5b9a4cad1a9518d4914e8d0e42bd025e5ce6b5c",
256
+ "variants/cpu-int4/train/optimizer_model.onnx": "8ed3b548bce52bf25c2154b1452314b3c48d079e28c52822dd348034f59dc054",
257
+ "variants/cpu-int4/train/trainable_parameters.json": "05fddff407edb77426860ce8637b046bffd277bb053b33aa6600a321ac993958",
258
+ "variants/cpu-int4/train/training_config.json": "4319ebadece693ac91fd1c0d5678e849521b9b3b54b70462df560653f5b1cb74",
259
+ "variants/cpu-int4/train/training_model.onnx": "8e6de37f3cccf0ccaa02815bbeaf774fa1fd0d0bc7e514102950937186a830cb"
260
+ },
261
+ "supportedTasks": [
262
+ "text-generation-with-past"
263
+ ],
264
+ "trainableParameterCount": 368640,
265
+ "trainingParameterCount": 268835639,
266
+ "transformersVersion": "4.57.6",
267
+ "trustRemoteCode": false,
268
+ "variants": [
269
+ {
270
+ "abi": null,
271
+ "executionProvider": "cpu",
272
+ "features": [
273
+ "core",
274
+ "inference",
275
+ "train"
276
+ ],
277
+ "id": "cpu-int4",
278
+ "minimumAndroidApi": 28,
279
+ "paths": {
280
+ "inference": "variants/cpu-int4/inference",
281
+ "tokenizer": "shared/tokenizer",
282
+ "train": "variants/cpu-int4/train"
283
+ },
284
+ "quantization": "int4",
285
+ "recommendedDeviceMemoryMb": null,
286
+ "supportedEngines": [
287
+ "native"
288
+ ],
289
+ "weightHandoff": "variants/cpu-int4/inference/weight_handoff_map.json"
290
+ }
291
+ ],
292
+ "weightHandoff": "variants/cpu-int4/inference/weight_handoff_map.json"
293
+ }
optimum/export_report.json ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "androidRuntime": {
3
+ "minimumAndroidApi": 28,
4
+ "recommendedDeviceMemoryMb": null,
5
+ "requiredAbis": []
6
+ },
7
+ "architectures": [
8
+ "gemma3_text"
9
+ ],
10
+ "license": {
11
+ "baseModelWeights": null,
12
+ "framework": null,
13
+ "noticeFile": null
14
+ },
15
+ "mobiletransformersVersion": "0.1.0",
16
+ "onnxRuntimeTrainingVersion": "1.23.0+cpu",
17
+ "optimumOnnxVersion": "0.1.0",
18
+ "peftMethods": [
19
+ "lora"
20
+ ],
21
+ "quantization": [
22
+ "int4"
23
+ ],
24
+ "selectedTask": "text-generation-with-past",
25
+ "sourceParameterCount": 268098176,
26
+ "supportedTasks": [
27
+ "text-generation-with-past"
28
+ ],
29
+ "trainInferenceLossDeltaNats": 0.8837,
30
+ "trainableParameterCount": 368640,
31
+ "trainingParameterCount": 268835639,
32
+ "trainingQuantizedParameterCount": 253296731,
33
+ "transformersVersion": "4.57.6",
34
+ "trustRemoteCode": false
35
+ }
optimum/supported_tasks.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ [
2
+ "text-generation-with-past"
3
+ ]
shared/tokenizer/chat_template.jinja ADDED
@@ -0,0 +1,279 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- macro format_parameters(properties, required) -%}
2
+ {%- set standard_keys = ['description', 'type', 'properties', 'required', 'nullable'] -%}
3
+ {%- set ns = namespace(found_first=false) -%}
4
+ {%- for key, value in properties | dictsort -%}
5
+ {%- if key not in standard_keys -%}
6
+ {%- if ns.found_first %},{% endif -%}
7
+ {%- set ns.found_first = true -%}
8
+ {{- key }}:{description:<escape>{{ value['description'] }}<escape>
9
+ {%- if value['type'] | upper == 'STRING' -%}
10
+ {%- if value['enum'] -%}
11
+ ,enum:{{ format_argument(value['enum']) }}
12
+ {%- endif -%}
13
+ {%- elif value['type'] | upper == 'OBJECT' -%}
14
+ ,properties:{
15
+ {%- if value['properties'] is defined and value['properties'] is mapping -%}
16
+ {{- format_parameters(value['properties'], value['required'] | default([])) -}}
17
+ {%- elif value is mapping -%}
18
+ {{- format_parameters(value, value['required'] | default([])) -}}
19
+ {%- endif -%}
20
+ }
21
+ {%- if value['required'] -%}
22
+ ,required:[
23
+ {%- for item in value['required'] | default([]) -%}
24
+ <escape>{{- item -}}<escape>
25
+ {%- if not loop.last %},{% endif -%}
26
+ {%- endfor -%}
27
+ ]
28
+ {%- endif -%}
29
+ {%- elif value['type'] | upper == 'ARRAY' -%}
30
+ {%- if value['items'] is mapping and value['items'] -%}
31
+ ,items:{
32
+ {%- set ns_items = namespace(found_first=false) -%}
33
+ {%- for item_key, item_value in value['items'] | dictsort -%}
34
+ {%- if item_value is not none -%}
35
+ {%- if ns_items.found_first %},{% endif -%}
36
+ {%- set ns_items.found_first = true -%}
37
+ {%- if item_key == 'properties' -%}
38
+ properties:{
39
+ {%- if item_value is mapping -%}
40
+ {{- format_parameters(item_value, value['items']['required'] | default([])) -}}
41
+ {%- endif -%}
42
+ }
43
+ {%- elif item_key == 'required' -%}
44
+ required:[
45
+ {%- for req_item in item_value -%}
46
+ <escape>{{- req_item -}}<escape>
47
+ {%- if not loop.last %},{% endif -%}
48
+ {%- endfor -%}
49
+ ]
50
+ {%- elif item_key == 'type' -%}
51
+ {%- if item_value is string -%}
52
+ type:{{ format_argument(item_value | upper) }}
53
+ {%- else -%}
54
+ type:{{ format_argument(item_value | map('upper') | list) }}
55
+ {%- endif -%}
56
+ {%- else -%}
57
+ {{ item_key }}:{{ format_argument(item_value) }}
58
+ {%- endif -%}
59
+ {%- endif -%}
60
+ {%- endfor -%}
61
+ }
62
+ {%- endif -%}
63
+ {%- endif -%}
64
+ ,type:<escape>{{ value['type'] | upper }}<escape>}
65
+ {%- endif -%}
66
+ {%- endfor -%}
67
+ {%- endmacro -%}
68
+ {% macro format_function_declaration(tool_data) -%}
69
+ declaration:{{- tool_data['function']['name'] -}}
70
+ {description:<escape>{{- tool_data['function']['description'] -}}<escape>
71
+ {%- set params = tool_data['function']['parameters'] -%}
72
+ {%- if params -%}
73
+ ,parameters:{
74
+ {%- if params['properties'] -%}
75
+ properties:{ {{- format_parameters(params['properties'], params['required']) -}} },
76
+ {%- endif -%}
77
+ {%- if params['required'] -%}
78
+ required:[
79
+ {%- for item in params['required'] -%}
80
+ <escape>{{- item -}}<escape>
81
+ {{- ',' if not loop.last -}}
82
+ {%- endfor -%}
83
+ ],
84
+ {%- endif -%}
85
+ {%- if params['type'] -%}
86
+ type:<escape>{{- params['type'] | upper -}}<escape>}
87
+ {%- endif -%}
88
+ {%- endif -%}
89
+ }
90
+ {%- endmacro -%}
91
+ {% macro format_argument(argument, escape_keys=True) -%}
92
+ {%- if argument is string -%}
93
+ {{- '<escape>' + argument + '<escape>' -}}
94
+ {%- elif argument is boolean -%}
95
+ {%- if argument -%}
96
+ {{- 'true' -}}
97
+ {%- else -%}
98
+ {{- 'false' -}}
99
+ {%- endif -%}
100
+ {%- elif argument is mapping -%}
101
+ {{- '{' -}}
102
+ {%- set ns = namespace(found_first=false) -%}
103
+ {%- for key, value in argument | dictsort -%}
104
+ {%- if ns.found_first %},{% endif -%}
105
+ {%- set ns.found_first = true -%}
106
+ {%- if escape_keys -%}
107
+ {{- '<escape>' + key + '<escape>' -}}
108
+ {%- else -%}
109
+ {{- key -}}
110
+ {%- endif -%}
111
+ :{{- format_argument(value, escape_keys=escape_keys) -}}
112
+ {%- endfor -%}
113
+ {{- '}' -}}
114
+ {%- elif argument is sequence -%}
115
+ {{- '[' -}}
116
+ {%- for item in argument -%}
117
+ {{- format_argument(item, escape_keys=escape_keys) -}}
118
+ {%- if not loop.last %},{% endif -%}
119
+ {%- endfor -%}
120
+ {{- ']' -}}
121
+ {%- else -%}
122
+ {{- argument -}}
123
+ {%- endif -%}
124
+ {%- endmacro -%}
125
+ {{ bos_token }}
126
+ {%- set ns = namespace(prev_message_type=None) -%}
127
+ {#- Tool Declarations -#}
128
+ {%- set loop_messages = messages -%}
129
+ {%- if tools or messages[0]['role'] == 'system' or messages[0]['role'] == 'developer' -%}
130
+ {{- '<start_of_turn>developer\n' -}}
131
+ {%- if messages[0]['role'] == 'system' or messages[0]['role'] == 'developer' -%}
132
+ {%- if messages[0]['content'] is string -%}
133
+ {{- messages[0]['content'] | trim -}}
134
+ {%- elif messages[0]['content'] is sequence -%}
135
+ {%- for item in messages[0]['content'] -%}
136
+ {%- if item['type'] == 'text' -%}
137
+ {{- item['text'] | trim -}}
138
+ {%- endif -%}
139
+ {%- endfor -%}
140
+ {%- endif -%}
141
+ {%- set loop_messages = messages[1:] -%}
142
+ {%- endif -%}
143
+ {%- if tools -%}
144
+ {%- for tool in tools %}
145
+ {{- '<start_function_declaration>' -}}
146
+ {{- format_function_declaration(tool) | trim }}
147
+ {{- '<end_function_declaration>' -}}
148
+ {%- endfor %}
149
+ {%- endif -%}
150
+ {{- '<end_of_turn>\n' }}
151
+ {%- endif %}
152
+ {#- Loop through messages. -#}
153
+ {%- for message in loop_messages -%}
154
+ {%- if (message['role'] == 'assistant') -%}
155
+ {#- Rename "assistant" to "model". -#}
156
+ {%- set role = "model" -%}
157
+ {%- else -%}
158
+ {%- set role = message['role'] -%}
159
+ {%- endif -%}
160
+ {%- if role != 'tool' -%}
161
+ {%- if ns.prev_message_type != 'tool_response' -%}
162
+ {{- '<start_of_turn>' + role + '\n' }}
163
+ {%- endif -%}
164
+ {%- set ns.prev_message_type = None -%}
165
+ {%- if 'content' in message and message['content'] is not none -%}
166
+ {%- if message['content'] is string -%}
167
+ {{ message['content'] | trim }}
168
+ {%- elif message['content'] is sequence -%}
169
+ {%- for item in message['content'] -%}
170
+ {%- if item['type'] == 'image' -%}
171
+ {{ '<start_of_image>' }}
172
+ {%- elif item['type'] == 'text' -%}
173
+ {{ item['text'] | trim }}
174
+ {%- endif -%}
175
+ {%- endfor -%}
176
+ {%- else -%}
177
+ {{ raise_exception("Invalid content type in user/assistant message") }}
178
+ {%- endif -%}
179
+ {%- set ns.prev_message_type = 'content' -%}
180
+ {%- endif -%}
181
+ {%- if 'tool_calls' in message and message['tool_calls'] and message['tool_calls'] is iterable -%}
182
+ {#- Tool Calls -#}
183
+ {%- for tool_call in message['tool_calls'] -%}
184
+ {% set function = tool_call['function'] %}
185
+ {{- '<start_function_call>call:' + function['name'] + '{' -}}
186
+ {%- if 'arguments' in function -%}
187
+ {%- if function['arguments'] is mapping -%}
188
+ {%- set ns = namespace(found_first=false) -%}
189
+ {%- for key, value in function['arguments'] | dictsort -%}
190
+ {%- if ns.found_first %},{% endif -%}
191
+ {%- set ns.found_first = true -%}
192
+ {{- key -}}:{{- format_argument(value, escape_keys=False) -}}
193
+ {%- endfor -%}
194
+ {%- elif function['arguments'] is string -%}
195
+ {# This handles string-JSON, just in case #}
196
+ {{ function['arguments'] }}
197
+ {%- endif %}
198
+ {%- endif -%}
199
+ {{- '}<end_function_call>' -}}
200
+ {%- endfor -%}
201
+ {%- if loop.last -%}
202
+ {{ '<start_function_response>' }}
203
+ {%- endif -%}
204
+ {%- set ns.prev_message_type = 'tool_call' -%}
205
+ {%- endif -%}
206
+ {%- else -%}
207
+ {#- Tool Responses -#}
208
+ {%- if 'content' in message and message['content'] -%}
209
+ {%- if message['content'] is mapping -%}
210
+ {%- if 'name' in message['content'] and 'response' in message['content'] -%}
211
+ {{ '<start_function_response>response:' + message['content']['name'] | trim + '{' }}
212
+ {%- set response_ns = namespace(found_first=false) -%}
213
+ {%- for key, value in message['content']['response'] | dictsort -%}
214
+ {%- if response_ns.found_first %},{% endif -%}
215
+ {%- set response_ns.found_first = true -%}
216
+ {{- key -}}:{{- format_argument(value, escape_keys=False) -}}
217
+ {%- endfor -%}
218
+ {{- '}<end_function_response>' -}}
219
+ {%- elif 'name' in message -%}
220
+ {{ '<start_function_response>response:' + message['name'] | trim + '{' }}
221
+ {%- set response_ns = namespace(found_first=false) -%}
222
+ {%- for key, value in message['content'] | dictsort -%}
223
+ {%- if response_ns.found_first %},{% endif -%}
224
+ {%- set response_ns.found_first = true -%}
225
+ {{- key -}}:{{- format_argument(value, escape_keys=False) -}}
226
+ {%- endfor -%}
227
+ {{- '}<end_function_response>' -}}
228
+ {%- else -%}
229
+ {{ raise_exception("Invalid tool response mapping: must contain 'name' and 'response' keys, or 'name' must be in the message.") }}
230
+ {%- endif -%}
231
+ {%- elif message['content'] is string -%}
232
+ {%- if 'name' in message -%}
233
+ {{ '<start_function_response>response:' + message['name'] | trim + '{value:' + format_argument(message['content'], escape_keys=False) + '}<end_function_response>' }}
234
+ {%- else -%}
235
+ {{ raise_exception("Invalid tool response: 'name' must be provided.") }}
236
+ {%- endif -%}
237
+ {%- elif message['content'] is sequence -%}
238
+ {%- for item in message['content'] -%}
239
+ {%- if item is mapping -%}
240
+ {%- if 'name' in item and 'response' in item -%}
241
+ {{ '<start_function_response>response:' + item['name'] | trim + '{' }}
242
+ {%- set response_ns = namespace(found_first=false) -%}
243
+ {%- for key, value in item['response'] | dictsort -%}
244
+ {%- if response_ns.found_first %},{% endif -%}
245
+ {%- set response_ns.found_first = true -%}
246
+ {{- key -}}:{{- format_argument(value, escape_keys=False) -}}
247
+ {%- endfor -%}
248
+ {{- '}<end_function_response>' -}}
249
+ {%- elif 'name' in message -%}
250
+ {{ '<start_function_response>response:' + message['name'] | trim + '{' }}
251
+ {%- set response_ns = namespace(found_first=false) -%}
252
+ {%- for key, value in item | dictsort -%}
253
+ {%- if response_ns.found_first %},{% endif -%}
254
+ {%- set response_ns.found_first = true -%}
255
+ {{- key -}}:{{- format_argument(value, escape_keys=False) -}}
256
+ {%- endfor -%}
257
+ {{- '}<end_function_response>' -}}
258
+ {%- else -%}
259
+ {{ raise_exception("Invalid tool response mapping: must contain 'name' and 'response' keys, or 'name' must be in the message.") }}
260
+ {%- endif -%}
261
+ {%- else -%}
262
+ {{ raise_exception("Invalid tool response message: multiple responses must all be mappings") }}
263
+ {%- endif -%}
264
+ {%- endfor -%}
265
+ {%- else -%}
266
+ {{ raise_exception("Invalid content type in tool message: must be mapping, sequence of mappings, or string.") }}
267
+ {%- endif -%}
268
+ {%- endif -%}
269
+ {%- set ns.prev_message_type = 'tool_response' -%}
270
+ {%- endif -%}
271
+ {%- if ns.prev_message_type not in ['tool_call', 'tool_response'] -%}
272
+ {{ '<end_of_turn>\n' }}
273
+ {%- endif -%}
274
+ {%- endfor -%}
275
+ {%- if add_generation_prompt -%}
276
+ {%- if ns.prev_message_type != 'tool_response' -%}
277
+ {{- '<start_of_turn>model\n' -}}
278
+ {%- endif -%}
279
+ {%- endif -%}
shared/tokenizer/mobiletransformers_tokenizer_config.json ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model": {
3
+ "bos_token_id": null,
4
+ "context_length": 2048,
5
+ "num_attention_heads": 12,
6
+ "num_hidden_layers": 12,
7
+ "num_key_value_heads": 12,
8
+ "eos_token_id": [
9
+ 1,
10
+ 50,
11
+ 106
12
+ ],
13
+ "pad_token_id": 1,
14
+ "type": "unknown",
15
+ "vocab_size": 262146
16
+ }
17
+ }
shared/tokenizer/special_tokens_map.json ADDED
@@ -0,0 +1,34 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "boi_token": "<start_of_image>",
3
+ "bos_token": {
4
+ "content": "<bos>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false
9
+ },
10
+ "eoi_token": "<end_of_image>",
11
+ "eos_token": {
12
+ "content": "<eos>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false
17
+ },
18
+ "image_token": "<image_soft_token>",
19
+ "pad_token": {
20
+ "content": "<pad>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false
25
+ },
26
+ "sfr_token": "<start_function_response>",
27
+ "unk_token": {
28
+ "content": "<unk>",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false
33
+ }
34
+ }
shared/tokenizer/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b6b09a0b4a803ad453063ca4bb49a784540e8120004e2450e025df2b27d41fb2
3
+ size 33384899
shared/tokenizer/tokenizer_config.json ADDED
The diff for this file is too large to render. See raw diff
 
variants/cpu-int4/checksums.json ADDED
@@ -0,0 +1,93 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "variants/cpu-int4/inference/chat_template.jinja": "db61fb01017bd82401d3ffca4f8e066cd56ff6d38d2a30c4058770c0bf7ab49b",
3
+ "variants/cpu-int4/inference/config.json": "88a795b85d50b1642e66193935a64c34326fa76efe107f9fceeafd31aa3e2af9",
4
+ "variants/cpu-int4/inference/frozen_base.onnx.data": "9c14b3237169fd6023199017e1949b02bbfe0ef1e80ac10c2d72bba670ebfa82",
5
+ "variants/cpu-int4/inference/genai_config.json": "acac04545d607a84f676049824d2e113f9367179741494a46279551b438c6efc",
6
+ "variants/cpu-int4/inference/generation_config.json": "90f3b346c07b98c000b6de297e805a83f504d6856b01c405a746d805c05444c2",
7
+ "variants/cpu-int4/inference/merger_lora_fpin_fpout.onnx": "953a8ad4704235c19ec06150788107921ba7f2578a5880a32cd59b473544712c",
8
+ "variants/cpu-int4/inference/model.layers.0.self_attn.q_proj.MatMul.weight.bin": "de5441ec8bb19c3a0038e7f3bd3cb25d9b81599d1e6977ceecdc1d894ede2fa1",
9
+ "variants/cpu-int4/inference/model.layers.0.self_attn.q_proj.MatMul.weight.bin.sha256": "b303f006c606265b707027dccaa1c22f9f19b7920a8d51e442cbcfde886d3b65",
10
+ "variants/cpu-int4/inference/model.layers.0.self_attn.v_proj.MatMul.weight.bin": "aced1f8faec6416a6cbee9227400021af9b1f4c0d25feddc54e255f1210575fb",
11
+ "variants/cpu-int4/inference/model.layers.0.self_attn.v_proj.MatMul.weight.bin.sha256": "e41d2d8d47eaea11a011971b193d1ea7f15222a23dc2c2d7342c983d5fad5d58",
12
+ "variants/cpu-int4/inference/model.layers.1.self_attn.q_proj.MatMul.weight.bin": "cc0c039a777068e664ea4916413b822ce497007dc31f22f7e7cc27f9ed0bd820",
13
+ "variants/cpu-int4/inference/model.layers.1.self_attn.q_proj.MatMul.weight.bin.sha256": "1e1651baa468edd1ec0a787271c362646bcf4669d06816750fb9907bda6161ed",
14
+ "variants/cpu-int4/inference/model.layers.1.self_attn.v_proj.MatMul.weight.bin": "e2cde258f00749809d3b5e6a3838dbccb866dfdc228c86e5375a1c3fb3be14b2",
15
+ "variants/cpu-int4/inference/model.layers.1.self_attn.v_proj.MatMul.weight.bin.sha256": "d670153828a24efe77701e8491f5bc182fd88c9f27b924c750b9c20ca4fbf89c",
16
+ "variants/cpu-int4/inference/model.layers.10.self_attn.q_proj.MatMul.weight.bin": "864537f8b063185411edf60cc77bb704f3ddd3d39b6f19f04bd0da2bf1ba5b5e",
17
+ "variants/cpu-int4/inference/model.layers.10.self_attn.q_proj.MatMul.weight.bin.sha256": "286bfb981d94d988671f4572b770728b9a285f400ec4edc6ad83c9ea499d2c40",
18
+ "variants/cpu-int4/inference/model.layers.10.self_attn.v_proj.MatMul.weight.bin": "f8aedd8998857b74a4e93e3182911873cd9ad547f8f0ed8212856e8f0beaef00",
19
+ "variants/cpu-int4/inference/model.layers.10.self_attn.v_proj.MatMul.weight.bin.sha256": "528f124d7e9657be8b57303ec6248483bb0c6dde4c84084575e3c9afd0b8188b",
20
+ "variants/cpu-int4/inference/model.layers.11.self_attn.q_proj.MatMul.weight.bin": "534363c1244562ba5cc4043f27b4da97fa3a48dc2a60772532816b5f11865b91",
21
+ "variants/cpu-int4/inference/model.layers.11.self_attn.q_proj.MatMul.weight.bin.sha256": "19a221e71a3a79d359f1168ea1a3f88003340820f8ecffae5f5472d71be6700c",
22
+ "variants/cpu-int4/inference/model.layers.11.self_attn.v_proj.MatMul.weight.bin": "9b3cc8d43d80eae0f674534447d7845b17a4c779fa74adb1de07254f65de03f0",
23
+ "variants/cpu-int4/inference/model.layers.11.self_attn.v_proj.MatMul.weight.bin.sha256": "643bbdcb8d0381f20eaf58b1d8d9ff7c3f6435ec71d5ee7d4a93d7b60d541a83",
24
+ "variants/cpu-int4/inference/model.layers.12.self_attn.q_proj.MatMul.weight.bin": "5213011869521ddedba25e37944d311b30632ec3b2c061caeb6671d4f4b9596f",
25
+ "variants/cpu-int4/inference/model.layers.12.self_attn.q_proj.MatMul.weight.bin.sha256": "92e8c3fb1b76481b06c2ff2dd3ae3a99ad52477b274511b4ba3944dc9f376858",
26
+ "variants/cpu-int4/inference/model.layers.12.self_attn.v_proj.MatMul.weight.bin": "8dfd86b40b9e613538f07c8297cf3fb54cb3ebf99fe4918c30c8c51bea4da5c0",
27
+ "variants/cpu-int4/inference/model.layers.12.self_attn.v_proj.MatMul.weight.bin.sha256": "ca5d5b8d2723dcb8a31ed895ebd6f72d3111de86981d1f10a57fd64d393a5ceb",
28
+ "variants/cpu-int4/inference/model.layers.13.self_attn.q_proj.MatMul.weight.bin": "3e58dbbd1a83c01943979052519976f15dd2e13b4502fe72edf522f86960d618",
29
+ "variants/cpu-int4/inference/model.layers.13.self_attn.q_proj.MatMul.weight.bin.sha256": "8a75590cded9ed4eb8518b18d4f5694fd4131d8ef225bfff9d148412716fe7c6",
30
+ "variants/cpu-int4/inference/model.layers.13.self_attn.v_proj.MatMul.weight.bin": "21666b649c20492830d5f37db41e2ba54d270228c20a532eea2104fa3c61396e",
31
+ "variants/cpu-int4/inference/model.layers.13.self_attn.v_proj.MatMul.weight.bin.sha256": "327d2d9dd53b71a1662bb5751d554ef719aff537095da4bd8ac2aaaea47de555",
32
+ "variants/cpu-int4/inference/model.layers.14.self_attn.q_proj.MatMul.weight.bin": "ef8d82392a5ac5cf37f65c638a965c7c949111ec783b0c947ecfc8d261e1ad9f",
33
+ "variants/cpu-int4/inference/model.layers.14.self_attn.q_proj.MatMul.weight.bin.sha256": "c183ea62d344b6331a35070bdac7091347be27f629a00a3569acf40f01c87dd2",
34
+ "variants/cpu-int4/inference/model.layers.14.self_attn.v_proj.MatMul.weight.bin": "c1c727d66921e1dceec4142a2a2c36254189234bd90b09db4125927e70b8e458",
35
+ "variants/cpu-int4/inference/model.layers.14.self_attn.v_proj.MatMul.weight.bin.sha256": "75c2d4a4ac3606a197b259bb309ada6a55341a4c3c3e68ab2b5d29985699cd88",
36
+ "variants/cpu-int4/inference/model.layers.15.self_attn.q_proj.MatMul.weight.bin": "31f7849f39f1b1fe0d7ef163aae56279a8f48eac0ec81dceae7dd7fda77dd293",
37
+ "variants/cpu-int4/inference/model.layers.15.self_attn.q_proj.MatMul.weight.bin.sha256": "c314704b435abe1cd3ba672eb1a408e7c6f7c808a215bfb7938d27f5c4583dbe",
38
+ "variants/cpu-int4/inference/model.layers.15.self_attn.v_proj.MatMul.weight.bin": "8f5878be79b76139fc9cc5724e91074c742fffd021b2b7bd6f0406775ec69365",
39
+ "variants/cpu-int4/inference/model.layers.15.self_attn.v_proj.MatMul.weight.bin.sha256": "17e8852305a2dd0be3788629535015c3ba347bd8a84a318c1a2740846ea656d3",
40
+ "variants/cpu-int4/inference/model.layers.16.self_attn.q_proj.MatMul.weight.bin": "cc87b80c44df0b66e627944acd40e87a0b63c77a86d1cfa55085dad5c5610de7",
41
+ "variants/cpu-int4/inference/model.layers.16.self_attn.q_proj.MatMul.weight.bin.sha256": "331eece77b3bc5bea26c224ee395a6c864e2e98667cd6bd0ef8a20c48aeaa27c",
42
+ "variants/cpu-int4/inference/model.layers.16.self_attn.v_proj.MatMul.weight.bin": "0c1f91d46f8f885ca1970f61ada3e69fe7e2a44c4f103beb456e84fb0f39cb13",
43
+ "variants/cpu-int4/inference/model.layers.16.self_attn.v_proj.MatMul.weight.bin.sha256": "c25e491df56cb7902bc53dfe2fdc5f69cc6c94d56a73a1e354920ace0cdedf8c",
44
+ "variants/cpu-int4/inference/model.layers.17.self_attn.q_proj.MatMul.weight.bin": "75aa1f482c3a16420a092f912a073abc981fc584038eead4fc4dad219c4a18df",
45
+ "variants/cpu-int4/inference/model.layers.17.self_attn.q_proj.MatMul.weight.bin.sha256": "5dccc067c68edcfb2aafd0ecfe966965f274afc3455699e6add6a2f4b3c0151f",
46
+ "variants/cpu-int4/inference/model.layers.17.self_attn.v_proj.MatMul.weight.bin": "c331cf9885fc9a470111e2e8e59018a2ffa235ee1441aa005789e321aee80df3",
47
+ "variants/cpu-int4/inference/model.layers.17.self_attn.v_proj.MatMul.weight.bin.sha256": "b3058d47101dd69928aa1b6e78574dd47e8a9eedda003972accac1fcbc50c53f",
48
+ "variants/cpu-int4/inference/model.layers.2.self_attn.q_proj.MatMul.weight.bin": "0aab3b51d26366fda266a9a030b64fa0097a3c9fe0f2ac8b45cd455ed6d00844",
49
+ "variants/cpu-int4/inference/model.layers.2.self_attn.q_proj.MatMul.weight.bin.sha256": "0c44bd5f5b21e9b5292a13d4069c6b313c9b9aa80881c3b469ab0034c0bedd1e",
50
+ "variants/cpu-int4/inference/model.layers.2.self_attn.v_proj.MatMul.weight.bin": "f995a03e11e45e51c9038b610940e2388dea7f89b186c9c32bf0774c7ad56676",
51
+ "variants/cpu-int4/inference/model.layers.2.self_attn.v_proj.MatMul.weight.bin.sha256": "b7f5aa1ce6d2e1be556f65fb6ca619e7591e63dd975534dae239e71290ac5a7b",
52
+ "variants/cpu-int4/inference/model.layers.3.self_attn.q_proj.MatMul.weight.bin": "5f256353f99cd60325cd88345d8095a1d8bf3ec2b2ed8b7dab08368018fd8335",
53
+ "variants/cpu-int4/inference/model.layers.3.self_attn.q_proj.MatMul.weight.bin.sha256": "00990e7e4ec37977e4c3024989b9dedab6f8c94a76755aa59ff21bdde5eae41e",
54
+ "variants/cpu-int4/inference/model.layers.3.self_attn.v_proj.MatMul.weight.bin": "7412ad9549d423c2cfbe4922299f5759b2e30e6839ef89cf99f8880ea8872ee5",
55
+ "variants/cpu-int4/inference/model.layers.3.self_attn.v_proj.MatMul.weight.bin.sha256": "e832cd832df1e7c9249d968196ff8da050eb1f4bb733e6ed1f3df80988037bad",
56
+ "variants/cpu-int4/inference/model.layers.4.self_attn.q_proj.MatMul.weight.bin": "ad0a0d5c5db67a43ea9793d34d5ef09e081772e8e5fba2b771ee95983ecda45a",
57
+ "variants/cpu-int4/inference/model.layers.4.self_attn.q_proj.MatMul.weight.bin.sha256": "6baba6d67fea470291b3da65d1cf9834f6e103b9a6845c8d753c958a2d873b5e",
58
+ "variants/cpu-int4/inference/model.layers.4.self_attn.v_proj.MatMul.weight.bin": "f59967d4272b6bd14e86c0cdc9fed9f1bcb21f67faa8a1d769935f710247f1b3",
59
+ "variants/cpu-int4/inference/model.layers.4.self_attn.v_proj.MatMul.weight.bin.sha256": "294e36d47d7ea2796ec74012dba0d6e672eaade5d8fed5c4fc15885d735fd35a",
60
+ "variants/cpu-int4/inference/model.layers.5.self_attn.q_proj.MatMul.weight.bin": "ff1c5fe943f8959a8b9d09640aec44d881f035f771416ec88ac929e5d4fe6b4b",
61
+ "variants/cpu-int4/inference/model.layers.5.self_attn.q_proj.MatMul.weight.bin.sha256": "aa4e396ee86ecd0b1538e96230fc842659e7ec66c927493682fa914efce3047b",
62
+ "variants/cpu-int4/inference/model.layers.5.self_attn.v_proj.MatMul.weight.bin": "4ee8c2d65e018f46002157c99b0fd025cfc523e03b07ca6fe26da71659f51c70",
63
+ "variants/cpu-int4/inference/model.layers.5.self_attn.v_proj.MatMul.weight.bin.sha256": "31352189895dd2e1d92b88721c0950070cdd2caa8c68c8cae31233825ec5b667",
64
+ "variants/cpu-int4/inference/model.layers.6.self_attn.q_proj.MatMul.weight.bin": "353d2ac53c60dfce33c57e501010ec0dac799d9050617cf6f9f818d37e1a2296",
65
+ "variants/cpu-int4/inference/model.layers.6.self_attn.q_proj.MatMul.weight.bin.sha256": "e869d030595d5ed13ac046acfc08fd9487b3a9ba9e940edbb4724eba90b84167",
66
+ "variants/cpu-int4/inference/model.layers.6.self_attn.v_proj.MatMul.weight.bin": "e5d43600c7a68cc5889563120b46780f107945e26c9385b5fca04c3a22960843",
67
+ "variants/cpu-int4/inference/model.layers.6.self_attn.v_proj.MatMul.weight.bin.sha256": "fd3c65bcc2e5652073fdf480a8be666611a1e20d26662e648e5cdd7fb4e0340c",
68
+ "variants/cpu-int4/inference/model.layers.7.self_attn.q_proj.MatMul.weight.bin": "e9d722fef4ecc4e3dffddd5b1527bcca70af70978eb812df6d7c4452c4c10b90",
69
+ "variants/cpu-int4/inference/model.layers.7.self_attn.q_proj.MatMul.weight.bin.sha256": "5d11e119c112c0f2cd8a46a10856aa4e3a467048928ddd0de0891dec95164b18",
70
+ "variants/cpu-int4/inference/model.layers.7.self_attn.v_proj.MatMul.weight.bin": "d50fe1d3c9c265fc38d169bd63fa727062f145719419564a124a5c06fbe59b04",
71
+ "variants/cpu-int4/inference/model.layers.7.self_attn.v_proj.MatMul.weight.bin.sha256": "bb5995d820468db5e2c676bf0960061c85fff3ccaeebb0c76b89f905b400f0bc",
72
+ "variants/cpu-int4/inference/model.layers.8.self_attn.q_proj.MatMul.weight.bin": "eee1373ed2ee35607cda09ceeef4c92b7ab331a419ffb795c821a58c563e884b",
73
+ "variants/cpu-int4/inference/model.layers.8.self_attn.q_proj.MatMul.weight.bin.sha256": "ebf030acd3a19f40d679c3e8917c15a97e122f4f8076eff7043758fefbef1b4c",
74
+ "variants/cpu-int4/inference/model.layers.8.self_attn.v_proj.MatMul.weight.bin": "3b9e52ee458990d65fd339e4e544d5f379b1600dd4852fc3084a32e96bc00a0b",
75
+ "variants/cpu-int4/inference/model.layers.8.self_attn.v_proj.MatMul.weight.bin.sha256": "6142f9fd4e7ee9f118757cb582ce4d16c3105cb389343ee8aef7dbf4b83b939f",
76
+ "variants/cpu-int4/inference/model.layers.9.self_attn.q_proj.MatMul.weight.bin": "68c5c1b401f009fd33795a986edbef642ed6bb2e8cfdee72f8ef5d9930a608b4",
77
+ "variants/cpu-int4/inference/model.layers.9.self_attn.q_proj.MatMul.weight.bin.sha256": "ec94b226ebc6be2d0eff5dfd48edb112728ec2d7fb4968481f811a87b31a16ae",
78
+ "variants/cpu-int4/inference/model.layers.9.self_attn.v_proj.MatMul.weight.bin": "b029f98f39c8aef6ec3676ae87106eab90d9521ff05ab1c7812963f8a1929555",
79
+ "variants/cpu-int4/inference/model.layers.9.self_attn.v_proj.MatMul.weight.bin.sha256": "9004e7234ef3e8231241955e74f0cfd7880d2ba47a4b6f4f43b3e86759ea7793",
80
+ "variants/cpu-int4/inference/model.onnx": "7816d03ce0be987e340c51ab64916f13a0f70a68630c3a2de82370a32a6fabfa",
81
+ "variants/cpu-int4/inference/model.onnx_data": "984595b520feb718e9dde76cc6dcd2bcc590bb67987c4da01b15a6ea54cade05",
82
+ "variants/cpu-int4/inference/optimum_config.json": "78abd079b5f45b977fad47a63f173701f6681d555c1372a800221831ce811e4f",
83
+ "variants/cpu-int4/inference/special_tokens_map.json": "d39df6f8ddb132a622394f05a603b166ba5f22d485483e30d06c1821ef4ac9b4",
84
+ "variants/cpu-int4/inference/tokenizer.json": "b6b09a0b4a803ad453063ca4bb49a784540e8120004e2450e025df2b27d41fb2",
85
+ "variants/cpu-int4/inference/tokenizer_config.json": "c33878647bdc69ce85d4b9aeef953815afd8405f8d85001ab0bbb89aa14d2554",
86
+ "variants/cpu-int4/inference/weight_handoff_map.json": "2b809307b885be7b6bd2908831ab37e5f18ea6d867b6fa8be2d284138e94581f",
87
+ "variants/cpu-int4/train/checkpoint": "cd9ea126423c721891377e36136904e64aaa4c04db73f82b8f2732aec482b1be",
88
+ "variants/cpu-int4/train/eval_model.onnx": "b66f69dca13c1f8a43a760f9f5b9a4cad1a9518d4914e8d0e42bd025e5ce6b5c",
89
+ "variants/cpu-int4/train/optimizer_model.onnx": "8ed3b548bce52bf25c2154b1452314b3c48d079e28c52822dd348034f59dc054",
90
+ "variants/cpu-int4/train/trainable_parameters.json": "05fddff407edb77426860ce8637b046bffd277bb053b33aa6600a321ac993958",
91
+ "variants/cpu-int4/train/training_config.json": "4319ebadece693ac91fd1c0d5678e849521b9b3b54b70462df560653f5b1cb74",
92
+ "variants/cpu-int4/train/training_model.onnx": "8e6de37f3cccf0ccaa02815bbeaf774fa1fd0d0bc7e514102950937186a830cb"
93
+ }
variants/cpu-int4/inference/chat_template.jinja ADDED
@@ -0,0 +1,279 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- macro format_parameters(properties, required) -%}
2
+ {%- set standard_keys = ['description', 'type', 'properties', 'required', 'nullable'] -%}
3
+ {%- set ns = namespace(found_first=false) -%}
4
+ {%- for key, value in properties | dictsort -%}
5
+ {%- if key not in standard_keys -%}
6
+ {%- if ns.found_first %},{% endif -%}
7
+ {%- set ns.found_first = true -%}
8
+ {{- key }}:{description:<escape>{{ value['description'] }}<escape>
9
+ {%- if value['type'] | upper == 'STRING' -%}
10
+ {%- if value['enum'] -%}
11
+ ,enum:{{ format_argument(value['enum']) }}
12
+ {%- endif -%}
13
+ {%- elif value['type'] | upper == 'OBJECT' -%}
14
+ ,properties:{
15
+ {%- if value['properties'] is defined and value['properties'] is mapping -%}
16
+ {{- format_parameters(value['properties'], value['required'] | default([])) -}}
17
+ {%- elif value is mapping -%}
18
+ {{- format_parameters(value, value['required'] | default([])) -}}
19
+ {%- endif -%}
20
+ }
21
+ {%- if value['required'] -%}
22
+ ,required:[
23
+ {%- for item in value['required'] | default([]) -%}
24
+ <escape>{{- item -}}<escape>
25
+ {%- if not loop.last %},{% endif -%}
26
+ {%- endfor -%}
27
+ ]
28
+ {%- endif -%}
29
+ {%- elif value['type'] | upper == 'ARRAY' -%}
30
+ {%- if value['items'] is mapping and value['items'] -%}
31
+ ,items:{
32
+ {%- set ns_items = namespace(found_first=false) -%}
33
+ {%- for item_key, item_value in value['items'] | dictsort -%}
34
+ {%- if item_value is not none -%}
35
+ {%- if ns_items.found_first %},{% endif -%}
36
+ {%- set ns_items.found_first = true -%}
37
+ {%- if item_key == 'properties' -%}
38
+ properties:{
39
+ {%- if item_value is mapping -%}
40
+ {{- format_parameters(item_value, value['items']['required'] | default([])) -}}
41
+ {%- endif -%}
42
+ }
43
+ {%- elif item_key == 'required' -%}
44
+ required:[
45
+ {%- for req_item in item_value -%}
46
+ <escape>{{- req_item -}}<escape>
47
+ {%- if not loop.last %},{% endif -%}
48
+ {%- endfor -%}
49
+ ]
50
+ {%- elif item_key == 'type' -%}
51
+ {%- if item_value is string -%}
52
+ type:{{ format_argument(item_value | upper) }}
53
+ {%- else -%}
54
+ type:{{ format_argument(item_value | map('upper') | list) }}
55
+ {%- endif -%}
56
+ {%- else -%}
57
+ {{ item_key }}:{{ format_argument(item_value) }}
58
+ {%- endif -%}
59
+ {%- endif -%}
60
+ {%- endfor -%}
61
+ }
62
+ {%- endif -%}
63
+ {%- endif -%}
64
+ ,type:<escape>{{ value['type'] | upper }}<escape>}
65
+ {%- endif -%}
66
+ {%- endfor -%}
67
+ {%- endmacro -%}
68
+ {% macro format_function_declaration(tool_data) -%}
69
+ declaration:{{- tool_data['function']['name'] -}}
70
+ {description:<escape>{{- tool_data['function']['description'] -}}<escape>
71
+ {%- set params = tool_data['function']['parameters'] -%}
72
+ {%- if params -%}
73
+ ,parameters:{
74
+ {%- if params['properties'] -%}
75
+ properties:{ {{- format_parameters(params['properties'], params['required']) -}} },
76
+ {%- endif -%}
77
+ {%- if params['required'] -%}
78
+ required:[
79
+ {%- for item in params['required'] -%}
80
+ <escape>{{- item -}}<escape>
81
+ {{- ',' if not loop.last -}}
82
+ {%- endfor -%}
83
+ ],
84
+ {%- endif -%}
85
+ {%- if params['type'] -%}
86
+ type:<escape>{{- params['type'] | upper -}}<escape>}
87
+ {%- endif -%}
88
+ {%- endif -%}
89
+ }
90
+ {%- endmacro -%}
91
+ {% macro format_argument(argument, escape_keys=True) -%}
92
+ {%- if argument is string -%}
93
+ {{- '<escape>' + argument + '<escape>' -}}
94
+ {%- elif argument is boolean -%}
95
+ {%- if argument -%}
96
+ {{- 'true' -}}
97
+ {%- else -%}
98
+ {{- 'false' -}}
99
+ {%- endif -%}
100
+ {%- elif argument is mapping -%}
101
+ {{- '{' -}}
102
+ {%- set ns = namespace(found_first=false) -%}
103
+ {%- for key, value in argument | dictsort -%}
104
+ {%- if ns.found_first %},{% endif -%}
105
+ {%- set ns.found_first = true -%}
106
+ {%- if escape_keys -%}
107
+ {{- '<escape>' + key + '<escape>' -}}
108
+ {%- else -%}
109
+ {{- key -}}
110
+ {%- endif -%}
111
+ :{{- format_argument(value, escape_keys=escape_keys) -}}
112
+ {%- endfor -%}
113
+ {{- '}' -}}
114
+ {%- elif argument is sequence -%}
115
+ {{- '[' -}}
116
+ {%- for item in argument -%}
117
+ {{- format_argument(item, escape_keys=escape_keys) -}}
118
+ {%- if not loop.last %},{% endif -%}
119
+ {%- endfor -%}
120
+ {{- ']' -}}
121
+ {%- else -%}
122
+ {{- argument -}}
123
+ {%- endif -%}
124
+ {%- endmacro -%}
125
+ {{ bos_token }}
126
+ {%- set ns = namespace(prev_message_type=None) -%}
127
+ {#- Tool Declarations -#}
128
+ {%- set loop_messages = messages -%}
129
+ {%- if tools or messages[0]['role'] == 'system' or messages[0]['role'] == 'developer' -%}
130
+ {{- '<start_of_turn>developer\n' -}}
131
+ {%- if messages[0]['role'] == 'system' or messages[0]['role'] == 'developer' -%}
132
+ {%- if messages[0]['content'] is string -%}
133
+ {{- messages[0]['content'] | trim -}}
134
+ {%- elif messages[0]['content'] is sequence -%}
135
+ {%- for item in messages[0]['content'] -%}
136
+ {%- if item['type'] == 'text' -%}
137
+ {{- item['text'] | trim -}}
138
+ {%- endif -%}
139
+ {%- endfor -%}
140
+ {%- endif -%}
141
+ {%- set loop_messages = messages[1:] -%}
142
+ {%- endif -%}
143
+ {%- if tools -%}
144
+ {%- for tool in tools %}
145
+ {{- '<start_function_declaration>' -}}
146
+ {{- format_function_declaration(tool) | trim }}
147
+ {{- '<end_function_declaration>' -}}
148
+ {%- endfor %}
149
+ {%- endif -%}
150
+ {{- '<end_of_turn>\n' }}
151
+ {%- endif %}
152
+ {#- Loop through messages. -#}
153
+ {%- for message in loop_messages -%}
154
+ {%- if (message['role'] == 'assistant') -%}
155
+ {#- Rename "assistant" to "model". -#}
156
+ {%- set role = "model" -%}
157
+ {%- else -%}
158
+ {%- set role = message['role'] -%}
159
+ {%- endif -%}
160
+ {%- if role != 'tool' -%}
161
+ {%- if ns.prev_message_type != 'tool_response' -%}
162
+ {{- '<start_of_turn>' + role + '\n' }}
163
+ {%- endif -%}
164
+ {%- set ns.prev_message_type = None -%}
165
+ {%- if 'content' in message and message['content'] is not none -%}
166
+ {%- if message['content'] is string -%}
167
+ {{ message['content'] | trim }}
168
+ {%- elif message['content'] is sequence -%}
169
+ {%- for item in message['content'] -%}
170
+ {%- if item['type'] == 'image' -%}
171
+ {{ '<start_of_image>' }}
172
+ {%- elif item['type'] == 'text' -%}
173
+ {{ item['text'] | trim }}
174
+ {%- endif -%}
175
+ {%- endfor -%}
176
+ {%- else -%}
177
+ {{ raise_exception("Invalid content type in user/assistant message") }}
178
+ {%- endif -%}
179
+ {%- set ns.prev_message_type = 'content' -%}
180
+ {%- endif -%}
181
+ {%- if 'tool_calls' in message and message['tool_calls'] and message['tool_calls'] is iterable -%}
182
+ {#- Tool Calls -#}
183
+ {%- for tool_call in message['tool_calls'] -%}
184
+ {% set function = tool_call['function'] %}
185
+ {{- '<start_function_call>call:' + function['name'] + '{' -}}
186
+ {%- if 'arguments' in function -%}
187
+ {%- if function['arguments'] is mapping -%}
188
+ {%- set ns = namespace(found_first=false) -%}
189
+ {%- for key, value in function['arguments'] | dictsort -%}
190
+ {%- if ns.found_first %},{% endif -%}
191
+ {%- set ns.found_first = true -%}
192
+ {{- key -}}:{{- format_argument(value, escape_keys=False) -}}
193
+ {%- endfor -%}
194
+ {%- elif function['arguments'] is string -%}
195
+ {# This handles string-JSON, just in case #}
196
+ {{ function['arguments'] }}
197
+ {%- endif %}
198
+ {%- endif -%}
199
+ {{- '}<end_function_call>' -}}
200
+ {%- endfor -%}
201
+ {%- if loop.last -%}
202
+ {{ '<start_function_response>' }}
203
+ {%- endif -%}
204
+ {%- set ns.prev_message_type = 'tool_call' -%}
205
+ {%- endif -%}
206
+ {%- else -%}
207
+ {#- Tool Responses -#}
208
+ {%- if 'content' in message and message['content'] -%}
209
+ {%- if message['content'] is mapping -%}
210
+ {%- if 'name' in message['content'] and 'response' in message['content'] -%}
211
+ {{ '<start_function_response>response:' + message['content']['name'] | trim + '{' }}
212
+ {%- set response_ns = namespace(found_first=false) -%}
213
+ {%- for key, value in message['content']['response'] | dictsort -%}
214
+ {%- if response_ns.found_first %},{% endif -%}
215
+ {%- set response_ns.found_first = true -%}
216
+ {{- key -}}:{{- format_argument(value, escape_keys=False) -}}
217
+ {%- endfor -%}
218
+ {{- '}<end_function_response>' -}}
219
+ {%- elif 'name' in message -%}
220
+ {{ '<start_function_response>response:' + message['name'] | trim + '{' }}
221
+ {%- set response_ns = namespace(found_first=false) -%}
222
+ {%- for key, value in message['content'] | dictsort -%}
223
+ {%- if response_ns.found_first %},{% endif -%}
224
+ {%- set response_ns.found_first = true -%}
225
+ {{- key -}}:{{- format_argument(value, escape_keys=False) -}}
226
+ {%- endfor -%}
227
+ {{- '}<end_function_response>' -}}
228
+ {%- else -%}
229
+ {{ raise_exception("Invalid tool response mapping: must contain 'name' and 'response' keys, or 'name' must be in the message.") }}
230
+ {%- endif -%}
231
+ {%- elif message['content'] is string -%}
232
+ {%- if 'name' in message -%}
233
+ {{ '<start_function_response>response:' + message['name'] | trim + '{value:' + format_argument(message['content'], escape_keys=False) + '}<end_function_response>' }}
234
+ {%- else -%}
235
+ {{ raise_exception("Invalid tool response: 'name' must be provided.") }}
236
+ {%- endif -%}
237
+ {%- elif message['content'] is sequence -%}
238
+ {%- for item in message['content'] -%}
239
+ {%- if item is mapping -%}
240
+ {%- if 'name' in item and 'response' in item -%}
241
+ {{ '<start_function_response>response:' + item['name'] | trim + '{' }}
242
+ {%- set response_ns = namespace(found_first=false) -%}
243
+ {%- for key, value in item['response'] | dictsort -%}
244
+ {%- if response_ns.found_first %},{% endif -%}
245
+ {%- set response_ns.found_first = true -%}
246
+ {{- key -}}:{{- format_argument(value, escape_keys=False) -}}
247
+ {%- endfor -%}
248
+ {{- '}<end_function_response>' -}}
249
+ {%- elif 'name' in message -%}
250
+ {{ '<start_function_response>response:' + message['name'] | trim + '{' }}
251
+ {%- set response_ns = namespace(found_first=false) -%}
252
+ {%- for key, value in item | dictsort -%}
253
+ {%- if response_ns.found_first %},{% endif -%}
254
+ {%- set response_ns.found_first = true -%}
255
+ {{- key -}}:{{- format_argument(value, escape_keys=False) -}}
256
+ {%- endfor -%}
257
+ {{- '}<end_function_response>' -}}
258
+ {%- else -%}
259
+ {{ raise_exception("Invalid tool response mapping: must contain 'name' and 'response' keys, or 'name' must be in the message.") }}
260
+ {%- endif -%}
261
+ {%- else -%}
262
+ {{ raise_exception("Invalid tool response message: multiple responses must all be mappings") }}
263
+ {%- endif -%}
264
+ {%- endfor -%}
265
+ {%- else -%}
266
+ {{ raise_exception("Invalid content type in tool message: must be mapping, sequence of mappings, or string.") }}
267
+ {%- endif -%}
268
+ {%- endif -%}
269
+ {%- set ns.prev_message_type = 'tool_response' -%}
270
+ {%- endif -%}
271
+ {%- if ns.prev_message_type not in ['tool_call', 'tool_response'] -%}
272
+ {{ '<end_of_turn>\n' }}
273
+ {%- endif -%}
274
+ {%- endfor -%}
275
+ {%- if add_generation_prompt -%}
276
+ {%- if ns.prev_message_type != 'tool_response' -%}
277
+ {{- '<start_of_turn>model\n' -}}
278
+ {%- endif -%}
279
+ {%- endif -%}
variants/cpu-int4/inference/config.json ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_sliding_window_pattern": 6,
3
+ "architectures": [
4
+ "Gemma3ForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "attn_logit_softcapping": null,
9
+ "bos_token_id": 2,
10
+ "dtype": "float32",
11
+ "eos_token_id": [
12
+ 1,
13
+ 50
14
+ ],
15
+ "final_logit_softcapping": null,
16
+ "head_dim": 256,
17
+ "hidden_activation": "gelu_pytorch_tanh",
18
+ "hidden_size": 640,
19
+ "initializer_range": 0.02,
20
+ "intermediate_size": 2048,
21
+ "layer_types": [
22
+ "sliding_attention",
23
+ "sliding_attention",
24
+ "sliding_attention",
25
+ "sliding_attention",
26
+ "sliding_attention",
27
+ "full_attention",
28
+ "sliding_attention",
29
+ "sliding_attention",
30
+ "sliding_attention",
31
+ "sliding_attention",
32
+ "sliding_attention",
33
+ "full_attention",
34
+ "sliding_attention",
35
+ "sliding_attention",
36
+ "sliding_attention",
37
+ "sliding_attention",
38
+ "sliding_attention",
39
+ "full_attention"
40
+ ],
41
+ "max_position_embeddings": 32768,
42
+ "model_type": "gemma3_text",
43
+ "num_attention_heads": 4,
44
+ "num_hidden_layers": 18,
45
+ "num_key_value_heads": 1,
46
+ "pad_token_id": 0,
47
+ "query_pre_attn_scalar": 256,
48
+ "rms_norm_eps": 1e-06,
49
+ "rope_local_base_freq": 10000.0,
50
+ "rope_scaling": null,
51
+ "rope_theta": 1000000.0,
52
+ "sliding_window": 512,
53
+ "transformers_version": "4.57.6",
54
+ "use_bidirectional_attention": false,
55
+ "use_cache": true,
56
+ "vocab_size": 262144
57
+ }
variants/cpu-int4/inference/frozen_base.onnx.data ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9c14b3237169fd6023199017e1949b02bbfe0ef1e80ac10c2d72bba670ebfa82
3
+ size 1684498944
variants/cpu-int4/inference/genai_config.json ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model": {
3
+ "bos_token_id": 2,
4
+ "eos_token_id": [
5
+ 1,
6
+ 50
7
+ ],
8
+ "pad_token_id": 0,
9
+ "context_length": 32768,
10
+ "type": "gemma3_text",
11
+ "vocab_size": 262144,
12
+ "decoder": {
13
+ "session_options": {
14
+ "provider_options": []
15
+ },
16
+ "filename": "model.onnx",
17
+ "head_size": 256,
18
+ "hidden_size": 640,
19
+ "num_attention_heads": 4,
20
+ "num_key_value_heads": 1,
21
+ "num_hidden_layers": 18,
22
+ "inputs": {
23
+ "input_ids": "input_ids",
24
+ "attention_mask": "attention_mask",
25
+ "position_ids": "position_ids",
26
+ "past_key_names": "past_key_values.%d.key",
27
+ "past_value_names": "past_key_values.%d.value"
28
+ },
29
+ "outputs": {
30
+ "logits": "logits",
31
+ "present_key_names": "present.%d.key",
32
+ "present_value_names": "present.%d.value"
33
+ }
34
+ }
35
+ },
36
+ "search": {
37
+ "max_length": 32768,
38
+ "min_length": 0,
39
+ "do_sample": false,
40
+ "top_k": 1,
41
+ "top_p": 1.0,
42
+ "temperature": 1.0,
43
+ "repetition_penalty": 1.0
44
+ }
45
+ }
variants/cpu-int4/inference/generation_config.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "cache_implementation": "hybrid",
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 1,
6
+ 50,
7
+ 106
8
+ ],
9
+ "top_k": 64,
10
+ "top_p": 0.95,
11
+ "transformers_version": "4.57.6"
12
+ }
variants/cpu-int4/inference/merger_lora_fpin_fpout.onnx ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:953a8ad4704235c19ec06150788107921ba7f2578a5880a32cd59b473544712c
3
+ size 811
variants/cpu-int4/inference/model.layers.0.self_attn.q_proj.MatMul.weight.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:de5441ec8bb19c3a0038e7f3bd3cb25d9b81599d1e6977ceecdc1d894ede2fa1
3
+ size 2621440
variants/cpu-int4/inference/model.layers.0.self_attn.q_proj.MatMul.weight.bin.sha256 ADDED
@@ -0,0 +1 @@
 
 
1
+ de5441ec8bb19c3a0038e7f3bd3cb25d9b81599d1e6977ceecdc1d894ede2fa1
variants/cpu-int4/inference/model.layers.0.self_attn.v_proj.MatMul.weight.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aced1f8faec6416a6cbee9227400021af9b1f4c0d25feddc54e255f1210575fb
3
+ size 655360
variants/cpu-int4/inference/model.layers.0.self_attn.v_proj.MatMul.weight.bin.sha256 ADDED
@@ -0,0 +1 @@
 
 
1
+ aced1f8faec6416a6cbee9227400021af9b1f4c0d25feddc54e255f1210575fb
variants/cpu-int4/inference/model.layers.1.self_attn.q_proj.MatMul.weight.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cc0c039a777068e664ea4916413b822ce497007dc31f22f7e7cc27f9ed0bd820
3
+ size 2621440
variants/cpu-int4/inference/model.layers.1.self_attn.q_proj.MatMul.weight.bin.sha256 ADDED
@@ -0,0 +1 @@
 
 
1
+ cc0c039a777068e664ea4916413b822ce497007dc31f22f7e7cc27f9ed0bd820
variants/cpu-int4/inference/model.layers.1.self_attn.v_proj.MatMul.weight.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e2cde258f00749809d3b5e6a3838dbccb866dfdc228c86e5375a1c3fb3be14b2
3
+ size 655360
variants/cpu-int4/inference/model.layers.1.self_attn.v_proj.MatMul.weight.bin.sha256 ADDED
@@ -0,0 +1 @@
 
 
1
+ e2cde258f00749809d3b5e6a3838dbccb866dfdc228c86e5375a1c3fb3be14b2
variants/cpu-int4/inference/model.layers.10.self_attn.q_proj.MatMul.weight.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:864537f8b063185411edf60cc77bb704f3ddd3d39b6f19f04bd0da2bf1ba5b5e
3
+ size 2621440
variants/cpu-int4/inference/model.layers.10.self_attn.q_proj.MatMul.weight.bin.sha256 ADDED
@@ -0,0 +1 @@
 
 
1
+ 864537f8b063185411edf60cc77bb704f3ddd3d39b6f19f04bd0da2bf1ba5b5e
variants/cpu-int4/inference/model.layers.10.self_attn.v_proj.MatMul.weight.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f8aedd8998857b74a4e93e3182911873cd9ad547f8f0ed8212856e8f0beaef00
3
+ size 655360
variants/cpu-int4/inference/model.layers.10.self_attn.v_proj.MatMul.weight.bin.sha256 ADDED
@@ -0,0 +1 @@
 
 
1
+ f8aedd8998857b74a4e93e3182911873cd9ad547f8f0ed8212856e8f0beaef00
variants/cpu-int4/inference/model.layers.11.self_attn.q_proj.MatMul.weight.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:534363c1244562ba5cc4043f27b4da97fa3a48dc2a60772532816b5f11865b91
3
+ size 2621440
variants/cpu-int4/inference/model.layers.11.self_attn.q_proj.MatMul.weight.bin.sha256 ADDED
@@ -0,0 +1 @@
 
 
1
+ 534363c1244562ba5cc4043f27b4da97fa3a48dc2a60772532816b5f11865b91
variants/cpu-int4/inference/model.layers.11.self_attn.v_proj.MatMul.weight.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9b3cc8d43d80eae0f674534447d7845b17a4c779fa74adb1de07254f65de03f0
3
+ size 655360
variants/cpu-int4/inference/model.layers.11.self_attn.v_proj.MatMul.weight.bin.sha256 ADDED
@@ -0,0 +1 @@
 
 
1
+ 9b3cc8d43d80eae0f674534447d7845b17a4c779fa74adb1de07254f65de03f0
variants/cpu-int4/inference/model.layers.12.self_attn.q_proj.MatMul.weight.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5213011869521ddedba25e37944d311b30632ec3b2c061caeb6671d4f4b9596f
3
+ size 2621440
variants/cpu-int4/inference/model.layers.12.self_attn.q_proj.MatMul.weight.bin.sha256 ADDED
@@ -0,0 +1 @@
 
 
1
+ 5213011869521ddedba25e37944d311b30632ec3b2c061caeb6671d4f4b9596f
variants/cpu-int4/inference/model.layers.12.self_attn.v_proj.MatMul.weight.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8dfd86b40b9e613538f07c8297cf3fb54cb3ebf99fe4918c30c8c51bea4da5c0
3
+ size 655360
variants/cpu-int4/inference/model.layers.12.self_attn.v_proj.MatMul.weight.bin.sha256 ADDED
@@ -0,0 +1 @@
 
 
1
+ 8dfd86b40b9e613538f07c8297cf3fb54cb3ebf99fe4918c30c8c51bea4da5c0
variants/cpu-int4/inference/model.layers.13.self_attn.q_proj.MatMul.weight.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3e58dbbd1a83c01943979052519976f15dd2e13b4502fe72edf522f86960d618
3
+ size 2621440
variants/cpu-int4/inference/model.layers.13.self_attn.q_proj.MatMul.weight.bin.sha256 ADDED
@@ -0,0 +1 @@
 
 
1
+ 3e58dbbd1a83c01943979052519976f15dd2e13b4502fe72edf522f86960d618
variants/cpu-int4/inference/model.layers.13.self_attn.v_proj.MatMul.weight.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:21666b649c20492830d5f37db41e2ba54d270228c20a532eea2104fa3c61396e
3
+ size 655360
variants/cpu-int4/inference/model.layers.13.self_attn.v_proj.MatMul.weight.bin.sha256 ADDED
@@ -0,0 +1 @@
 
 
1
+ 21666b649c20492830d5f37db41e2ba54d270228c20a532eea2104fa3c61396e
variants/cpu-int4/inference/model.layers.14.self_attn.q_proj.MatMul.weight.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ef8d82392a5ac5cf37f65c638a965c7c949111ec783b0c947ecfc8d261e1ad9f
3
+ size 2621440
variants/cpu-int4/inference/model.layers.14.self_attn.q_proj.MatMul.weight.bin.sha256 ADDED
@@ -0,0 +1 @@
 
 
1
+ ef8d82392a5ac5cf37f65c638a965c7c949111ec783b0c947ecfc8d261e1ad9f
variants/cpu-int4/inference/model.layers.14.self_attn.v_proj.MatMul.weight.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c1c727d66921e1dceec4142a2a2c36254189234bd90b09db4125927e70b8e458
3
+ size 655360
variants/cpu-int4/inference/model.layers.14.self_attn.v_proj.MatMul.weight.bin.sha256 ADDED
@@ -0,0 +1 @@
 
 
1
+ c1c727d66921e1dceec4142a2a2c36254189234bd90b09db4125927e70b8e458
variants/cpu-int4/inference/model.layers.15.self_attn.q_proj.MatMul.weight.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:31f7849f39f1b1fe0d7ef163aae56279a8f48eac0ec81dceae7dd7fda77dd293
3
+ size 2621440
variants/cpu-int4/inference/model.layers.15.self_attn.q_proj.MatMul.weight.bin.sha256 ADDED
@@ -0,0 +1 @@
 
 
1
+ 31f7849f39f1b1fe0d7ef163aae56279a8f48eac0ec81dceae7dd7fda77dd293
variants/cpu-int4/inference/model.layers.15.self_attn.v_proj.MatMul.weight.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8f5878be79b76139fc9cc5724e91074c742fffd021b2b7bd6f0406775ec69365
3
+ size 655360
variants/cpu-int4/inference/model.layers.15.self_attn.v_proj.MatMul.weight.bin.sha256 ADDED
@@ -0,0 +1 @@
 
 
1
+ 8f5878be79b76139fc9cc5724e91074c742fffd021b2b7bd6f0406775ec69365
variants/cpu-int4/inference/model.layers.16.self_attn.q_proj.MatMul.weight.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cc87b80c44df0b66e627944acd40e87a0b63c77a86d1cfa55085dad5c5610de7
3
+ size 2621440